Functioning and Capability: The Foundations of Sen's and Nussbaum's Development Ethic, Part 2

My task here is to consider in some detail the capability ethic's foundational concepts of functioning and capability, its structure, and its relevance for a reconstruction of the social ideals of freedom, rights, and justice.

A truly developed society, maintains Amartya Sen, would enable humans to be and do, and to live and act, in certain valuable ways. He employs the general terms ‘valuable functionings’ and ‘achievements’ to cover these intrinsically good ‘beings and doings’. Let us first clarify Sen's theory of the nature of functioning before proceeding to discuss his closely related concept of capability.

Sen frequently tries to explain his concept of human functioning by the example of riding a bicycle (see Sen, 1984: 334, Sen, 1985a: 10). Important difference exists between the bicycle, the riding, and any mental state or utility that accompanies the riding. The bicycle itself is a mere object, a commodity. I may own the bike, be near it, and be sitting on it (even when it is moving), and yet not be riding it. To be riding the bike is to be engaged in a purposive human activity with or by means of the bike. The bike is necessary but not sufficient for the cycling. The cycling, as both process and result, is an ‘achievement’ of the rider—as any parent knows when their child first begins to peddle the new bike. While riding, the cyclist may or may not be enjoying herself, satisfying some desire, or getting something out of the activity.

The bicycle example is somewhat misleading if it suggests that intentionality and purposiveness are necessary conditions for human functionings. Sen extends the concept of functioning beyond intentional action to include any ‘state of existence of a person’ (Sen, 1985a: 10). Included would be not only the choosing that initiates the riding but also the mental state—whether one of joy, boredom, or fear—that happens to accompany the activity. Moreover, also included under the concept are states of a person such as their being physically or psychologically fit (caused by the riding). Or to switch examples, the following are all distinguishable functionings related to food: (i) the choosing to eat (ibid.: 69–70; 1987b: 37, 1988c: 282, 294), (ii) the intentional activity of actually eating, (iii) the enjoyment (or its lack) in eating, (iv) the process of digesting the food, (v) the state of being nourished or free from malnourishment, and (vi) the subsequent intentional activities, such as working or playing, causally made possible by being nourished.

In relation to Sen's list of the types of functioning, Martha Nussbaum's concept of functioning is somewhat narrower. First, although Sen conceives of choosing (category i) as a distinguishable (intentional, mental, inner) functioning, Nussbaum understands choosings as not more than the voluntary or chosen dimension of an intentional human functioning. For Nussbaum, choosings without functionings would be more transcendental than human: the acts of will or disembodied angels, godlings, or Cartesian egos. Likewise, processes without choosings (category iv) would be less than human; for example, ‘the sleeper's life of non‐guided digestive functioning’ (Nussbaum, forthcoming a), the lives of pigs (Nussbaum, 1990b: 211), and presumably the movements of compleat robots. One reason for Nussbaum's view is that she seems uneasy about a model in which choosings are inner acts of will.¹ As we shall see, this difference between Sen and Nussbaum has implications for the way in which they individuate functionings and for their views on human well‐being and its relation to agency.

Although Nussbaum is reluctant to endorse Sen's view that choosing is a separate functioning, she does assume that choosing, as a component of human functioning, is made possible by a distinguishable personal capabilitythat can be more or less fully ‘developed’ and exercised: ‘One of the capabilities Aristotelian government most centrally promotes is the capability of choosing: of doing all these [valuable] functions in accordance with one's own practical reasoning’ (Nussbaum, 1990a: 214, note omitted). Moreover, the following statement indicates her willingness to view practical wisdom, of which choosing is a part, as a functioning: ‘The exercise of practical wisdom is itself a human excellence, an activity of intrinsic value apart from its tendency to produce virtuous actions’ (Nussbaum, 1990b: 92).

A second difference in their respective concepts of human functioning concerns the mental states (category iii, above) of happiness or pleasure (or their opposites). Sen conceives such mental states as distinguishable functionings that have intrinsic value and can be pursued as such. Nussbaum, on the other hand, takes what she believes to be a less utilitarian and more Aristotelian position. Although she counts ‘being able to have pleasurable experiences’ as one of the valuable human functional capabilities, she refuses to make pleasure a separable functioning. Pleasure or satisfaction, argues Nussbaum, is supervenient on functioning rather than itself a functioning:

According to EN X [Aristotle, Nicomachean Ethics, Bk X] pleasure supervenes upon the activity to which it attaches, like the bloom on the cheek of a healthy young person, completing or perfecting it. Here pleasure is not identical with the activity; but it cannot be identified without reference to the activity to which it attaches. It cannot be pursued on its own without conceptual incoherence, any more than blooming cheeks can be cultivated in isolation from the health and bodily fitness with which they belong. Still less could there be a single item, pleasure, that is separable from all the activities and yielded up by all of them in differing quantities. (Nussbaum, 1990b: 57; see Nussbaum 1986a: 294–5)

Regardless of the extension of the concept, both Sen and Nussbaum conceive of humans as functioning in a variety of ways. Sen defines a person's ‘achieved living’ (Sen, 1990c: 113) as the person's combined ‘doings and beings’ (ibid.: 113), ‘the set of functions a person actually achieves’ (Sen, 1985d: 198):

Given n different types of functionings, an n‐tuple of functionings represents the focal features of a person's living, with each of its n components reflecting the extent of the achievements of a particular functioning. (Sen, 1990c: 114)²

What is the ethical significance of this notion of functioning? The concept of functioning couple with the (about to be discussed) notion of capability for functioning, provides Sen with a complex interpretation of human well‐being and deprivation: the ‘primary feature of a person's well‐being is the functioning vector that he or she achieves’ (Sen, 1985d: 198).³ Moreover, this interpretation ‘builds on the straightforward fact that how well a person is must be a matter of what kind of life he or she is living, and what the person is succeeding in “doing” or “being’ ” (Sen, 1985a: 46).

In contrast, the rival approaches are restricted to other, less urgent sorts of information. What the commodity approach values is, at best, only a means to well‐being. Given interpersonal variability, different amounts and kinds of goods can result in the same sort and level of functioning (and freedom to function). And the same kinds and amounts of goods can result in wildly different levels of achievement (and freedom to achieve). A focus on functioning enables us to keep very clear about the constant ends and the variable means of development. The welfarist perspective, concerned only with the goal of utilities, neglects or ‘muffles’ all other sorts of human functioning. Happiness may be coupled with malfunctioning, and discontent may accompany or spur the most important of activities. Even the discipline of development economics has been one‐sided, for it has emphasized quantity of life (longevity) and neglected the quality of the lives that are led, for example, being healthy and being educated (Sen, 1985a: 46). At this point, we have not treated Sen's and Nussbaum's views concerning which achievements are genuine. We do know, however, that development is for people and the lives they lead rather than merely a matter of whether they possess certain goods or satisfy certain preferences.

Before analysing Sen's and Nussbaum's related notion of capability for functioning, it is important to stress the normative role of functioning. G. A. Cohen, although correctly seeing how important capabilities are in Sen's ethics, fails to recognize that Sen also gives independent and intrinsic value to certain functionings.⁴ Sen does put more normative emphasis—with respect to responsible adults—on ‘freedom to achieve valuable ways of functioning’ (Sen, 1990a: 52) than he does on the valuable functionings themselves. He does recognize, however, that the more valued (Sen, 1985d: 20) and valuable are one's functionings, the greater is the extent of one's freedom to function. One reason that reducing someone's freedom is bad is that it decreases her opportunities for achieving valuable functionings.

Nussbaum is both more explicit about the intrinsic value of certain functionings and gives a more balanced account of the comparative value of capabilities and functionings. For Nussbaum, valuable functionings and valuable capabilities to function are interconnected and are distinguishable sources of value:

Capability‐needs are important because of the value of the functionings in which they naturally terminate; functionings are valuable, in part, for the way in which they realize capabilities. We cannot and should not prise the two apart. (Nussbaum, 1988a: 169)

The capability for good health is valuable because inter alia healthy bodily functioning is valuable. The capability of concern for others derives part of its value from the value of compassion as an action or achievement. ‘Freedom for what?’, we might say, is a question that cannot be replaced by ‘Is there freedom?’

Nussbaum also stresses actual functioning as a platform for free choice. Public action should be concerned that human beings actually function at certain minimal levels such that they are free to choose to advance beyond or retreat from that level. A very sick person may not even be in a position to decide whether to strive for some level or other of healthy functioning. Only if a young person can read at some level, is she sufficiently informed to be able to decide to improve or abandon her reading. Responsible government not only gets citizens to minimal levels of actual functioning and capability; it also non‐coercively encourages people to choose to function in valuable rather than trivial or evil ways.

We shall see below that this dual source of value reveals a difference between Sen and Nussbaum. It also presents, I argue, a problem of the relative value of capabilities and functionings that neither Sen nor Nussbaum fully confront. That problem only emerges because these two thinkers are sensitive to the ways in which both actual achievements and the freedom to achieve are intrinsically valuable ingredients in a good human life.

It is not enough to single out certain functionings as the content of the good human life, as the ultimate end of development. As Aristotle says, a distinction should be made between actuality and possibility. An important difference exists, for example, between a stone and a sleeping human, with respect to some activity like cycling. Neither the stone nor the sleeping cyclist are engaged in riding. Only the cyclist, however, can ride or is capable of cycling.⁵ For Sen and Nussbaum, development is the promotion and expansion of valuable capabilities. Accordingly, to grasp their development ethic we must be clear on their concept of the nature, importance, and varieties of human capabilities.

We must ask several questions, not all of which have been answered by Sen or Nussbaum. What, precisely, is meant by ‘capability?’ How do capabilities relate to functionings, on the one hand, and to freedoms, on the other? Given the high evaluation, just analysed, of actual functionings, why posit capabilities and insist on their intrinsic importance? Is there a problem—in general or in specific cases—of the relative worth of actual and possible functionings? And, most importantly, for the construction of an action‐guiding ethic, which functional capabilities are most valuable and why? Finally, we shall identify, in a provisional way, some strengths and weaknesses of this approach.

What sorts of things are the capabilities that Sen proposes? A person's combination of actual functionings, her ‘functioning vector’, is the particular life she actually leads. The person is leading this life of ‘beings and doings’ but could lead alternative lives. The person's ‘capability set’ (Sen, 1985d: 201; Sen, 1988c: 289) is the total set of functionings that are ‘feasible’, that are within her reach, that the person could choose:

While the entitlement of a person is a set of alternative commodity bundles, the capability of a person is a set of alternative functioning bundles. (Sen and Drèze, 1989: 13, n. 21; emphasis in text)

A person's capability set can be defined as the set of functioning vectors within his or her reach. In examining the well‐being aspect of a person, attention can legitimately be paid to the capability set of the person and not just to the chosen functioning vector. This has the effect of taking note of the positive freedoms in a general sense (the freedom ‘to do this’, or ‘to be that’) that a person has. (Sen, 1985d: 200–1; emphasis in text)

The notion of ‘capability’ was introduced . . . to refer to the extent of the freedom that people have in pursuing valuable activities or functionings. (Sen and Drèze, 1989: 42; see also Sen, 1990c: 114)

On this conception, two people could have the same capability set and choose different bundles of actual functionings. Conversely, they could have different capability sets and choose the same (sorts of) functionings (Sen, 1990c: 116). One of Sen's One of Sen's favourite examples of the latter also amounts to an argument for adding capabilities to the moral space of functionings. Both the person starving and the person fasting—for example, the Somali refugee, and the Irish hunger striker—exemplify the functioning of being severely undernourished. But, it is clear, the two do not enjoy ‘the same level of well‐being’ (Sen, 1988b: 17). The difference lies in the absence of certain options for the one and the presence of these options for the other. The former is not free not to be severely undernourished nor to function in many other undesirable ways. The latter, in contrast, has the significant capability not to starve: ‘B [the faster] could have in a straightforward sense, chosen an alternative life style which A [the non‐faster] could not have chosen’ (Sen, 1985b: 201; emphasis in text).

Sen entertains but does not decide between two alternative ways of describing this sort of contrast (Sen, 1985a: 17–18; Sen, 1988c: 290–1; Sen, 1985d: 201–2). On the one hand, we can say that the faster's ‘freedom to choose’ or ability to choose not to starve herself is part of the faster's enhanced well‐being in comparison to the person involuntarily starving. Here the list of functionings does not include the act of choosing: the value of the choosing is reflected both in the actual functioning chosen (the starving) and in ‘the nature and the range of the capability set itself’ (Sen, 1988b: 17). Alternatively, we can view the act of ‘choosing’ as itself a functioning. Then, we can say that the faster's choosing to starve is one of her functionings. It can be called a ‘refined functioning’ (ibid.: 18) and her functioning of starving—but not the non‐faster's—can then be redescribed as the functioning of ‘fasting’. Only starving that is chosen in the context of other alternatives can be called fasting: ‘Choosing to do x when one could have chosen any member of a set S, can be defined as a “refined functioning’ ” (ibid.: 18; see Sen, 1987b: 37).⁶

Regardless of how described, what is important about Sen's discussion is that it gives us several reasons, in interpreting human well‐being or deprivation, to add the category of ‘capability to function’ to the category of (unrefined) functioning. One reason that valuable functionings are valuable is that they realize valuable capabilities. Moreover, valuable functionings gain some of their value from the fact that they are chosen, ‘done in accordance with practical reason’ (Nussbaum, 1990a: 214), rather than determined or necessitated. Further, even though I am not now functioning in a valuable way, it is good that I have an array of options and even better when this array includes valuable functionings. Capabilities add something intrinsically and not merely instrumentally valuable to a human life, namely, positive freedom in the sense of available and worthwhile options. Finally, capabilities as well as functionings are important in grasping the aim and limits of good government. Responsible law‐makers and development policy‐makers aim at getting people over a threshold—of minimal human and valuable functionings—so that they are able, if they so choose, to function in more fully human ways. The purpose is not, as Rawls fears, to impose a certain conception of the good life on human beings but to enable them to cross a threshold so that they have certain choices. Nussbaum puts it well:

The conception [Aristotelian social democracy] does not aim directly at producing people who function in certain ways. It aims, instead, at producing people who are capable of functioning in these ways; who have both the training and the resources to so function, should they choose. The choice itself is left to them. And one of the capabilities Aristotelian government most centrally promotes is the capability of choosing; of doing all these functions in accordance with one's own practical reason . . . The government aims at capabilities, and leaves the rest to the citizens. (Nussbaum, 1990a: 214; emphasis in text)

We need, however, to probe further Sen's and Nussbaum's conceptions of capability. Not only is the notion of capability susceptible to different interpretations, but a close reading reveals an important difference between the two thinkers. Let us begin with Sen's conception. What sorts of things are the capabilities that Sen describes? At least five interpretations are possible. Capabilities might be construed as (i) inclinations or desires, (ii) needs, (iii) concrete skills, (iv) general character traits, or (v) possibilities or opportunities. Let us look at each candidate in turn.

(i) It is clear that Sen does not identify capabilities with either inclinations or desires. The faster, who is capable (in Sen's sense) of being well‐nourished, is not, all things considered, favourably disposed to be well‐nourished or desires to be so. (ii) Likewise, capabilities need not be needs. Someone could have the capability of fasting but no need to fast unless she had certain political commitments. Also someone might have a biological need to be well‐nourished; but, if she had decided to fast, Sen would say she was capable of being well‐nourished but had no such need.

(iii) The relation of capabilities and concrete abilities or skills is more complicated. We need to be cautious here, because one ordinary use of ‘capability’ is that of ‘ability’ or ‘skill’. A good midfielder in soccer must be capable of accurate passing, playmaking, and dribbling. And, Sen sometimes explicitly defines capabilities as abilities: ‘A functioning is an achievement, whereas a capability is the ability to achieve.’ (Sen, 1987b: 36) This definition, however, does not help much because Sen is using ‘ability’ in this context in a way that is similar in breadth to his use of the term ‘capability’. For Sen, to say that someone has the capability or ability to move about freely (Sen and Williams, 1982: 20) is to speak not of powers, skills, or other traits possessed by the person but rather of possibilities or options facing the person. Moreover, for Sen, many of the functionings of which humans are capable are not intentional actions at all. Rather, they are states of being, such as being healthy or being free of malaria, that are not identical with skills. (There are certain skills useful for becoming or staying healthy, but being healthy is functioning in a healthy way and not itself a skill.)

(iv) A fourth theory of capability and its relation to functioning would be to conceive capabilities not as concrete skills, such as a surgeon's ability to use a scalpel, but as more general personal traits such as the capacity or power of a person to move about, imagine, or reason. So understood, capabilities would exhibit what Nussbaum calls different ‘levels’: they would be formed from an ‘undeveloped’ or latent state (a capacity for a capacity), maintained and exercised in one's maturity, and diminished or lost in old age. In my judgement, it is this interpretation of capabilities, as general powers of the person, that best fits Nussbaum's concept of internal capability (Nussbaum, 1988a: 160–72). Notice, for instance, how Nussbaum (misleadingly, I later argue), distinguishes between what she calls ‘internal’ and ‘external’ capabilities:

Internal capabilities are conditions of the person (of body, mind, character) that make that person in a state of readiness to choose the various valued functions. External capabilities are internal capabilities plus the external material and social conditions that make available to the individual the option of that valued function. (Nussbaum, 1990a: 228)

Nussbaum's ‘state of readiness to choose’ an actual functioning would be based on or include general powers that can be nurtured, acquired, developed, maintained, exercised, impeded, diminished, lost and (sometimes) restored.⁷ There personal powers are (or fail to be) realized, embodied or expressed concrete activities. Good actions, which for Nussbaum (following Aristotle) compose ‘flourishing living’ (eudaimonia) (Nussbaum, 1986a: 6), would embody the best of these internal potentials.

Having these internal powers is necessary but not sufficient for good functioning, for one must also have available certain ‘external and social conditions’. Suppose that the skill of riding a bicycle were one of valuable general capabilities, as conceived by Nussbaum. To perform the function of riding requires that one has (or immediately acquires) the internal ability to ride, access to a bike, and no social conditions that hinder bike riding. Instead of saying that external capabilities include both internal capabilities and external conditions, it would be more perspicuous if Nussbaum said that functioning both realizes internal capabilities and requires externalopportunities. These external options would depend on access to resources, the presence of enabling conditions (such as legal rights), and the absence of preventing conditions (such as legal prohibitions or threatening bayonets). As we shall see, on this view the task of government is to help its citizens acquire the actual or developed capability (as internal power) and ‘to make sure that all citizens have the necessary resources and conditions’ for acting, if they so choose, in the valued ways (Nussbaum, forthcoming b). Nussbaum's account appropriately emphasizes that good societies and good development promote, through various institutions and practices, good human development. Responsible institutions promote the formation, exercise, maintenance, strengthening, and restoration of certain good human powers.⁸

(v) The best interpretation of what Sen means by ‘capability’ is that it connotes a certain sort of possibility or opportunity for functioning. My claim is that, in contrast to Nussbaum, Sen restricts capabilities to opportunities. For Sen, capabilities are options or choices open to the person, possible functionings from which a person may choose.

What sort of possibility? Obviously not logical possibility, for it is not a logical contradiction that precludes the starving person from eating. Here an interchange between Bernard Williams and Sen is instructive (Sen, 1987b: 99–100, 109). Sen would say that someone living in Los Angeles lacks the capability of breathing unpolluted air. Williams, on the other hand, thinks Sen should say that this inhabitant lacks the ‘ability’ (ibid.: 99) of breathing unpolluted air ‘here and now’ (ibid.: 99), but has the general capability to breathe unpolluted air and could realize the capability by migrating to another location.

Sen's response is brief but revealing. First he agrees with another point that Williams makes, namely, that we must not think of capabilities singly but rather about ‘sets of co‐realizable capabilities’ (ibid.: 109). Sen's way of putting the point is that capabilities are members of sets of capabilities and that such sets are ‘sets of n‐tuple functionings from which the person can choose any one n‐tuple’ (Sen, 1990c: 114; see Sen, 1987b: 109).

Sen means that we cannot simply ask whether a Los Angeles inhabitant has the capability of breathing fresh air. For the question would have to address a particular Los Angeles resident's set of co‐realizable possibilities. One of those n‐tuples presumably would include staying, due to irremediable lack of means, in an area that remains permanently beset with pollution. Another set would include the resident's possibility, due to (present or potential) wealth or (reckless) desperation, of migrating to a locale with fresh air. About the resident so conceived, Sen says that we can say that prior to migration she had the requisite capability to breathe unpolluted air because ‘that alternative must be seen in terms of the post‐migration n‐tuple of all functionings’ (ibid.: 109)—obviously including the living in a place with fresh air. Depending on her external options, however, there will be some point at which we can say that the Los Angeles resident has no (or little) capability for breathing fresh air due to the fact that her lack of real options makes it impossible for her to leave Los Angeles (or possible only with extreme risk or cost). On Sen's view, the issue for this Los Angeles resident is not whether to migrate from Los Angeles or clean it up so that she can exercise some internal ability to breathe clean air. The issue is, given that something can be done, whether or not it is worth giving up other options and gaining the option of breathing clean air by, for example, reducing the pollution in Los Angeles or moving somewhere with clean air.⁹

Hence, for Sen, capabilities are not powers of the person that might or might not be realized in different situations. They are, rather, options (sets of compossible options) for actions. Sen's concept of capability as possibility focuses rather on options for actions. These options may refer to but are not identical with traits of a person.

Sen, of course, could take personal traits into consideration to specify a person's capabilities as opportunities. One's internal powers would be relevant as means that make us free to do or be in certain ways. This response, however, would indicate a misunderstanding of Nussbaum's view and weaken Sen's own view. For Nussbaum, it is important conceptually to characterize and institutionally to promote valuable internal features of human beings if these citizens are to be able to take advantage of valuable external opportunities (resources and options). Moreover, Nussbaum contends that valuable capabilities to function, including the power of choice, are internal constitutive conditions of valuable functionings rather than merely contingent means to do and be. In valuable functioning we realize, exercise, and celebrate in action our valuable capabilities rather than merely use them, like Rawlsian social primary goods, as means to other things. In riding the bike, the child, having acquired the capability to ride the bike, realizes the power in riding rather than merely using it as a means to ride. (Riding, of course, also may be a means to other ends.) Moreover, if our internal powers were construed as merely means to good ends, then Nussbaum would rightly argue that the view threatens to ‘commodify the self’ (Nussbaum, forthcoming b). Just as economic utilitarianism commodifies the self by putting monetary value on human functionings, so construing capabilities only as means to functionings would alienate our powers from our activities.

Sen, I think, makes a minor mistake in his interchange with Williams. Williams had raised the question of whether ‘everything that Sen counts as a capability is directly related, at least, to choice’ (Sen, 1987b: 98). Sen responds by arguing that ‘we sometimes have the ability to do things in a genuine sense without being able to do their opposite’ (ibid.: 111). Sen's argument is that ‘Ann has the ability not to marry Bill’ (ibid.: 111) without her having the capability of marrying him. And Bill has the capability of ending his life without always being able to keep on living (ibid.: 111). To use the above Los Angeles pollution example, Sen would say that Magic Johnson has the capability of breathing polluted air in Los Angeles without implying that he can breathe unpolluted air in Los Angeles. Sen's mistake here is to confuse choice of the opposite or contrary capability with choice of a contradictory option.

Sen is correct that Ann's capability not to marry Bill does not entail her marriage to Bill. If, however, she has the capability of rejecting Bill's proposal, she also has the capability of not rejecting it: either by changing the subject, asking for more time, or accepting it. Likewise, Bill's capability to take his life does not ensure survival, for he may be dying rapidly of inoperable cancer. But if Bill is capable of suicide he must be capable of non‐suicide. He must have some choice, even though it does not include being able to continue living for more than a short time. He can decide not to kill himself and instead postpone his decision or decide to ‘let himself’ die of natural causes. If the latter, he can still decide whether to follow Dylan Thomas and ‘rage against the dying of the light’, adopt Nussbaum's stance and moderate the rage but still be fearful (Nussbaum, 1990b: 379), look forward to eternal bliss, or postpone this decision. That Magic has no capability of breathing unpolluted air in Los Angeles does not eliminate his choice to breathe. The general conclusion is that although the capability to do X does not imply the ability to do the opposite; choice of some sort is essential to capabilities as Sen conceives them. That our team is able not to lose, because it can attempt a tying field goal, does not necessarily entail winning.

Sen and Nussbaum sketch several distinct types of functionings and types of capabilities. We have already seen one distinction on which Nussbaum insists but that Sen, neglecting the internal dimensions of capabilities, fails to make, namely, a distinction between ‘levels’ of capability (see Nussbaum, 1988a: 160–72). In Nussbaum's typology, a basic capability is an undeveloped or potential capacity. When this potential is actualized, through nurture and maturation, the result is an ‘internal’ capability, which can be exercised or realized in the correlative functioning. An agent's capability is also ‘external’ to the extent that there are no external circumstances that block or prevent the realization in action of the internal capability. Although Sen construes capabilities as opportunities rather than personal powers, he does make a distinction analogous to Nussbaum's levels. As we have seen, he distinguishes between those opportunities that are more or less feasible. Here feasibility concerns not only empirical likelihood but also normative costs and benefits.

Sen identifies several additional types of functionings and capabilities. First, functionings and capabilities may be referred to either positively or negatively. For instance, not being diseased would be part of the positive functioning of being healthy. Secondly, actual and possible functionings can be described more or less hierarchically. The general capability of being able to be free of avoidable morbidity is further specified by being capable of being free from malaria. Being able to ride a bicycle presupposes and specifies being able to move about. The most inclusive or general capability would be the ‘capability to function well’ (Sen, 1985d: 200) or, as Nussbaum expresses it, the ‘capability to live a rich and fully human life, up to the limit permitted by natural possibilities’ (Nussbaum, 1990a: 217). Thirdly, functionings and capabilities can differ with reference to the activities of others. We have seen that the capability to appear in public without shame has a reference to the judgements of others in a way that is not true of the capability to be able to move freely. Moreover, some functionings and capabilities are more or less universal, shared or shareable by (almost) all human beings. Some, like the capability to play wide receiver, and not just the culturally relative goods that contribute to them, are specific to particular times, places, and physical abilities.

Sen's fourth classification of capabilities and functionings is based on a distinction between the well‐being and agency dimensions of human beings (see Sen, 1985d; 1987a). When applied to functionings and capabilities, this dichotomy issues in two distinctions, one between well‐being and agency functionings, and the other between well‐being and agency capabilities. The basis for the distinction is one of part and whole. Well‐being freedom and well‐being functioning involve choices concerning one's own advantage. Agency freedom and achievements include the agent's well‐being but can also be concerned with people and purposes other than and maybe at odds with the agent's well‐being.¹⁰ Hence, Sen provides conceptual space for a Kantian component of agency and breaks decisively with any psychological (or ethical) egoism that claims that humans are no more than ‘strict maximizers of a narrowly defined self‐interest’ (Sen, 1990a: 54). Insofar as humans can devote themselves to people and causes beyond and against their own welfare, Sen can answer a sceptical realist's concern about any normative theory that proposes a just treatment of conflicting interests or freedoms:

If conflicts of interest are very sharp and extensive, the practical feasibility and actual emergence of just social arrangements may pose deep problems. There are reasons for skepticism here, but the extent and force of that skepticism must depend on the view we take of human beings as social persons. If individuals do, in fact, incessantly and uncompromisingly advance only their narrow self‐interests, then the pursuit of justice will be hampered at every step by the opposition of everyone who has something to lose from any proposed change. If, on the other hand, individuals as social persons have broader values and objectives, including sympathy for others and commitment to ethical norms, then the promotion of social justice need not face unremitting opposition at every move. (Sen, 1990a: 54)

Sen goes on to provide empirical filling for this sort of altruistic conceptual space by referring to his own empirical work (Sen, 1982a) and that of many other social scientists, such as Albert Hirschman (Sen, 1990a: 54; Sen, 1987a: 16–28). He also marshals evidence from momentous events suggested by the names ‘Prague or Paris or Warsaw or Beijing or Little Rock or Johannesburg’ as evidence that ‘among the things that seem to move people . . . are concern for others and regard for ideas’ (Sen, 1990a: 54).

Nussbaum, who agrees with Sen about the complexities of human motivation, softens his sharp distinction between well‐being (capabilities and achievements) and agency (capabilities and achievements). Human well‐being, she argues, consists only of those functionings that are both chosen and valuable. Choosing, rather than a separate functioning, is integral to all valuable functioning. Hence, she finds that she has no need to distinguish two different types of human achievement. Secondly, she does not restrict human well‐being to personal advantage and, hence, there is no need to open conceptual space for the human agent to be able to choose other‐directed actions. Instead, she persuasively includes two sorts of social virtues—social attachments and social affiliation (acting ‘with and toward others’)—as two sorts of good functionings that compose well‐being (Nussbaum, 1990b: 98; 1990a: 226). She even goes so far as to designate affiliation, along with practical reason (including choice), as one of the two ‘architectonic functionings’ (ibid.: 226) that pervade all other valuable functionings. Sen conceives well‐being achievements as a (self‐interested) sub‐class within agency achievement. Nussbaum, more Aristotelian and less Kantian, understands practical reason (which includes agency) and affiliation as two supreme human functions that ‘infuse’ and ‘arrange all of the others, giving them in the process a characteristically human shape’ (ibid.: 226). Hence, although dualities exist in both Sen's and Nussbaum's perspectives, they are drawn along different lines and serve different purposes. Perhaps Sen, addressing economists, has to start with a conception of humans as motivated by self‐interest and defend a more expansive view that makes room for altruism and sacrifice. Nussbaum, more at home with the world of Greek thought, finds it difficult to draw the sharp modern distinction between individual and communal good.

Finally, and most importantly, capabilities and functionings can be ranked from the trivial to the important (Sen, 1987b: 108). Good development achieves the expansion of (valuable) capabilities and the promotion of valuable functionings. In the next section, I review the procedure that Sen and Nussbaum advocate for determining value and what results from their use of that method. Before doing so, however, I pause to raise a general issue about the valuation of capabilities. Sen views capabilities as ranging from the very valuable to the trivial, but he does not construe any capabilities as morally bad or evil. Why?

So far, Sen has neglected this question in his writings. However, I believe he would want to argue against the existence of morally evil capabilities in the following way.¹¹ Capabilities, he might argue, exist only in the realm of possibility. We are capable of doing many terrible things such as strangling our new born babies, running over pedestrians, poisoning wells, and torturing people. The evil, however, should be ascribed to the functionings, the actions, and not to the capabilities. Possible evil acts are not evil. Evil comes into the picture only with actuality; and it is compounded in an evil act, such as torturing, when that act is freely chosen rather than coerced. The capability for torturing would not be bad in itself, even if one could not resist—say, because of psychological conditioning—realizing that capability. Only the actual torturing would be evil: it would reveal that the torturer lacked the valuable capability of being able to refrain from torturing people. Likewise, on the issue of gun control, Sen would construe the argument against gun control to be not that guns give people an intrinsically bad capability of shooting people, but that this neutral capability is likely to be realized in disastrous acts (because the ability to shoot people will not be inhibited by the valuable capability of not shooting).

I have two problems with this argument on the level of general capabilities (in contrast to specific abilities). First, if we accept Sen's premise that what is evil occurs only on the level of functioning and not the level of capability, then we will have to say the same thing about the good and the valuable. Good would occur only in actual functioning and never in capabilities for functioning. Being able to be healthy, being able to choose, and being able to be concerned about the environment would not be good (or bad) in themselves. What is good would only be the healthy functioning, the actual choosing, the demonstrated concern. But Sen correctly emphasizes that it is good to have the capability or freedom for good health just as it is beneficial to function in a healthy way. It is good both to be able to choose and actually to choose. It is good to be able to be a friend as well as have friendships.

A second problem is that Sen's argument hinges on his incomplete assumption that capabilities are (no more than) possible functionings rather than (also) actual powers of persons. The capability, say for healthy bodily functioning, is in the realm of possibility rather than actuality. A capability for bodily health is just the same as healthy bodily functioning, with the subtraction of one feature, namely, actual existence. (One is reminded of Kant's refutation of the ontological argument by appealing to the difference between 100 conceived thalers and actual thalers in the pocket.) Why call something bad if it is only possible—unless it is tied by probability relations to bad actuality?

Sen's mistake here, I believe, is precisely his assumption that capabilities are possibilities facing the agent rather than, with Nussbaum, powers possessed by the agent. If capabilities are powers, then they are also, in a sense, actualities that we can, do, and should evaluate on a scale from good through trivial to bad. A difference exists between a latent or undeveloped human power and an actual or developed one. Some powers are worth developing and having and some are not. Among the latter, some are bad to have and this defect is not due merely to what their outcomes might or likely will be. We want our children to acquire good character and valuable abilities even though they may never have the chance to display the former or exercise the latter. (We certainly hope that they can and will display their virtue and excellence.) We oppose the nurturing of abilities to torture, rape, and degrade; even though they would never issue in their counterpart actions. For we believe that it is bad to be of bad character and have bad abilities even when the person cannot and even could not express them in action. Nothing may be wrong with the gourmet's capability of enjoying exotic foods unless his exercise of that capability actually cuts into the capability of others to have enough to eat. The ability to shoot, even to shoot other people, is not necessarily bad and may be justified for police and the military. But there is something wrong with having the unexercised capability for child molesting or, more generally, the capacity for injustice. In between (i) possible functionings and (ii) actual functionings, are (iii) actual abilities to function. Even if we accept that (i) can not be bad, we can and must be able to say that some examples of both (ii) and (iii) are bad.¹²

If this second criticism of Sen is sound, we have another reason for modifying Sen's ‘capabilities as possibilities’ in the direction of Nussbaum's view of ‘capabilities as powers’. The latter does more justice to our intuitions that some powers are bad: not just to exercise, but also to have, acquire, or nurture.

It is not enough to carve out the space of functionings and capabilities; for, as we have seen, these actual and possible functionings differ in value. What international and national development should do is to expand capabilities, especially valuable ones, and promote valuable functionings.¹³ What are Sen's and Nussbaum's lists of good functionings and capabilities? What is the basis for the list and the rankings?

Sen clearly recognizes the importance of the task of developing his capabilities approach to include ‘different evaluation exercises’ (Sen, 1987b: 107): ‘It is valuation with which we are ultimately concerned in the functionings approach’ (Sen, 1985a: 32). However, he also recognizes the difficulty of the project: it is hard to put the right questions let alone get the right answers, and in 1985 he confessed that he had ‘no magic solution to offer in dealing with these complex questions’ (ibid.: 48). Moreover, as discussed earlier in Sen's critique of Rawls, Sen emphasizes a social commitment to the freedom that individuals ‘actually enjoy to choose between different ways of living that they can have reason to value’ (Sen, 1990c: 115). Sen evaluates positive (and negative) freedom as more basic than the identification of the more determinate valuable capabilities and functionings because the former provides the framework or format for the latter. Unlike Rawls, Sen is not opposed to a thick conception of the good and a ‘perfectionist’ conception of good government. He believes, however, that the norms of valuable capabilities and functionings presuppose and require the prior defence of the ‘moral space’ of freedoms and capabilities.

These disclaimers notwithstanding, Sen accepts that a development ethic can and should go beyond ‘actual freedom’ and (i) assess the living standard or well‐being of people; that is, evaluate or rank how well people in the same or different countries are in fact living; and (ii) identify a cut‐off point or threshold, beneath which people are in poverty or seriously deprived. These purposes are served by formulating a theory of ‘basic functionings’ and ‘basic capabilities’ or ‘primary capabilities’ (Sen, 1984: 320). A basic capability is defined as ‘being able to do certain basic things’ (Sen, 1980: 218); and such basic doings are precisely supremely valuable functionings.¹⁴

Although Sen so far has offered only scattered suggestions and examples of valuable capabilities and functionings, Nussbaum has emphasized this evaluative task and proposed the outline of a systematic conception. Her aim, often employing Sen's concepts and arguments, is to articulate what she calls an Aristotelian view of ‘good human functioning’ that precedes and is the basis for considering the responsibilities and structures of a just political arrangement. I pointed out earlier that in Rawls's liberal (and ‘political’) theory, the right, with certain qualifications, is prior to the good. Rawls proposes what he takes to be a fair framework—albeit informed by ideals of moral personality and social co‐operation—in which people, within limits, are free to pursue their own conception of the ultimate good. For Nussbaum, the good is prior to the right in that the aim of government goes beyond fairly distributing Rawls's primary goods and Sen's positive freedoms, as important as both these tasks are. The more determinate and guiding aim of just legislators should be that of promoting ‘the capability to live a rich and fully human life’ (Nussbaum, 1990a: 217). Nussbaum, taking the space of capabilities and functioning as settled, invites us to join her in working out a general and systematic account of good human functioning in two levels or stages. Who is the ‘we’, what are her two levels, and what are her tentative results?

The ‘we’ is best conceived if and when it includes people who discover or create—through an international dialogue—a provisional, revisable consensus on what it means to be human and to live well. Participants consult their own experience, the stories and self‐understandings of their respective groups, and the insights of other groups and dialogue partners. International interdependence and boundary crossings of various kinds make it particularly imperative to forge together a global ethic and a conception, as widespread as possible, of being human and human flourishing.¹⁵

The first step in constructing an account of good human functioning is to work out an ‘outline sketch’ (Nussbaum, 1988b: 38–9; 1990b: 70) of being human, a ‘thick vague conception of the human being’ (1990a: 205) or ‘the shape of the human form of life’ (ibid.: 219). The second step, the thick vague conception of good human functioning (ibid.: 205), goes further and provisionally identifies, in a more determinate but still general way, the most important or ‘basic human functional capabilities’ (ibid.: 225), ‘the totality of functionings that constitute the good human life’ (ibid.: 209).

Nussbaum offers her theory as non‐metaphysical in two senses: (i) it is not ‘externalist’ in the sense that ‘it is not a theory that is arrived at in detachment from the actual self‐understandings and evaluations of human beings in society’ (ibid.: 217), and (ii) it is not a ‘theory peculiar to a single metaphysical or religious tradition’ (ibid.: 217). Nussbaum describes her own theory as ‘internalist’, in the sense of drawing on ‘searching participatory dialogue’ in and for human history and experience, and ‘essentialist’, in the sense of offering a normative yet revisable conception of human life's ‘defining features’ (Nussbaum, forthcoming b; see also forthcoming a). It seems clear, however, that Rawls would call Nussbaum's theory a ‘metaphysical’ rather than merely a ‘political’ conception of justice. Nussbaum is explicitly striving for a ‘general and comprehensive’ (in Rawls's senses) conception of the human good—albeit one that aims to be an ‘overlapping consensus’ concerning ‘our recognitions of members of very different traditions as human across religious and metaphysical gulfs’ (ibid.: 217). Rawls uses ‘overlapping political consensus’ in a much narrower sense to refer to agreement about the political domain for an already democratic society. Nussbaum, in contrast, aims to forge an international consensus about good human functioning (ibid.: 205). The common vision is to be general and comprehensive. It is also to be fallible, ‘historically sensitive’, and shareable regardless of one's particular religious or metaphysical commitments.¹⁶ If the project succeeds, we will have ‘the basis for a global ethic and a fully international account of distributive justice’ (forthcoming b).

Let us look in more detail at Nussbaum's argumentation and results at each level. At the first level she tries to identify the ‘constitutive conditions’ or ‘shape of the human form of life’ (ibid.: 219; forthcoming b). What, according to Nussbaum, does or should the cross‐cultural consensus count as human? What follows is her ‘first approximation’ (1990a: 219) (see 1988b: 48–9; 1990a: 219–24; 1991: 17–22; forthcoming b):

Level 1 of the Thick, Vague Conception: The Constitutive Circumstances of the Human Being (or: The Shape of the Human Form of Life)

1.

Mortality (Fact and awareness of death and aversion to it)

2.

The Human body

2.1.

The need/appetite for food/drink

2.2.

The need/desire for shelter

2.3.

Sexual desire

2.4.

The need/desire for mobility

3.

Capacity for Pleasure and Pain

4.

Cognitive Capability: Perceiving, Imagining, Thinking

5.

Early Childhood Development (Helplessness, need, and dependence)

6.

Practical Reason (Each evaluates, chooses, plans, and executes a conception of the good life)

7.

Affiliation and Concern for Other Humans (Each ‘lives to and with others’)

8.

Dependence on and Respect for Other Species and Nature

9.

Humour and Play

10a.

Separateness (Each is ‘one in number’)

10b.

Strong Separateness (Each has her own peculiar context)

The list is not a value‐neutral report from a biologist or externalist metaphysician; it is already evaluative in a most general way. Moreover, ‘this list of capabilities is a kind of ground‐floor, or minimal conception of the good’, (1990a: 224) for a creature that lacked any one (let alone all) of these traits ‘would be too lacking, too impoverished, to be human at all’ (ibid.: 224) let alone have a good human life:

It [the list] recognizes certain aspects of human life that have a special importance. Without them, we would not recognize ourselves or others as the sort of beings we are; and they provide the basis for our recognition of beings unlike ourselves in place, time, and concrete ways of life as members of our very own kind. (Ibid.: 218)

We say that a life without these items is not recognizable as human; and, given that any life we can coherently wish for ourselves or for others will have to be, at least, human, it sets an outline around our aspirations. (Ibid.: 224; see also 1990b: 379 ff; 1991: 15; forthcoming b)

One problem here is that it is unclear how to interpret Nussbaum's language of essentialism.¹⁷ Suppose we interpret it in a strong way. Then each part of each of the above conditions are viewed as necessary or essential (and together sufficient) for something being counted as human being. A creature would be ‘lacking in humanity’ if it were blind. However, this strong interpretation would entail that Stevie Wonder was not human. We certainly do want to say that perceiving in general and sight in particular enhances the quality of life and even that it is part of human flourishing, but that seems to take us to Nussbaum's second level.

We can stay on the first level and give a weaker and more plausible interpretation of Nussbaum's argument (see 1990a: 221). Instead of saying that an irremediably sightless person was not human, we should adopt a more Wittgensteinian ‘family resemblance’ view of definition and say that the more these central properties are irremediably missing, the less confidence we have in calling a creature human. Permanent blindness should not by itself disqualify someone from humanness. Total lack of perception would come closer to a disqualifying condition, but even then we might want to count a non‐perceiving being that could imagine and think as at least partially human. In at least early episodes of Star Trek, Spock, born of human mother and Vulcan father, is not viewed as fully human because of his inability to have emotions. He was, however, initially counted as partially human and, in the later episodes, as increasingly human when he gradually acquired or permitted himself emotional capabilities. Nussbaum should replace her sharp ‘threshold’ of necessary and sufficient conditions with a ‘more or less’ minimum for what counts as human.

Nussbaum's Level 1 list includes a heterogeneous mixture of human traits, ‘both limits against which we press and capabilities through which we aspire’ (forthcoming b). In effect, Nussbaum includes in this first threshold of human capacities elements that are entirely absent from Sen, namely, the context of certain limits, vulnerabilities, and needs that are also part of the human mode of existence. It might be argued that Nussbaum should cleanly separate these limits and so forth from the first level capabilities. Nussbaum, however, emphasizes that our human powers only make sense in intimate connection with our human limitations. As humans we exercise our powers and capacities as we struggle against certain vulnerabilities and seek to meet characteristic sorts of needs. We are beings who must die, and in order to live and/or live well must meet certain bodily, psychological, and social needs. Being human, then, includes not just having and exercising certain capacities but doing so in the context of ‘countering [our] limitations’ and meeting our related needs in ‘a humanly good way’ (1990a: 224). For example, we ‘press against’ our bodily needs for food by preferring ‘recurrent hunger plus eating to a life with neither hunger nor eating’ (ibid.: 224). We struggle against our mortality by weakening our desire for other‐worldly transcendent permanence and affirming that there is intrinsic value precisely in those human functionings in which we struggle against the limits, vulnerabilities, and transitoriness of human life (1986a: 2; 1990a: 224; 1990b: 365–91).

In seeking to deepen a needs‐based development ethic, Sen's gloss of needs by capabilities unfortunately loses something that only a concept of need can supply. Nussbaum improves on Sen's account of human capabilities not only by making them personal traits but also by locating them within the context of other human features: the limits, vulnerabilities, and needs of lives we count as human. Our good human functionings, however we conceive them, occur in peculiarly human circumstances in which we cope—humanly and humanely—with certain features peculiar to the sort of creatures we find ourselves to be.

Level 2 of Nussbaum's proposal for a ‘thick vague conception of the good’ is her list of those human functional capabilities that are supremely valuable. We can call each item a virtue or excellence, for Nussbaum explicitly draws inspiration from Aristotle's conception of human flourishing as virtuous dispositions and activities. This conception of good human capabilities (and functioning) is ‘based on’ the above normative conception of human existence in the sense that the first‐level perspective provides the starting‐point and frame within which the second‐level ethical inquiry identifies more specific—but still general—valuable human functionings and a second threshold of capability with respect to each functioning. Although this second‐level list could be either one of valuable functionings or valuable capabilities for these functionings, Nussbaum chooses the latter. As we have seen, the aim of good government and good development is to promote valuable capabilities, that is, ensure that people get to a second threshold so that they have the positive freedom to choose whether or not to advance further (1990a: 224; see Sen and Drèze, 1989: 42–3; Nussbaum, forthcoming b):

We want to describe two distinct thresholds: a threshold of capability to function beneath which a life will be so impoverished that it will not be human at all; and a somewhat higher threshold, beneath which those characteristic functions are available in such a reduced way that, though we may judge the form of life a human one, we will not think it a good human life. The latter threshold is the one that will eventually concern us, when we turn to public policy. (Nussbaum, 1991: 23; emphasis in text)

We are now ready to summarize Nussbaum's list of ‘Basic Human Functional Capabilities’ (1990a: 225; see Nussbaum, forthcoming b; Nussbaum, 1991: 24) and to map Sen's scattered remarks onto Nussbaum's list. I have labelled the items in Level 2 with names of virtues, both to make explicit the relationships to the parallel items in Level 1 and to make the items easier to refer to and remember. N and S stand for ‘Nussbaum’ and ‘Sen’, respectively. The quoted items come from Nussbaum 1990a: 225 unless otherwise noted.

Level 2 of the Thick, Vague Conception: Basic Human Functional Capabilities

1.

Virtues in Relation to Mortality

1.1.

N and S: ‘Being able to live to the end of a complete human life, so far as is possible’ (see Sen, 1988b: 13)

1.2.

N: Being able to be courageous (see Nussbaum, 1990b: 374, 378)

2.

Bodily Virtues

2.1.

N and S: ‘Being able to have good health’ (see Sen, 1985b: 197)

2.2.

N and S: ‘Being able to be adequately nourished’ (see ibid.: 197)

2.3.

N and S: ‘Being able to have adequate shelter’ (see Sen, 1980: 218)

2.4.

N: ‘Being able to have opportunities for sexual satisfaction’¹⁸

2.5.

N and Being able to move about from place to place' (see Sen, 1980: 218; Sen and Williams, 1982: 20; Sen, 1982c: 200; Sen, 1985b: 199; Sen, 1987a: 64)

3.

Virtue of Pleasure

3.1.

N and S: ‘Being able to avoid unnecessary and non‐useful pain and to have pleasurable experiences’ (see Sen, 1985b: 195–6, 1987a: 64)

4.

Cognitive Virtues

4.1.

N: ‘Being able to use the five senses’

4.2.

N: ‘Being able to imagine’

4.3.

N: ‘Being able to think and reason’

4.4.

N and S: Being ‘acceptably well‐informed’ (Sen, 1985b: 199; see Nussbaum, 1995: 24)

5.

Virtues of Affiliation I (Compassion)

5.1.

N: ‘Being able to have attachments to things and persons outside ourselves’

5.2.

N: ‘Being able to love, grieve, to feel longing and gratitude’

6.

Virtue of Practical Reason (Agency)

6.1.

N: ‘Being able to form a conception of the good’ (see Sen and Williams, 1982: 13)

S: ‘Capability to choose’ (ibid.: 13); ‘ability to form goals, commitments, values’ (Sen, 1985b: 218; Sen, 1987a: 41)

6.2.

N and S: ‘Being able to engage in critical reflection about the planning of one's own life’ (see Sen, 1985b: 218)

7.

Virtues of Affiliation II (Friendship and Justice)

7.1.

N: ‘Being able to live for and to others, to recognize and show concern for other human beings, to engage in various forms of familial and social interaction’

7.1.1.

N: Being capable of friendship (Nussbaum, 1995: 24)

S: Being able to visit and entertain friends (Sen, 1985b: 199)

7.1.2.

S: Being able to participate in the community (ibid.: 199).

7.1.3.

N: Being able to participate politically (Nussbaum, 1990a: 233) and be capable of justice (Nussbaum, 1988a: 161)

8.

Ecological Virtue

8.1.

N: ‘Being able to live with concern for and in relation to animals, plants, and the world of nature’

9.

Leisure Virtues

9.1.

N: ‘Being able to laugh, to play, to enjoy recreational activities’

10.

Virtues of Separateness

10.1.

N: ‘Being able to live one's own life and nobody else's’

10.2.

N: ‘Being able to live in one's very own surroundings and context’

11.

Virtue of Self‐respect

11.1.

S: ‘Capability to have self‐respect’ (Sen and Williams, 1982: 20)

11.2.

S: ‘Capability of appearing in public without shame’ (Sen, 1985b: 199)

12.

Virtue of Human Flourishing

12.1.

N: ‘Capability to live a rich and fully human life, up to the limit permitted by natural possibilities’ (Nussbaum, 1990a: 217)

12.2.

S: ‘Ability to achieve valuable functionings’ (Sen, 1985b: 200)

Four comments are in order with respect to the list. First, the items on Level 2 nicely parallel the items on Level 1 with one notable exception: Level 2's item 5, the Virtues of Affiliation (Compassion), are not as specific as they might be. Given that we want to count as human only those kinds of creatures that develop from infancy to young adulthood, it would be appropriate to specify further, by adding as 5.1.1 or 5.1.2, the virtue of being able to grow and be nurtured by adults as well as the virtue of good parenting. These virtues would fit nicely in an Aristotelian account in which good habits are a matter of training and education.¹⁹

Secondly, in order to simplify and help remember Level 2's list, I propose the labels (‘virtue of . . . ’) for the types of valuable capabilities. Are there any even more general categories by which the virtues might be organized? Neither Sen nor Nussbaum offers any, although we have seen that Nussbaum suggests that practical reason and affiliation (what I have called ‘Virtues of Affiliation II’) are ‘architectonic’ in that they ‘infuse’ and ‘organize’ the other excellences. ‘Infuse’ suggests that the other capabilities can be exercised humanely just in case humans reason practically in and for others. (I would suggest that Virtue 11, Sen's ‘virtue of self‐respect’, also should play this role.) ‘Organize’ suggests that these two capabilities give some structure to the long list. Perhaps Nussbaum intends that all the other virtues be classified as varieties of either practical rationality or affiliation (but not both), but that would be at odds with her notion of ‘infusion’. Worth considering would be an ordering with three general types of virtue (with sub‐types in parenthesis): Physical Virtues (1, 2); Individuality (or Agency) Virtues (3, 4, 6, 10, 11); and Reciprocity (or Social) Virtues (5, 7, 8, 9).²⁰ This trichotomy may be more trouble than it is worth, for it might suggest dichotomies that Nussbaum wants to avoid, e.g., physical/mental or individual/community. On the other hand, the excessive length of the present list makes some additional organization desirable. To facilitate this ordering, it might be better for practical rationality and affiliation to ‘infuse’ but not ‘organize’ the other virtues.

Thirdly, Sen's and Nussbaum's lists differ at a few points. For Sen, the bodily capabilities and functionings (2) are intrinsically good and not, as they are in some dualistic theories of the good life, merely instrumental means to other (higher) goods. In interpreting Aristotle, Nussbaum distinguishes between bodily functionings that are chosen and intentional, for instance, ‘chosen self‐nutritive and reproductive activities that form part of a reason‐guided life’ and those that are non‐intentional, such as digestion and other ‘functioning of the bodily system in sleep’ (forthcoming a). She may want to say that intentional bodily actions that lead to being well‐nourished and healthy are intrinsically good, but that being healthy or having good digestion are not functionings (because not intentional) and are valuable only because of what they enable us to do. Another option open to her would be to adopt Sen's view that bodily states and processes, whether intentional or not, both as intrinsically and instrumentally good but as less valuable than other inherently good capabilities/functionings.

Furthermore, Nussbaum has included items 5 and 8–10, for which Sen has no counterparts. These items are welcome features. Item 8, which I have called ‘ecological virtue’, is an especially important recent addition to Nussbaum's outlook. In a period when many are exploring ways of effecting a convergence between environmental ethics and development ethics, it is important that an essentially anthropocentric ethic ‘make room’ for respect for other species and for ecological systems. Worth considering is whether Nussbaum's ‘ecological virtue’ is strong enough. Perhaps it should be formulated to read: ‘Being able to live with concern for and in relation to animals, plants, and nature as intrinsically valuable.’ Item 9 injects some appealing playfulness in a list otherwise marked by the ‘spirit of seriousness’. What explains the presence of these items on Nussbaum's list, their absence on Sen's list, and, more generally, the more concrete texture often displayed in Nussbaum's descriptions? One hypothesis is that the differences are due to Nussbaum's greater attention, in her Level 1, to the limits, vulnerabilities, and needs of human existence. Further, it may be that Nussbaum's richer conception of human beings derives from making use ‘of the story‐telling imagination far more than the scientific intellect’ (Nussbaum, 1990a: 217; see 1986a: 69; 1990b). On the other hand, Sen helpfully includes the good of self‐respect, a virtue that enables him to find common ground with Rawls and to establish links with the Kantian ethical tradition, in which moral agents have the obligation to respect all persons, including themselves, as ends‐in‐themselves.

Fourthly, both thinkers make it clear that these goods are not only many but each is ‘distinct in quality’ and of ‘central importance’ (Nussbaum, forthcoming b). Valuable capabilities are, then, ‘incommensurable’ in two senses. First, ‘we cannot satisfy the need for one of them by giving a larger amount of another one’ (ibid.). Hence there are strict limits on the ‘trade‐offs’ common in quantitative cost‐benefit analysis. Secondly, the plural goods are incommensurable in the sense that they are irreducible to some common and ‘deeper’ measure such as utility. Nussbaum emphasizes that a science of rules cannot replace a practical wisdom that improvises in unique contexts, exercises the ‘ability to discern, acutely and responsively, the salient features of one's particular situation’ (1990b: 37), and affirms the ethical value of the emotions and imagination (ibid.).

One objection to Sen and Nussbaum here might be the following. On the surface, this concept of the good human life is attractive, but is not such a pluralistic vision of the good life going to run into insoluble problems when two or more of the capabilities cannot be simultaneously chosen as actual functionings? Sen and Nussbaum agree on an instructive and convincing response. First, neither of them dodge the ‘conflict of principles’ problem. Nussbaum stresses that two or more of the items, such as the capabilities for mobility and nutrition, often support each other (1990a: 225); she recognizes, however, that the ‘components may in principle conflict with one another’ (ibid.: 225). Sen, whose early fame rested on his important work in social choice theory, makes the point that it is better to be ‘vaguely right’ than ‘precisely wrong’ (1987b: 34). It is better to be correct in identifying the diversity of good functionings and be beset with the problem of ordering them than in using one homogeneous quality like utility that, at best, does justice to only one intrinsic good and, at worst, is wildly inaccurate with respect to human well‐being and other goods.

Moreover, Sen argues that a pluralist approach can yield various sorts of ‘partial orderings’—short of a complete ordering that would lay down ahead of time adjudication rules for every disagreement among lists of valuable functionings, priority rules for conflicts within the best list, and application of the one list to two functioning bundles (Sen and Williams, 1982: 17; Sen, 1984: 288; Sen, 1985a: 16, 54; Sen, 1987b: 29, 108). A thick, vague conception of the good, although general, is not indeterminate; by critical reflection it rules out many lists, ‘weightings’, and functioning vectors as pretty obviously inadequate. Here the appeal is to persistent, widely shared, and consistent intuitions.

Nussbaum especially addresses the ordering problem in relation to clashes between two valuable capabilities or two valuable functionings. Two component functionings may not be (equally) co‐possible: ‘Concern for other species may or may not fit well with our efforts to feed ourselves, to be mobile, to be healthy’ (Nussbaum, 1990a: 226). Nussbaum attacks the ‘conflict of goods’ problem in two ways. First, it may turn out that by acting and, thereby, changing the world, we can have our cake and eat it too (see ibid.: 212). What at first seem to be ‘tragic choices’ may in fact be caused by a social order that can be changed (at a cost that does not outweigh the benefit of having both initially conflicting functionings). Here Nussbaum is assuming, contrary to Rawls and the liberal tradition, that it is possible to forge a national and international agreement on good human functionings. She is also assuming, again in contrast to Rawls and other liberals, that it is at least sometimes possible to resolve conflicts between two (or more) components of a generally agreed‐upon conception of the good life.²¹ Once we discern such a tension, we can identify its source and determine whether or not it can be eliminated or at least mitigated. For example, more efficient use of present energy may make it unnecessary to build a hydroelectric plant that would destroy the valley home of an indigenous tribe and an endangered mammal species. We may find that the conflict between preserving rainforests and ensuring that everyone is well nourished is not a ‘tragic choice’ but can be eliminated by a structural change in a nation's land tenure system. We judge both fulfilling work and loving child care to be intrinsically valuable human functionings, but we find that current and remediable institutional arrangements force into unnecessary tragic conflicts those women who pursue both ends.

Secondly, Nussbaum urges that we respond to apparent conflicts of goods by eschewing a priori priority rules. Giving a particularist spin to her conception of reflective equilibrium, she urges us to try to resolve conflicts among incommensurable goods by ‘penetrating into the particularity’ of mutable, complex, unique situations (Nussbaum, 1990b: 66–75). In doing so sometimes it will be quite clear which goods, in the particular context, should take precedent. We can construct provisional, working rules from repeated concrete deliberations, and these rules can guide us when our intuitions are confused, hazy, or inconsistent. Antecedent priority rules even help correct bias in concrete deliberations. There is, however, no rule for when priority rules should win out. The final appeal, as in doctoring or navigation, is to discern or discriminate in specific situations (ibid.: 73, 97).

Nussbaum concedes and even insists that it may turn out that ‘one strand in our common humanness may not be harmoniously related to another, given the circumstances of life’ (1990a: 226). Then we may have to live, in that situation at least, with a tragic value‐conflict. Neither Sen nor Nussbaum offer an algorithm for resolving conflicts within their conception of the plural and diverse good. Conflicts within our conception of the good, like disagreements among conceptions of the good, may be one of the limits on our collective humanness and a creative expression of our cultural and individual particularity. But before we resign ourselves to moral impasses and tragic choices, we should consider changing the world, so as to eliminate conflict, and penetrating more deeply into the world, so as to discern the most pressing good in that context.

There is one ordering or valuation problem that so far neither Sen nor Nussbaum have addressed, namely, the relative weight to be given to (valuable) capabilities and (valuable) functionings. Nussbaum, we have seen, makes two points about the relation of a valuable capability to function and the actual functioning. On the one hand, each is valuable in itself and as related to the other: a valuable functioning is valuable, in part, because it realizes a valuable capability; and a capability is valuable, in part, because it is realized in a functioning that is valuable (see Nussbaum, 1988a: 169; 1988b: 15–16). Moreover, from the perspective of the lawgiver, capabilities have priority because good government has the task not of making people function in a certain way but of getting people to a second (capability) threshold where they are able to choose whether or not to function in those ways. Or to put the same point in the language of (positive) freedom, it is good both to be free to choose among options and to choose the best option.

This nice balancing of our ‘perfectionist’ and ‘liberal’ intuitions, however, ignores the issue of those cases where we have to decide on the relative importance of good capability and the correlative good functioning. Is it better to award the prize (praise or a scholarship) to X who has greater capability for good functioning but is unlikely to choose to so function (and most likely will function at a lower level) or to Y who has less capability but likely will realize the capability that she does have in a good functioning (that surpasses what X is likely to achieve). When we cannot do both, should we reward X's greater capability relative to Y or Y's greater past achievement relative to X? Is there a general priority rule, even a rule of thumb, for this case? Do we penetrate into particularly and consider more specific types of cases or individual cases? Perhaps there is a general difference between (i) a forward‐looking selection of a ‘promising prospect’ and (ii) a backward‐looking award for past achievement? Frequently, however, awards such as scholarships have both forward‐ and backward‐looking aspects. When we have to choose between a relatively superior capability (of one person) and a relatively superior functioning (of another person), which should we choose? Are there any reasonable principles or should we just improvise? A liberal will go with capability. A perfectionist will go with achievement. What of the liberal Aristotelian? Is there a difference that makes a difference between Sen and Nussbaum? My hunch is that Sen, more of a liberal, will rank positive freedom and capability more highly than actual achievement and that Nussbaum, more of an Aristotelian, will put achievement first.

A final issue: at several points Sen says that he is providing the basic needs approach to development (BNA) with needed ‘foundations’, that the BNA and other ‘quality of life’ approaches ‘can be more explicitly developed, conceptually defended, and empirically applied’ (Sen, 1988b: 16) by his capabilities approach. Still to be investigated, however, is whether a concept of needs has a role to play that cannot be accomplished by either capabilities and functionings. Has Sen avoided ‘commodity fetishism’ only to fall into ‘capabilities fetishism’?²²

Nussbaum identifies two non‐reducible roles that need play in the capability ethic. First, humans need to develop their nascent valuable capacities into mature ones. Their ‘undeveloped’, implicit, or embryonic capabilities are ‘needs for functioning’ (Nussbaum, 1995: 31): ‘the very powers of this being exist as needs for fulfillment’ (1990a: 243). A need is satisfied when these implicit or potential capabilities become explicit or actual capabilities:

On this account, B‐capabilities [Nussbaum's term for undeveloped or potential internal capabilities] are needs for functioning: they give rise to a claim because they are there and in a state of incomplete realization. They are conditions that reach towards, demand fulfillment in, a certain mode of activity. If that activity never arrives, they are cut off, fruitless, incomplete. As Aristotle insists, their very being makes reference to functioning; so without the possibility of functioning, they are only in a shadowy way even themselves. (Nussbaum, 1988a: 169; see 1995: 31; emphasis in text)

As she makes clear, Nussbaum's appeal to needs here is not to subjective desires or preferences (see 1988a: 169). She means more than that we should value and promote the development of our own and others' potential capabilities and then realize them in functioning. Talk of our human need for actual capabilities and actual functionings is a way of saying that actuality is prior to possibility in the ethical sense that (i) actual capabilities are more valuable than merely latent ones and ‘have a claim to be developed’ (1991: 31) and (ii) actual (internal) valuable capabilities essentially refer forward to functioning and, hence, make a claim to be realized in functioning. This is not to say that valuable capabilities are not also valuable in themselves, but ‘if functioning never arrives on the scene they are hardly even what they are’ (ibid.: 31). In any case Bernard Williams's following question is to be answered affirmatively: ‘Do we come out of the terminology of capabilities again, when we turn to their natural basis?’ (Sen, 1987b: 101). It is not that we have found some value‐neutral fact about our being that metaphysically entails an ethical duty. Rather, our cross‐cultural human self‐interpretations are such that we deem ourselves obliged to promote the acquisition and realization of certain capabilities (in ourselves and others). And we view it as especially tragic when a young person, full of promise, dies before having the chance to develop and realize her excellent powers (Nussbaum, 1995: 31).

Nussbaum gives the concept needs a second role; for, as we have seen, she argues that human excellence is acquired and displayed precisely in relation to certain human lacks and limits. A good athlete presses against and makes recede her human limitations. But to extinguish them altogether—for instance, by gaining infinite speed by divine steroids—would eliminate both competition and the competitor. The same is true of virtue. Without various vulnerabilities like death, we would not have the ability to be courageous.

Sen has not really taken up these question, but it seems clear that his intent is to push freedoms and capabilities as opportunities as far as he can without resorting to other concepts. Sen's theory of actual freedom would be more comprehensive and humanly nuanced if he followed Nussbaum and added internal powers to external opportunities and viewed humans not only as capable but as in need of nurture in a context of neediness. The decisions of how to grow and function—of how to develop, do and be—are to be made not just in relation to resources and opportunities but in relation to certain human limits and vulnerabilities that we must struggle against in humanly appropriate ways. If Nussbaum is correct, as important as the concept of capabilities proves to be, there still would be a continuing role for the language of needs.

Sen's concept of capability and his critique of Rawls led us to touch on Sen's conception of the nature and value of positive freedom. For ‘the capability to function is the thing that comes closest to the notion of positive freedom, and if freedom is valued then capability itself can serve as an object of value and moral importance’ (Sen, 1984: 316). For Sen, positive freedom (i) is defined as ‘the “capability” to achieve various alternative combinations of functionings, that is, doings and beings’ (Sen, 1990c: 116) and (ii) can add intrinsic value to the actual options and functionings chosen. The good human life is, among other things, a life of freedom (Sen, 1985a: 70). Valuable human functionings include the act of choosing from among valuable functionings. In some of his recent writings Sen tends to use the concept of freedom even more than the concept of capability. The reason, I think is twofold. Against right wing libertarians such as Hayek, Friedman, and Nozick, Sen is trying to (re)claim the concept of freedom for the democratic left—whether that of social democracy or democratic (self‐managing) socialism. In distinction from Nussbaum, Sen gives an account of positive freedom that is ‘thicker’ than Rawls's primary goods but ‘thinner’ than Nussbaum's (and his own) concept of the good life.

Sen offers two distinctions to grasp the full complexity of the concept of freedom: (i) negative freedom/positive freedom, and (ii) well‐being freedom/agency freedom. Let us briefly consider each. Sen roughly accepts Isaiah Berlin's classic distinction between negative and positive freedom but rejects Berlin's critique of the latter. A person is negatively free when one is not being interfered with by others—whether persons, governments, or institutions. Such interference, whether restraint or coercion, is bad because the interferer prevents the agent from doing what he chooses or choosing what he does. But Sen correctly sees that someone can be free from such external interference and still be radically unfree due to the absence of options in general and valuable options in particular. No one may be interfering with the starving person and yet she is not free not to starve, for bad fate is her only option. Nussbaum puts it well: ‘Some policies of non‐interference actually extinguish human freedom to choose what is valuable’ (Nussbaum, 1988a: 213). The concept of positive freedom is important because it marks out how a person is really able to act, live, function or achieve. Positive freedom is ‘what a person is actually able to do or be’ (Sen, 1989b: 770).

To be positively free is to be able to live as one chooses, to have the effective ‘power to achieve chosen results’ (Sen, 1985b: 208). It is this dimension of positive freedom that R. G. Peffer emphasizes in his recent and important interpretation of Marx. ‘By “positive freedom”,’ Peffer says, ‘I mean freedom in the sense of being able to determine or control one's life,’ that is, ‘have a significant impact or effect upon the direction of one's own life and the circumstances under which one must live.’²³ Sen, however, goes further. He correctly sees that positive freedom also includes the real availability of an array of options and that freedom is increased to the extent that the number and goodness of these options are increased: ‘the extent of freedom must not be judged only by the number of alternatives; it depends also on the goodness of the alternatives’ (Sen, 1987b: 36, n. 17). Though both persons A and B are starving, A is more free than B, if A (the faster) has the real option of being adequately nourished. A's freedom is enhanced not only because A has more options than B but because only A is capable of additional functioning that is valuable. Another step that Sen could and should but does not take is to see that positive freedom is also enhanced insofar as there is an increase in either the diversity or probability of options.²⁴

Sen, consistent with his dual (well‐being/agency) perspective on human beings, also distinguishes between well‐being freedom and agency freedom. If I have the former I have the real opportunity to choose and achieve well‐being; if I have the latter, I additionally have the opportunity to choose against my well‐being. A good society would provide the material and institutional conditions for both types of freedom. However, Sen's discussion of agency freedom as a capability illustrates the weakness we identified earlier. His view of agency (or autonomy) as capability speaks only of opportunities or possible functionings and fails to follow Nussbaum's view of capability as (also) a trait or power of the person. For Nussbaum, the capability of a person to choose depends not simply on external features but also on whether that person has a developed power of choice (Nussbaum, 1988a: 160–72). Neither stones nor human infants can choose right now. The baby, however, has a latent or undeveloped power and, if all goes well, eventually will acquire the actual or developed power as an adult. And, with ageing the power may decline and finally be extinguished altogether.²⁵

One purpose of good development is to ensure that this human ability for choice is acquired by the young, maintained by the mature, and restored (when possible) to those who lose it. Good governments, especially through education, facilitate the formation of good capabilities, remove impediments to their exercise, and provide means for their use. If the present argument is correct, Sen should deepen his notion of freedom so that it faces in two directions: toward both the power of choice and the options open to choice. One virtue of doing so would be to justify the educational dimension of good development (see Nussbaum, 1990a: 233–4; 1988a: 213). We shall see that another advantage would be that ‘capability rights’ could be related to features of the kinds of beings that humans find themselves to be and value.

Given this conception of the complex nature of positive freedom, how does Sen understand its value? Here the crucial distinction is between freedom's intrinsic and its instrumental value. Negative as well as positive freedom have both sorts of value. As we have seen, capabilities always add some value to actual functionings:

Freedom is valuable in itself, and not only because of what it permits us to achieve or do. The good life may be seen to be a life of freedom, and in that context freedom is not just a way of achieving a good life, it is constitutive of the good life itself. (Sen, 1989b: 770; emphasis in text)

It is intrinsically good to choose a way of life and not to fall or be forced into it (Sen, 1985a: 70). Sen's argument, for this claim, is an appeal to our intuitions about the starving/fasting examples discussed above (Sen, 1987a: 60). To say freedom (or autonomy) is intrinsically good is not to say, with Kantians, that freedom is the only good; there are many intrinsically valuable functionings. Due to individual variations, Sen argues that Rawls's primary goods are at best means to positive freedom and that equal positive freedom requires different amount of goods for different sorts of people, if they are to be free to pursue a full range of objectives. More determinately, he can join Nussbaum in saying that it is precisely our actual freedom that enables us to choose, weigh, and pursue valuable goals. Choice is a sort of super or ‘architectonic’ capability in that its exercise enhances the value, and thereby ‘humanizes’ other valuable functionings such as eating, loving, and communal participation (Nussbaum, 1990a: 226–8).

We see here a key element in Sen's and Nussbaum's strategy in meeting Rawls's objection to ‘perfectionism’. Recall that Rawls, in his thin theory of the good, objects to the use of state power to impose one (general, comprehensive) conception of the good. Within ‘wide limits, “justice as fairness” does not prejudge the choice of the sort of persons that men want to be’.²⁶ In going beyond primary goods to actual freedom Sen responds to the liberal challenge. Nussbaum fills in this space by building freedom as a personal power right into her concept of those important capacities that a good political arrangement is to promote. Moreover, although a good polity certainly promotes a conception of the good life, the ‘vagueness’ (generality) of this ‘outline sketch’ permits ample room for each society (‘local specification’) and each individual in a society (‘plural specification’) to exercise the capability for freedom by particularizing, in terms of its traditions and tastes, the thick conception.²⁷ Finally, individuals exercise their (capability) for autonomy when they decide the sphere(s) of life in which they will seek to realize their humanity, for example, family, employment, recreation, religion, or politics. Although the concept of the good life is relevant for all these domains, citizens are free to choose which one(s) are of special importance to them. In all these spheres, but especially in politics, their capability for choice can be realized by forming the group's goals and plans.

Sen correctly insists that the intrinsic value of freedom does not imply that it also cannot be instrumentally valuable in achieving other good functionings. As Aristotle and Dewey note, but many forget, something can be both a means and an end, an instrumental good and a good in itself. For example, some of Sen's important empirical work traces the causal link between India's press freedom and its lack of famines since Independence. Press freedom is both good in itself and promotes the good of avoiding starvation.

From the perspective of his capability approach, Sen has begun to sketch out a suggestive theory of moral rights as ‘capability rights’ (Sen, 1982c: 222). These rights have an important role to play but are defined in relation to ‘basic’ functionings and capabilities or freedoms. Such rights are viewed as consequentialist goals rather than Nozickian deontological side constraints on goal‐seeking action. Finally, Sen argues for positive or welfare rights in addition to negative rights or rights not to be interfered with. Let us take each point in turn.

In commenting on one of Sen's essay (Sen, 1987b), Bernard Williams makes an important observation and asks a significant question:

We actually believe that people have a basic right to breathe clean air without having to go somewhere else to do it, but we do not believe that they have a similar right to go to expensive winter holiday resorts. I am not very happy myself with taking rights as the starting point. The notion of a basic human right seems to me obscure enough, and I would rather come at it from the perspective of basic human capabilities. I would prefer capabilities to do the work, and if we are going to have a language or rhetoric of rights, to have it delivered from them, rather than the other way around. But I think that there remains an unsolved problem: how we should see the relations between these concepts. (Sen, 1987b: 100)

In response, Sen defines a basic right as a right to a minimum amount of basic capabilities or freedoms (see Sen 1982c: 199–200; 1984: 297):

Some of the relevant freedoms can also yield straightforward notions of rights. For example, minimal demands of well‐being (in the form of basic functionings, e.g., not to be hungry) and of well‐being freedom (in the form of minimal capabilities, e.g., having the means of avoiding hunger), can well be seen as rights that command attention and call for support. (Sen, 1985b: 217)

Rights, then, are defined as basic not because they are indispensable to the fulfilment of any other right but because they are a way of formulating the urgency of minimal levels of eminently valuable human (actual and possible) functionings.²⁸ It is intrinsically good for humans to be and do in certain ways and to have the freedom to so function, to be above the ‘threshold’ that enables them to so function. To justify something as a fundamental right is to identify a human functioning as basic, that is, as intrinsically and supremely good. Rights are grounded in the good in the sense that they are justified with reference to valuable human functionings. For instance, recall that Nussbaum proposes ‘separateness’ and ‘strong separateness’, our being able to live our own lives in our own surroundings, as two of our basic human capabilities. Then she goes on to identify a governmentally protected private sphere as indispensable for the functioning of this capability:

Separateness and strong separateness have been read here to require the protection, around each citizen, of a sphere of privacy and non‐interference within which what goes on will not be the business of political planning at all, though politics will protect its boundaries. (Nussbaum, 1990a: 239)

What this sphere consists of would be a matter of political argument that would result in a ‘local specification’ of the basic norm. Such specification, in contrast to the ‘plural specification’ referred to above, is done collectively by citizens in a particular community when, through ‘participatory dialogue’ they particularize—‘with rich sensitivity to the concrete context, to the characters of the agents and their social situation’—the general conception of the good life. It is clear, however, that Nussbaum herself approves of the Athenian consensus that included ‘almost all speech, above all political speech, most of family and sexual life, and most, though not all, one's dealing with personal friends’ (ibid.: 239; see forthcoming b). She also approves of the protection of religious freedom enshrined in the American separation of church and state. One has the right to ‘use’ one's separateness to choose (or not) to practice (with others) a religion.²⁹ Rights come on the scene, so to speak, as we further define and institutionally protect our human good of (strong) separateness: ‘a scheme of basic rights . . . will be justified with reference to the role they play in protecting a way of life that citizens have agreed to be good for them as human beings’ (Nussbaum, 1990a: 239).

Since some valuable functionings are more valuable than others, some rights are more fundamental than others. For instance, ‘the ability to retain bodily safety is quite a different type of right from the ability to keep one's financial accounts private’ (Sen, 1982c: 203). And Nussbaum remarks that in Aristotelian social democracy there is no absolute right to property: land, money, and possessions are not intrinsically valuable but are good only to the extent that they promote the good functioning of citizens and, especially, their moving across the threshold. Basic rights would be formulated in relation to minimally acceptable levels of the eminently valuable functionings. Rights formulate justified demands that people be above a certain threshold of valuable functioning and freedom.³⁰

The demand for this ‘moral minimum’³¹ is relevant to the tasks of good governments and the duties of other people, for example, the duty to eliminate or reduce poverty as an absolute level of deprivation (Sen, 1987b: 109; 1984: Essay 14). However, for Sen, such a relation between duty‐bearer and rights‐bearer is based finally on the relation of the right to certain valuable capabilities (1982c: 200). Hence, Sen calls his view a ‘capability rights system’ rather than a ‘two‐person (or two‐agent)’ rights system (ibid.: 200).³²

A second point about Sen's capability‐based theory of rights is that he offers it as one type of what he calls a ‘goal rights system’ (ibid.: 199). Such an ethics or ‘moral system’ is defined as one ‘in which fulfillment and non‐realization of rights are included among the goals, incorporated in the evaluation of states of affairs, and then applied to the choice of actions through consequential links’ (ibid.: 199). To decide on the moral rightness or wrongness of human actions (commissions or omissions), we must interalia consider the states of affairs that result from the action. Rights fulfilments and violations are then to be counted as ingredients in the states of affairs. Respecting of rights will be part of the value of a state of affairs, and violating of rights will be part of the disvalue: ‘a tortured body, an unfed belly, a bullied person, or unequal pay for equal work, is as much a part of the state [of] affairs as the utility and disutility occurring in that state’ (Sen, 1979: 488). Goal rights systems, although differing among themselves according to their lists of rights, would be ‘pure’ if they proposed no goals other than rights. They would be impure if they also included other goals, such as utility.

Sen situates his theory of rights as goals within his larger ethical project. This perspective, as worked out thus far, remains open either to an ‘impure’ and broad consequentialism, on the one hand, or a broad deontology with ‘consequence sensitivity’, on the other. Goal rights systems would be forms of ethical consequentialism, broadly conceived, if the rights and the other goals were all that mattered morally: ‘for the consequentialist, even a broad one, the rightness of actions has no real sense other than what reflects the goodness of states of affairs’ (Sen, 1983: 130). A goal rights system, however, also could be part of a ‘broad deontology’ if it were what Sen calls ‘consequence sensitive’ (Sen, 1987a: 76). In this sort of deontology, the consequences of rights fulfilments or violations would matter, but things other than good consequences also would be morally relevant.

However Sen finally decides to stand on this issue of broad consequentialism versus broad deontology, he argues that a broad consequentialism, a ‘goal rights system with consequence‐based reasoning’, is superior to any narrow deontology such as Nozick's. On Nozick's view, people have (negative) rights, such as a right not to have one's person or property interfered with (without one's consent). These are consequence‐independent ‘side constraints’ rather than goals. That is, Nozick claims, that rights are not part of end states to be achieved but rather ‘constrain your goal‐directed behavior’:

The side‐constraint view forbids you to violate these moral constraints in the pursuit of your goals; whereas the view whose objective is to minimize the violation of these rights allows you to violate the rights (the constraints) in order to lessen their total violation in the society.³³

Nozick means that rights as side constraints (almost) never can be violated justifiably or overridden even if the consequence would avert an even greater violation of rights (or promote an overwhelming fulfilment of rights).³⁴

In arguing for the superiority of his own view, Sen puts his finger exactly on the place where Nozick's view seems morally reprehensible. Sen has three compelling arguments. First, perfect fulfilment of Nozickian (negative) rights is compatible with ‘terrible hardships and miseries’ (Sen, 1988c: 275) in the lives of those who lack economic power. For example, as Sen's work on famine shows, ‘even terrible famines are entirely consistent with a fully operative and fully complied entitlement system of negative rights of the kind outlined and defended by Robert Nozick’ (ibid.: 275).

Secondly, Sen presents some actual and hypothetical cases in which he effectively appeals to ‘resilient moral intuitions’ (Sen, 1982c: 198) that an agent is justified in violating one Nozickian right as a positive means of respecting some other negative right, more pressing in the context. For example, in his ‘Ali case’ Sen persuasively argues that Donna is justified in breaking into Charles' room and, against his will, inspecting Charles' papers if that is the only way for Donna to prevent the ‘bashing up’ of her friend Ali. Our social ‘multilateral interdependences’ (ibid.: 190) are such that third parties should sometimes violate rights to promote other rights. Negative freedoms and negative rights sometimes require not only omissions but also commissions.

Thirdly, and more generally, Sen convincingly argues that by affirming rights as goals, he can explain a third party's guaranteeing a weightier right by means of violating a less important one:

It is not adequate for me to resist molesting you; it is necessary that I value the things I can do to stop others from molesting you. I would fail to value negative freedom if I were to refuse to consider what I could do in defence of negative freedom. (Sen, 1984: 314; see also 1982c: 222; 1985b: 217; 1987a: 88–9; 1988c: 274–5; emphasis in text)

Finally, Sen's capabilities perspective on rights provides a way of defending what have come to be called positive or welfare rights. The stress on basic capabilities, such as being able to be minimally well‐nourished, enables us to ‘blur the distinction between rights that relate to so‐called positive freedom and those related to negative freedom and non‐coercion’. We can say two things about one capability. Reflecting traditional negative rights, we can say people have a right not to have the capability eliminated or impeded. For example, a right not to be hungry reflects the crucial importance that there be no obstacles to having or realizing one's capability to be well‐nourished (see Sen, 1982e). Reflecting traditional positive rights, we can say people have the right to be provisioned in such a way that they can choose to be adequately nourished. And if we follow Nussbaum in supplementing Sen, we additionally can say that government and other development institutions have a duty to develop people's latent or embryonic capability so they are able and ready to exercise it.

In this connection there remain two areas for future investigation. We saw above that Nussbaum views the right to property as justified instrumentally only to the extent that it promotes valuable human capabilities and functionings. Lacking so far in Nussbaum's account is her consideration of the full range of basic human rights that could and should be related to her ample list of valuable functional capabilities. Whereas Sen makes it clear that basic rights should include welfare or subsistence rights, Nussbaum so far has restricted her attention to the rights to privacy and political speech. I see no reason, however, that Nussbaum could not join Sen and affirm that people have a basic right to the minimal level of all the most valuable capabilities. More positively, her most recent work suggests that undeveloped or ‘basic capabilities . . . give rise to correlated political duties’ and that ‘this idea . . . is the underlying basis, in the Western philosophical tradition, for many notions of human rights’ (Nussbaum, 1995: 31). On this positive basis and her full list of valuable functionings, Nussbaum can and should go on to work out a theory of positive rights.

Secondly, both Sen and Nussbaum should consider in greater depth how duties to respect the various rights are to be distributed. How, for example, should obligations to promote the welfare or personal security rights of persons in a poor country be apportioned among private individuals, the nation's government, rich nations and (foreign) individuals, and international organizations?³⁵ Moreover, how much is required (duty) and how much is praiseworthy (supererogatory) but not required.

Finally, Sen's approach enables us to justify overriding some less important or non‐basic negative right, like the right to have my caviar and eat it too, insofar as it will allow us to reduce the violation or increase the fulfilment of some more important or basic right, such as the right to be able to meet minimal nutritional requirements.

This multiplicity of rights is a part of Sen's more general ethical pluralism. In addition to the points made earlier about ‘partial ordering’ and weighing of conflicting elements, it must also be noted that Sen urges us to conceive capability rights not only as intrinsically good but also as instrumentally good (Sen, 1988d). Respecting rights can be justified instrumentally when such an act is causally linked to other goods. For instance, respecting the capability rights of small farmers and day‐labourers is not only good in itself but in many circumstances will causally promote the good goal of environmental respect. At this point Sen's view seems superior to Nussbaum's. For Nussbaum conceives of rights as purely instrumental (and negative): ‘they will be justified with reference to the role they play in protecting a way of life that citizens have agreed to be good for them as human beings’ (Nussbaum, 1990a: 239). Some rights theorists, such as Nozick and Shue, view rights as ‘neither means nor ends’ but as ‘constraints on both means and ends’.³⁶ Some view rights as means but not ends (or vice versa). Sen suggestively conceives rights as both means and ends.

A normative theory of distributive justice is concerned with such questions as what should be distributed, who or what should be the agent and recipient of the distributing, and what are the circumstances of and bases for just distributions. Such a theory is the least‐elaborated aspect of Sen's and Nussbaum's capability perspective. This judgement is true with respect to distributional issues both within and between nation‐states. So far Sen's major contribution has been to sort out some of the issues (Sen, 1984: Essay 12), criticize some leading theories (1980), argue for a conception of what is to be distributed (ibid.; 1990c), and consider some distributive or ‘combining’ principles (ibid.: 112).

Nussbaum recently has gone further with respect to ‘domestic’ justice by sketching out what she calls an Aristotelian conception of political distribution and social democracy (1988a; 1990a; 1990b). Although what they have done so far is suggestive and promising, they correctly recognize that there is much to accomplish before we have a comprehensive capability‐based theory of national justice, relevant to the internal structure of all nations, and international justice, relevant to the relations between nations and to global institutions.

It should come as no surprise how Sen answers the question of the ‘what’ of distribution and, what he calls, the ‘territory of justice’ (1990c: 115). Individuals' claims to just treatment are to be assessed not (merely) in terms of utility, Rawlsian social primary goods, or Nozickian negative rights but in terms of their capabilities or ‘the freedoms they actually enjoy to choose between different ways of living that they can have reason to value’ (ibid.: 115). Justice concerns how people are free to choose and able to function in valuable ways. Hence, what is to be distributed are, ultimately, freedoms and functionings. Supposing that an adequate theory of justice will be in some sense egalitarian, Sen as early as 1980 asked, ‘equality of what?’ His answer was ‘equality of “basic capabilities”: a person being able to do certain basic things’ (1980: 218). Nussbaum agrees with Aristotle that a government's allocation programmes and even its entire structure ‘will be chosen with a view to good human functioning’ (Nussbaum, 1990a: 230):

The aim of political planning is the distribution to the city's individual people of the conditions in which a good human life can be chosen and lived. This distributive task aims at producing capabilities. That is, it aims not simply at the allotment of commodities, but at making people able to function in certain human ways. A necessary basis for being a recipient of this distribution is that one should already possess some less developed capability to perform the functioning in question. The task of the city is, then, to effect the transition from one level of capability to another. This means that the task of the city cannot be understood apart from a rather substantial account of the human good and what it is to function humanly. (Nussbaum, 1988a: 145–6)

To elucidate Sen's and Nussbaum's position, consider a citizen who is crippled. If a government distributed on the basis of marginal utility, the cripple might be passed over altogether because restorative operations or special walkways probably would not be the most socially efficient use of public funds. Regardless of how utility is conceived, the combining principle of ‘sum‐ranking’ instructs us to assess the consequences simply by adding up the individual utilities. Although each unit of utility counts, what is irrelevant is the personal, separate, or individual locus of the utility: ‘Persons do not count as individuals in this any more than individual petrol tanks do in the analysis of the national consumption of petroleum’ (Sen and Williams: 4; see Nussbaum, forthcoming b).

The cripple might fare no better if her total utility (e.g., happiness) were somehow given priority. For, as we saw earlier, the cripple might have ‘a jolly disposition’, ‘a low aspiration level’ or be content with a disability she believed was due to God's will. Deficient functioning and limited freedom would be ‘muffled’ by contentment. The cripple's total utility might be best served by ignoring and even increasing her disability.

Moreover, Rawls's social primary goods will not do the distributive job needed; because Rawls, even if he included the good of health care, has no concern with human variability and the conversion of goods into freedoms or functionings. To enable the cripple to be able to choose to enjoy the non‐cripple's ‘ability to get around’ is going to take more or different commodities—whether wheelchairs, access ramps, or income supplements. Sen, then, helpfully shifts the discussions of justice ‘from goods to what goods do to human beings’ (Sen, 1980: 219), from income and wealth to freedom to function in desirable ways. Such a focus justifies Nussbaum calling such a theory ‘deep’, since it concerns the ways humans are living.

In what sense is this approach to justice egalitarian? What does it mean to advocate ‘basic equality capability’? (Sen, 1980: 220) First, both Sen and Nussbaum argue that equality means that a just society recognizes each citizen's justified claim to at least a minimal level of freedom and well‐being. Nussbaum argues that good government enables its citizens to cross the two thresholds of being human and being able to flourish. Secondly, Nussbaum adds that equality here concerns ‘width’. It means that all—rather than just some—citizens are empowered to cross over the lines; at least insofar that each is (or can become) capable of functioning at these minimal levels, and adequate resources are available. This demand rules out sexism, racism and other prejudicial applications of the norms (Nussbaum, 1995: 42–51). Thirdly, equality means that the government should make resources and education available such that each citizen has the opportunity to function ‘not just minimally, but well, insofar as natural circumstances permit (Nussbaum, 1990a: 228). Recall, further, that one of Sen's criticism of the BNA was that it left the impression of ‘basic needs and no more’.

Fourthly, when resources are limited, such that everyone cannot reach the two thresholds and advance to full human flourishing, equality means that it is more important to empower everyone—who so chooses—to get over the line than promote or permit an elite to ascend higher (ibid.: 229). Why? Because once one has crossed the line, more is not necessarily better; and, even if it were, further attainments can be left to the person's own resourcefulness.

What if an elite's attainment of (or monetary rewards for) higher levels of functioning is or can be causally linked to more people crossing the threshold? Then, as in Rawls's difference principle, such inequalities would be justified. But what if elite achievements (or rewards) prevent those barely above the (first? second?) threshold either from advancing further or from closing the gap between the top and the bottom? Should elite achievement (rewards) be governmentally promoted at the cost of limiting the further advance of those just over the threshold?

Nussbaum begins to take up these questions when she argues that since more is not always better, elite achievements are not likely to require more resources, at least those such as money and property, that the less well‐off need to get above the thresholds. Rather, elites can achieve either by making better use of what they already have or by excelling in ways that do not (or should not?) depend so much on income and wealth, for instance, friendship, aesthetic experience, and political participation. Or if some accomplishments—we might think of film making—require large sums, then those interested can pursue the needed resources on their own rather than receive governmental transfers from those barely over the threshold. In a good and just society the most important use of governmental expenditures will be of getting everyone ‘across the threshold into capability to choose well’ (ibid.: 229).

This answer is certainly a good start. It hinges, however, on certain assumptions that may not hold, even in a good and just society, with sufficient frequency, namely, (i) that the elite will choose to define achievement apart from their monetary rewards, bank accounts, and sports cars, and (ii) that genuine achievements in politics, science, and the arts will not require government subsidies. We look forward to further consideration by Sen and Nussbaum on the proper balance between the egalitarian and ‘meritocratic’ components in just distribution. We also anticipate that Nussbaum will clarify in more detail the nature and relation of the two thresholds and their implications for government responsibility.

Finally, who or what should receive the just benefit of basic capability equality? ‘Each and every citizen’, is Nussbaum's answer. Nussbaum effectively criticizes sexist and racist applications of the capability ethic, especially Aristotle's prejudicial exclusion of slaves and women for his concept of citizenship (see Nussbaum 1988a and 1995: 42–51). Moreover, she correctly recognizes that citizens who are incapable of minimal functioning right now can acquire or develop—with the appropriate public arrangements—the requisite capability (see Nussbaum, 1990a: 243). But suppose such ‘line crossings’ were only possible with outlays that put others below the line? What does justice or mercy require, if anything, with respect to genetic humans and other animals completely incapable of reaching the threshold?³⁷ What does (retributive) justice require with respect to those actions in which an agent's intentional or foreseeable action puts someone else below the lines or deprives them in some other way of human functioning? When should such harm be a crime and what would be an appropriate punishment? What does (compensatory) justice demand with respect to rectification owed to victims of rights violations? Again these further questions must be put on the agenda of the capability theory of distributive, redistributive, and compensatory justice.

Sen treats questions of the ‘ethics of international income distribution’ even more sketchily than he does domestic justice; and Nussbaum, so far, has only stated that rich countries should help foreign poor as well as help their own poor get above the capability threshold (Nussbaum, 1991: 25; forthcoming b). Sen restricts himself to four suggestive points but has failed to take up other related questions. First, he argues that we should avoid discussing inter‐state justice as if nations were human persons, for such an approach would both prevent consideration of the absolute and relative deprivations of human individuals and tend to have the bad consequences of strengthening unjust national elites in both recipient and donor countries (Sen, 1984: 292–3, 297–9). So far, however, Sen has not considered what less‐than‐personal moral standing states can be said to have and how this standing should be weighed against the moral claims of persons.

Secondly, Sen argues that we should not presume that a given nation‐state is entitled to its wealth and that it is only alterations in the status quo that require justification. For a nation's wealth may be due to unjust past and present actions in relation to other states and peoples (Sen, 1984: 293). Yet so far Sen has not reflected on the principles we should use to identify such actions and what compensatory justice requires.

Thirdly, although Sen is open to giving some moral weight to national self‐interest, he correctly sees that such a concern has to be balanced with ‘positional‐neutral’ evaluations of deprivation and well‐being (Sen, 1984: 295–6). Again, we look forward to his moral assessment of nationality and national self‐interest in an interdependent world.

Fourthly, Sen and Nussbaum both assert that a moral case for international redistribution from rich nations to poor nations can be made. Sen maintains that such a case should not be based on the fact that persons in richer countries have higher incomes (per capita) than persons do in poor countries. The variation in conversions of commodities to capabilities may mean that the (per capita) well‐being of persons in some poorer countries, such as Sri Lanka, is higher than that in some richer countries, such as Saudi Arabia or Brazil (ibid.: 294). Even apart from the problem that intranation distributions are hidden by national averages, we need to recognize that life expectancy and literacy would be better indices of functionings and capabilities than would per capita GNP. This important point, made possible by the capability perspective, does not yet make, however, the positive moral case for international redistribution and inter‐state promotion of just development. It will be important, for example, for Nussbaum to clarify and defend further what she recognizes to be the international implication of her ‘fundamental commitment . . . to bring each and every person across the threshold into capability for good functioning’:

This view also generates clear global imperatives: for the countries whose citizens have a material surplus are, by holding on to that surplus, doing injustice to citizens in countries that are too poor to move everyone across the threshold (Nussbaum, forthcoming b).

In the light of their powerful capability ethic, then, Sen and Nussbaum need to consider the bases, nature, and extent of the moral responsibilities of rich nations, multinationals, or international bodies to reduce the following unconscionable deprivations that continue to exist in the developing world:

There are still nearly 900 million adults in the developing world who cannot read or write, 1.5 billion people without access to primary health care, 1.75 billion people without safe water, around 100 million completely homeless, some 800 million people who still go hungry every day and more than a billion who survive in absolute poverty.³⁸

The capability ethic that Sen and Nussbaum are forging has given international development ethicists a challenging, richly nuanced, and fertile resource. It forcefully identifies both strengths and weaknesses of commodity‐based, utilitarian, and deontological rights‐based ethics. Building on and deepening the basic needs perspective, it promisingly evaluates development theory and practice by the plural criteria of valuable human capabilities and achievements. Recasting the traditional social ideals of human freedom, rights, and justice, the capability perspective has in effect launched a new development paradigm. To contribute to the paradigm's further evolution has been a major aim of the present essay.