The capability approach to assessing the quality of life, as developed by Amartya Sen and Martha C. Nussbaum, has enjoyed a considerable degree of influence not only in normative economics but also among liberal egalitarian political philosophers. However, there are aspects of that approach that have yet to be subjected to rigorous critical scrutiny from the perspective of the political liberal. The aspect that I shall be investigating here is the role of the value of freedom in the capability approach, and the extent to which this role serves to limit the scope for paternalist intervention. Although some political liberals are prepared to accept some forms or instances of paternalist intervention, the degree of paternalism inherent in the capability approach ought to be of interest to all liberals (or at least to all liberals who find the capability approach interesting), given the widely endorsed presumption that paternalist intervention stands in need of justification.
The object of my investigation is the capability approach understood as a normative political theory, and not merely as a non-political (or not-necessarily-political) theory of the quality of life. In other words, I am interested here in the capability approach understood as a theory that contains or entails certain political prescriptions. Therefore, in using expressions like ‘the quality of life’, ‘being well off’, and so on, I mean, in the present context, the quality of life, being well off, and so on, insofar as these are the kinds of things that ought to be promoted or distributed fairly by political institutions.
It might of course be claimed that the capability approach does not provide the whole story as far as normative political theory is concerned. Thus, it might be said that the capability approach supplies only a theory of the good, whereas liberalism imposes a priority of the right over the good. Sen himself has sometimes suggested that the capability approach needs to be tempered or complemented by an independent concern with Rawlsian basic liberties. If we accept this suggestion, however, we shall accept it up to a point and for certain reasons. It is important to clarify these reasons. After all, the normative requirements of liberalism are often assumed or argued to be built into the capability approach, given its freedom-based understanding of the good and its consequent focus on capabilities, where a person's capabilities are said to represent her freedom of choice among a range of possible achievements.Footnote 1 My inquiry is confined to this supposedly liberal aspect of the capability theorist's account of the good.
Three alternative accounts can be given of the way in which capability theorists have incorporated freedom into their conception of the good. Each of these accounts implies a different interpretation of the capability approach in terms of the way it views the normative relation between capabilities and functionings, and each of these different interpretations of the capability approach can in turn be shown to permit a certain degree of paternalism. I shall examine these three accounts of the value of freedom, and their normative implications, in section 3 of this paper. In section 4, I shall outline an account of the value of freedom that appears to be excluded by all three of the interpretations of the capability approach. I shall argue that this last account of the value of freedom is a necessary part of any wholly anti-paternalist, freedom-based account of the good. I shall also argue that, despite appearances, capability theorists have no good reason to reject this account of the value of freedom.
Sections 1 and 2 provide the groundwork for the three interpretations of the capability approach presented in section 3. Section 1 provides an initial account of the capability approach and of its ambiguities, while section 2 points to a tension between two premises of the capability approach: anti-fetishism and anti-paternalism.
1. FUNCTIONINGS AND CAPABILITIES
I shall begin this first section by setting out the relevant characteristics of the capability approach. I shall then note an ambiguity in the central notions of functionings and capabilities, an ambiguity that will motivate further investigation of the space for paternalist intervention.
1.1 The capability approach
The capability approach involves assessing the quality of people's lives in terms of functionings and capabilities. Functionings are ‘doings’ and ‘beings’ – that is, the actions a person performs and the states of being she achieves. Examples of valuable functionings are ‘walking’, ‘being well nourished’, and ‘associating with others’. A person's capabilities are her abilities to achieve functionings. I have the capability to walk if my internal abilities and external resources are such that I would succeed in walking if I tried. A life of quality is something more than a life made up of valuable functionings; it is also one in which the individual has available a certain range of alternative functionings, and is therefore able to exercise her faculty of choice. Measurements of the quality of life therefore need to take into account a person's ‘capability set’, which is the set of alternative functionings available to the individual, given her internal abilities and the resources at her disposal.
In order to clarify the reasons behind this understanding of the quality of life, it will be useful to contrast it with two older traditions in normative economics of which Sen, in particular, has provided powerful critiques: those of welfarism and resourcism.
Welfarism is the view that the quality of a person's life depends on the degree to which her preferences are satisfied. Sen has advanced two main objections to welfarism. First, its metric of well-being is purely subjective and leads to counterintuitive measurements from an egalitarian point of view. Most importantly, it fails to take into account the phenomenon of adaptive preferences: a person who is materially disadvantaged may adapt her preferences to her situation, as a result of which her disadvantage fails to show up on a welfarist metric of advantage (Sen Reference Sen1980: 202–5; Reference Sen1985b: 191; Reference Sen1999: 62–3). In order to accommodate this objection, we need a metric of individual advantage that takes into account more objective factors, such as a person's talents, resources and environment. The second objection advanced by Sen is that welfarism assesses well-being only in terms of outcomes and so fails to take into account the fact that the quality of a person's life also depends on the possibilities she has of bringing about certain outcomes. As Sen puts it, freedom has value ‘over and above . . . what freedom permits us to achieve’ (1988: 290). The good life ‘is, inter alia, a life of freedom’ (1985b: 202). Welfarism treats persons as patients rather than as agents, as passive recipients of benefits, rather than as choosers, each with his or her own life to lead.
Both of the above-mentioned inadequacies of welfarism – the lack of objectivity in its metric of advantage and the failure to recognize the independent value of freedom – appear to be avoided by resourcist egalitarian theories like those of John Rawls (Reference Rawls1971) and Ronald Dworkin (Reference Dworkin1981). Resources (like, for example, income and wealth or, more broadly, Rawlsian primary goods) give us a more objective metric and, being means to many alternative ends, seem also to guarantee their owners a degree of freedom. What, then, is wrong with the resourcist perspective, according to capability theorists? Its fault lies in its failure to take account more directly of those factors that really affect the quality of a person's life: the valuable functionings that resources help us to achieve. While resources are a necessary means to achieving functionings, they convert into those functionings (the ends to which they are means) at different rates in different people, given interpersonal differences in terms of internal and environmental factors. For example, if physically disabled people need wheelchairs in order to move around adequately, they convert resources into the functioning ‘moving around adequately’ less efficiently than do normally able people.
Which doings and beings, exactly, contribute to a life of quality? Nussbaum has provided a definite list: being healthy, using all five senses, reflecting critically, affiliating, enjoying humour, participating politically. . . . Sen has avoided drawing up such a list – ‘a wise move, or passing the buck – depending on your point of view’ (Beard Reference Beard2000: 6). But the exact nature of the functionings on the list need not concern us here. What matters, for present purposes, is that some kind of list is required – a list which, however vaguely, includes certain doings and beings and excludes others, or at least establishes an order of priority among them – for otherwise it will no longer be clear why the capability approach is not identical to welfarism (or at least to a theory prescribing opportunity for welfare), given the consequent necessity of falling back on preferences for the selection and/or the relative weighting of functionings. The list might be based on a broad cross-cultural consensus and be realizable in different ways in different contexts, as Nussbaum claims of her own list. Even a cross-culturally valid list, however, will include and/or prioritize some doings and beings and not others. If there are doings and beings that are valued within all cultures, there are likely also to be ones that are not valued or are disvalued within all cultures.
1.2 A capabilitarian trilemma
Capability theorists have not always been clear about the conceptual relation between functionings and the good life. Are functionings necessarily good ends? Or are there such things as ‘bad functionings’? How, moreover, will our answer to this last question affect our understanding of the notion of a ‘capability set’? Addressing these questions will bring out a basic ambiguity in the capability approach that underlies much of the discussion to follow.
In the value-laden sense deriving from Aristotle, a functioning is necessarily an appropriate or valuable doing or being, and the expression ‘bad functioning’ is an oxymoron. This interpretation of functioning probably fits best with ordinary language. As H.L.A. Hart points out, ‘[w]e say it is the function of the heart to circulate the blood, but not that it is the function of a cancerous growth to cause death’ (Hart Reference Hart1994: 191). It is true that ordinary language contains words like ‘malfunction’ and ‘dysfunction’, which might be thought to show that functionings can themselves consist in bad doings or beings. However, in ordinary language, to ‘dysfunction’ is not to achieve ‘bad functionings’; rather, it is to do or be the opposite of one or more of the good things in which functioning consists, where functioning is itself valuable. And while ‘malfunctioning’ does mean functioning in a certain way, it is not normally understood to mean functioning in a way that is bad in an absolute sense; rather, it is understood to mean functioning at a level that is intermediate between functioning fully and not functioning at all (as when we say ‘this car is malfunctioning’).
Nevertheless, it is possible to take a contrary view and to stipulate that the term ‘functioning’ is to be understood in an evaluatively neutral way, so that all conceivable doings and beings – including, say, suicide and homicide – count as functionings. On this second understanding of functioning (assumed, for example, by Peter Vallentyne Reference Vallentyne2005), only a subset of the set of all functionings will appear on the list referred to by the capability approach. (It should be noted that the question of which of these two senses of ‘functioning’ one assumes is independent of the question of the basis – Aristotelian or otherwise – on which the capability theorist is to select the relevant list of functionings.)
Nussbaum clearly assumes the value-laden sense of functioning. Sen is more equivocal, and perhaps with good reason, for the value-laden sense of functioning is one horn of a trilemma according to which no more than two of the following three claims can be made coherently:
1. The concept of functioning is value-laden in the above sense.
2. The concept of capability is dependent upon that of functioning, in the sense that a person's overall capability set is nothing more or less than the set of alternative functionings available to her. In Sen's words, ‘capability is defined in terms of the same focal variables as functionings. In the space of functionings, any point represents an n tuple of functionings. Capability is a set of such functioning n tuples . . . . [T]here is no difference as far as the space is concerned between focusing on functionings or on capabilities’ (Sen Reference Sen1992: 50).Footnote 2
3. Capabilities include not only the freedom to function in valuable ways (as specified by the capability theorist's list of functionings) but also the freedom to choose not to function in valuable ways – that is, the freedom to dysfunction.
If we accept claims 2 and 3, we must reject the value-laden understanding of functionings (that is, claim 1): if we accept Sen's claim about the identical space of functionings and capabilities, and we accept that a capability set includes the freedom not to function in the valuable ways specified by the list (claim 3), then we must also accept that not all functionings are valuable. If, on the other hand, we accept claims 1 and 3, we must reject claim 2, for the space of functionings and capabilities will then no longer be identical. Finally, if we accept claims 1 and 2, we must reject claim 3: although we can continue to say that having a capability entails having a choice, the relevant choice will only be between alternative functionings which (given claim 1) are all valuable in the sense of appearing on the relevant list.
I shall assume here that the capability theorist will not reject claim 3. This assumption is necessary in order for my investigation to get off the ground. If the capability theorist rejects claim 3 while nevertheless considering the prohibition of dysfunctional behaviour to be morally problematic, then she must rely, to a greater extent than is normally supposed necessary, on freedom-based normative requirements that are exogenous to the capability approach itself. My investigation concerns the rather different view according to which at least some of those freedom-based requirements are already built into the capability approach.
I shall remain agnostic, on the other hand, over which of the first two claims must be rejected. What concerns me here is the substantive question of how, and to what extent, the role of the value of freedom in the capability approach limits the scope for paternalist intervention. One can answer this substantive question while abstaining from making a choice between rejecting claim 1 and rejecting claim 2. I shall therefore avoid the question of whether functioning and capability are necessarily value-laden concepts by referring to the ‘valuable functionings’ on the list, to the capability ‘to achieve good functionings’, and so on, while admitting that such usages will be pleonastic on the respective value-laden interpretations.
2. ANTI-FETISHISM AND ANTI-PATERNALISM
On the basis of the foregoing account, the capability approach can be understood as arising out of two distinct theoretical moves. These two moves are illustrated by the capabilitarian critiques of resourcism and welfarism, and motivate the focus on functionings and capabilities respectively. I shall call these two moves anti-fetishism and anti-paternalism. In this section I shall spell out the nature of these moves and shall then point to a tension between them.
2.1 The value of ends and the value of freedom
The rejection of resourcism depends on an anti-fetishist move. I take fetishism to involve over-representing the true worth of a phenomenon by treating its extrinsic value as if it were intrinsic value.Footnote 3 I shall assume here that the intrinsic value of a phenomenon is the value it has as an end in itself, and that its extrinsic value is the value it has in virtue of its relation to some other valuable phenomenon. Where something has both intrinsic and extrinsic value, fetishism can arise from a misrepresentation of the degree to which its value is intrinsic. Sen has explicitly used the term ‘fetishism’ in criticising the Rawlsian focus on primary goods.Footnote 4 Resources are valuable, not as ends in themselves, but in virtue of their relation to valuable achievements, that relation being one of enablement (resources enable a person to achieve certain things).
Given the value of freedom, however, an account of the good is of course incomplete if it refers only to actual achievements. The aim is not to force people to function well. Therefore, in addition to the shift in focus from resources to functionings, capability theorists favour the shift in focus from functionings to capabilities. This last shift in focus constitutes an anti-paternalist move.
I shall assume here that the conduct of agent A is paternalist with respect to agent B if it constitutes a limitation of B's available options and is carried out in order to further B's interests. This assumption does not amount to a complete definition of paternalism. Rather, it states a sufficient condition. Some theorists of paternalism have claimed that an act can be paternalist without restricting a person's freedom.Footnote 5 In the present context, however, I shall use the term ‘anti-paternalism’ to refer exclusively to opposition to behaviour that restricts people's freedom in their own interests. Thus, I shall not be concerned here with exercises of social power (sometimes called ‘nudging’) that are said to influence people's choices by rearranging their options without removing any of them – although some liberals find these exercises of power objectionable too. Instead, I shall be concerned with forms of behaviour that are paternalist in a strong sense: coercion, the removal of certain options (including, but not limited to, the prevention of weak-willed or short-sighted behaviour), or the introduction of disincentives that limit overall freedom by imposing costs on certain options.Footnote 6
Why do liberals consider paternalism to be morally problematic? One well-known answer to this question is that paternalism conflicts with the liberal premise of equal respect for persons. To respect persons is to treat them as moral agents or, as Kantians put it, as ends in themselves. To treat persons as ends in themselves is to treat them as beings that set ends –that is, as ‘originators of ends’ or ‘points of origin of ends’ (Taylor Reference Taylor, Pelczynski and Gray1984: 114).Footnote 7 In Rawlsian terms, it is to treat them as enjoying a certain (equal) status as beings that possess the ‘moral power’ to form, revise and pursue a conception of the good. To act paternalistically toward another person is implicitly to deny that the other possesses this moral power at an adequate level (Quong Reference Quong2011: §3.5). It is also, standardly, to assume one's own superiority in terms of possession of this power. Thus, paternalism involves an implicit rejection of the basis of our moral equality.Footnote 8
What seems to be special about the capability approach is that it places importance both on the ends of life (hence the anti-fetishist move) and on respect for persons and for their freedom (hence the anti-paternalist move). Resourcist theories are anti-paternalist, but they are also fetishist; welfarism is anti-fetishist, but it is also compatible with paternalism. Only the capability approach appears to avoid both paternalism and fetishism.
Nevertheless, there is a tension between the anti-fetishist move and the anti-paternalist move. The simplest way to state this tension is as follows: whereas the anti-paternalist move points toward a focus on capabilities rather than on the realization of valuable functionings – it tells us to treat people as agents – the anti-fetishist move appears to point toward a focus on achieved functionings – it tells us to take account more directly of the degree to which the ends of life are actually realized.
The above statement of the tension between the anti-fetishist move and anti-paternalist move is plausible, but it does not, in fact, perfectly characterize that tension. In order to provide a more precise characterization of the tension, we now need to engage in a more fine-grained analysis of the ways in which one good can ‘convert’ into another.
2.2 Instrumental value, contributory value and conversion rates
It is true that anti-fetishism can point toward a focus on realized functionings, given that realized functionings are among our ends in life. But it is also true that anti-fetishism can point toward a focus on capabilities, since freedom can have intrinsic value and can itself therefore qualify as one of the ends of life. If freedom has intrinsic value, then we should be concerned not only with the rate of conversion from resources into functionings, but also with the rate of convertibility from resources into functionings, which is to say, with the rate of conversion from resources into capabilities to function. From the point of view of the capability theorist, resourcism is fetishist in the sense that it ignores both rates of conversion: from resources into functionings and from resources into capabilities.
We have seen that fetishism involves the over-representation of the true worth of a phenomenon by treating its extrinsic value as if it were intrinsic value. We should now take into account a distinction between two kinds of extrinsic value. This distinction will help in clarifying the capability theorists’ critique of resourcism, as well as serving us later in our examination of the value of freedom. The two kinds of extrinsic value are instrumental value and contributory value. Let us say that a phenomenon has instrumental value if it is a causal means to the realization of some other valuable phenomenon, and that a phenomenon has contributory value if it is a non-causally necessary but insufficient condition for some other valuable phenomenon.
The relation between resources and functionings is of the instrumental kind: resources are valuable as a necessary element in a causal chain leading to the realization of valuable functionings. The relation between resources and capabilities to function, on the other hand, is of the contributory kind: resources are valuable as a non-causally necessary but insufficient condition for being free to realize valuable functionings. (Sen himself claims that resources are ‘means to freedom’, as if the relation between resources and freedom were instrumental rather than contributory (Sen Reference Sen1990; Reference Sen1992: 37–8, 81–4). However, one does not possess resources and then achieve freedom with them; rather, one possesses freedom in virtue of one's possession of resources. Resources therefore ‘convert’ into capabilities in a non-causal, non-instrumental sense – not through an act of conversion but, rather, in the same sort of way as that in which having beautiful colours ‘converts’ into being a great painting, thanks to the presence of other factors such as shape and form. More technically, where contributory value is concerned, rates of ‘conversion’ are from subvening phenomena into supervening phenomena.)
The fact that anti-fetishism can lead us to prefer a focus on freedom to a focus on resources shows that anti-fetishism and anti-paternalism do not necessarily point in opposite directions. But while the two moves do not necessarily point in opposite directions, neither do they necessarily point in the same direction. This is because anti-fetishism can tell us to focus in part on realized functionings, whereas anti-paternalism necessarily draws us away from that focus.
The remaining tension between the two moves depends on differences in a third rate of conversion: that from capabilities into functionings. Like the other two conversion rates just mentioned (from resources into functionings and from resources into capabilities), the conversion rate from capabilities into functionings varies from one person to another – this time, depending on differences in the quality of people's actual choices. Consider the example of an individual who freely chooses to neglect her own cultural enrichment and/or her own future health and/or significant opportunities to affiliate with others, preferring instead to invest the bulk of her resources in an extremely expensive sports car. Let us assume that this outcome constitutes an example of dysfunctioning, at least in the sense of reflecting an erroneous choice regarding the proper order of priority among functionings, if not the choice of an alternative that is completely absent from the appropriate list of functionings. (If this assumption is denied, we shall always be able to find an alternative example, just as long as capability theorists assume some list of functionings and/or some order of priority among them – which, as we have seen, they surely must.) Cases of this sort are sometimes cited by capability theorists as examples of ‘functioning poverty’: although not poor either in terms of resources or in terms of capabilities, our sports-car enthusiast is poor in terms of achieved functionings (Balestrino Reference Balestrino1996; Sen Reference Sen1996: 123–4).Footnote 9
How should a normative political theory deal with failures or inferiorities in this third rate of conversion? If it were based exclusively on functionings, such a theory would prescribe that the government take the most efficient measures available to ensure that the individual in question functions better. The intrusiveness of such measures would not in itself be an issue for such a theory, since the optimal degree of intrusiveness would depend solely on its degree of efficiency in realizing the desired functionings. The capability approach is not, of course, a theory based exclusively on functionings; it is a theory based on capabilities to function instead of, or in addition to, functionings. Nevertheless, exactly how the capability approach would deal with such low conversion rates will only become clear once we have established in what ways, exactly, the approach places value on freedom. Does freedom have instrumental value as a means to the realization of valuable functionings, contributory value as a non-causally necessary but insufficient condition for valuable functionings, or value as an end in itself?
If freedom's value were wholly instrumental – if freedom were nothing more than a means to the realization of valuable functionings – it would be difficult to see why the existence of poor conversion rates from capabilities into valuable functionings would not lead us to take an interest more directly in the realization of valuable functionings (Sen Reference Sen2009: 236). Perhaps for this reason, capability theorists have also referred, implicitly or explicitly, both to the contributory value of freedom and to the intrinsic value of freedom. Our next task, then, is to consider the capability approach on the assumption that it ascribes to freedom one or another of these kinds of value.
3. THREE INTERPRETATIONS OF THE CAPABILITY APPROACH
I shall now present three views of the value of freedom that might plausibly be attributed to capability theorists. Each of these views motivates a distinct interpretation of the capability approach in terms of the normative relation it assumes to exist between functionings and capabilities.
3.1. The contributory value of freedom: functionings and capabilities as jointly necessary conditions for a life of quality
Saying that freedom has contributory value as a non-causally necessary condition for good functioning seems to provide a promising basis for opposing paternalist intervention, because it seems to make the forcible realization of valuable functionings a conceptual impossibility. Moreover, in the case of some particular functionings it is immediately plausible to ascribe this kind of value to freedom. Consider the examples of hospitality and love, both of which appear on Nussbaum's list. It is clearly impossible for one person to force another to achieve either of these particular functionings, for hospitality and love are properly interpreted not simply as certain forms of outward behaviour but as behaviour accompanied by appropriate intentions. In other words, hospitable or loving behaviour cannot be what they are without being the result of a free choice. Or consider an example repeatedly put forward by Sen: fasting is different from starving, given that the freedom to eat is a necessary condition for the former but not for the latter (Sen Reference Sen1992: 111–12; Reference Sen1999: 76; Reference Sen2009: 237). Where two individuals desire to fast, and the first is given food while the second is deprived of food, the first will fast whereas the second will merely starve. The freedom to choose whether or not to eat is therefore a constitutive part of the practice of fasting, and has value as long as fasting is in turn a constitutive part of some broader cultural practice that itself has value – for example, a religious or political practice. (Sen himself has claimed that this example demonstrates the ‘intrinsic’ value of freedom. In my view it demonstrates, rather, the contributory – and therefore extrinsic – value of the freedom to choose whether or not to eat. The point is that ‘[h]aving the option of eating makes fasting what it is, to wit, choosing not to eat when one could have eaten’ (Sen, Reference Sen1999: 76, my emphasis).Footnote 10)
This conceptual point might not be applicable to all of the functionings on the list, especially basic ones like moving about, being well nourished, using all five senses, and so on, which might be defined in purely behavioural terms and without reference to the process by which they are brought about. Nevertheless, the conceptual implication we have noted in the cases of hospitality and love can be extended, stipulatively, to cover the other functionings. We can stipulate, that is, that the free choice of any functioning on the relevant list is a necessary condition (in an analytic rather than a causal sense) for the achievement of that functioning. The set of valuable functionings will then coincide with the set of functionings that Sen would call ‘refined’, where refined functionings are functionings that are chosen from a set of possible functionings (Sen Reference Sen, Sen and Nussbaum1993: 40).
This view of freedom as possessing contributory value suggests a particular interpretation of the way the notions of capability and functioning are to be incorporated into the capability theorist's conception of the good: capabilities and functionings are jointly necessary conditions for a life of quality. Support for this interpretation can be found in Nussbaum's claim that ‘capability-needs are important because of the value of the functionings in which they naturally terminate’, while ‘functionings are valuable, in part, for the way in which they realize capabilities. We cannot and should not prise the two apart’ (Nussbaum Reference Nussbaum and Patzig1987: 173).Footnote 11 Similarly, David Crocker has suggested that ‘human well-being . . . consists [for Nussbaum] only in those functionings that are both chosen and valuable’ (Crocker Reference Crocker, Nussbaum and Glover1995: 166, my emphasis). And according to Serena Olsaretti, what explains ‘the capability approach's insistence on the freedom to achieve functionings, rather than on achieved functionings’, is ‘a model of well-being which views the endorsement of valuable functionings as constitutive of well-being’, together with the view that ‘endorsement of valuable functionings is best secured when people choose freely which functionings to achieve, rather than being forced to achieve them’ (Olsaretti Reference Olsaretti2005: 98–100).
Despite assigning a central role to freedom, this first interpretation of the capability approach leaves ample margin for paternalist intervention. Why? Because the goal remains the realization of functionings, albeit functionings of a certain kind (that is, freely chosen functionings). Freedom converts into these kinds of functionings at different rates in different individuals, given differences in the quality of their actual choices. If differential conversion rates should worry us in the case of the relation between resources and freedom (a relation which, it should be recalled, is of the contributory kind), they should also worry us in the case of the relation between freedom and freely chosen functionings.
Now it is true that there is an important limit on what we can do to compensate for a poor conversion rate from freedom into functionings – a limit that does not exist in the case of a poor conversion rate from resources into freedom. In the latter case, we can usually compensate for a poor conversion rate simply by increasing the person's level of resources. In the former case, such compensation is less easily achieved, for if we restrict the person's freedom in an attempt to direct her towards the achievement of valuable functionings we shall be removing a necessary element of those same valuable functionings. However, the introduction of disincentives, the prohibition of various personal vices, or even the use of direct force, might still be justified as the most efficient way of promoting voluntary valuable functioning in the long run. Such measures might serve an educational purpose, opening people's eyes to the value of the doings and beings that they are directed to achieve.Footnote 12 After a period of non-voluntary pursuit of certain ends, during which the quality of the individual's life is no better than when she was dysfunctioning (but also no worse, given that freedom is a necessary but insufficient condition for valuable functioning), the individual might come to endorse such ends, to form goals or projects that include those ends, and to pursue those goals or projects voluntarily. ‘Compulsion is justified by education for future insight’ (Fichte Reference Fichte1845: 578).Footnote 13 Whether such short-term measures can indeed be justified in this way will be an empirical question. They are not ruled out by the mere fact that freedom or voluntary endorsement is a constitutive part of valuable functioning.
3.2 The intrinsic value of freedom: functionings and capabilities as disjunctively necessary conditions for a life of quality
Let us assume, next, that the freedom to function well (and not to function well) has intrinsic value as well as contributory value. In this case, even the individual who freely chooses not to function well leads a life that has some value. Thus, a government should give importance, in its assessments of the quality of life, not only to actual (freely chosen) functionings, but also to freedom of choice itself, including the freedom to choose ‘none of the above’. (This is not to deny, on the other hand, that the best-off individual will be one who is capable of functioning well and chooses freely to do so.)
This second view of the value of freedom suggests a second interpretation of the capability theorist's view of the normative relation between capabilities and functionings: capabilities, on the one hand, and realized functionings, on the other, are disjunctively necessary conditions for well-being. Support for this interpretation can be found in Sen's claim that, given the intrinsic value of freedom, ‘[t]he focus of the capability approach is not just on what a person actually ends up doing, but also on what she is in fact able to do’ (2009: 235). Most importantly, Sen often seems to want to leave open whether public policy should concentrate more on the promotion of functionings or on that of capabilities, explicitly using the name ‘capabilities approach’ to refer to a perspective that concentrates ‘either on the realized functionings . . . or on the capability set of alternatives’ (1999: 75).Footnote 14 Emphasizing the complexity of his normative theory and the consequent difficulty of assigning relative weights to various functionings and various capabilities in order to arrive at overall measurements of the quality of life, Sen affirms that ‘the capability perspective is inescapably pluralist. First, there are different functionings, some more important than others. Second, there is the issue of what weight to attach to substantive freedom (the capability set) vis-à-vis the actual achievement (the chosen functioning vector)’ (Sen Reference Sen1999: 76). He further suggests that the existence of a plurality of evaluative criteria, and the consequent possibility of incommensurabilities and of incomplete orderings, should not be seen as a weakness or a source of embarrassment. The problem of assigning relative weights to capabilities and realized functionings cannot and should not be avoided simply by concentrating exclusively on one of them; rather, we need to recognize the complexity of our evaluations, and seek a consensus over the relative weights of their various dimensions: ‘this is a “social choice” exercise, and it requires public discussion and a democratic understanding and acceptance’ (Sen Reference Sen1999: 78–9).
What will be the fate, on this second interpretation, of our individual with the tendency freely to choose not to function well? The quality of her life will no longer be seen as zero, since the freedom she has (to choose whether or not to function well) will itself be of value regardless of what choices she actually makes. Nevertheless, a government strategy that includes, for the individual's own good, certain disincentives, or prohibitions, or even direct force, will remain a normative possibility. The question of whether a government may (or must) adopt such measures need not necessarily be answered in the negative, but will depend on the further question of the relative weights to be attached to freedom (including the freedom not to function well), on the one hand, and achievement (of valuable functionings), on the other, in determining the quality of lives. We have seen, in discussing the first interpretation, how a restriction of the freedom not to function well can improve a person's lot even in terms of freely chosen valuable functionings. Whether such a restriction of freedom is permissible on the second interpretation will depend on how much the intrinsic disvalue of that restriction counts in our overall evaluations. The answer to this last question is left open by our second interpretation of the capability approach, and the task of answering it is apparently to be devolved to the sphere of democratic deliberation. Moreover, as long as the starting point for that democratic deliberation includes the premise that realized functionings should have some weight in our overall evaluations, it is difficult to imagine the outcome of that deliberation not leaving some space for paternalist intervention.
3.3. The political value of freedom: the quality of life as the capability to achieve valuable functionings
The surest way to accommodate the anti-paternalist objection just moved against the second interpretation seems to be to resist any temptation to focus on realized functionings within the context of a normative political theory. We should specify that the list of capabilities expresses, in the words of Nussbaum, ‘not a complete account of the good . . . but a political account’ (Nussbaum Reference Nussbaum2000b: 148; see also Reference Nussbaum2006: 71). We should state more clearly that not all aspects of the good (in the comprehensive or ‘complete’ sense of the good) contribute to the quality of life in the politically relevant sense. And we should say that, at the political level, the intrinsic, instrumental and contributory value of capability is relevant in assessing the quality of life, whereas the value of realized functioning is not. In this strictly political sense, individuals who freely choose not to function well will not, as such, be considered any worse off than those who freely choose to function well: how well off a person is will depend exclusively on how far she has the capability to achieve valuable functionings.
Along these lines, Nussbaum has claimed that the capability approach ‘does not aim directly at producing people who function in certain ways’ but, rather, ‘at producing people who are capable of functioning in these ways, who have the training and the resources to so function, should they choose. The choice itself is left to them’ (Nussbaum Reference Nussbaum, Douglass, Mara and Richardson1990: 214; see also Reference Nussbaum2000a: 128–32). Nussbaum has also come out explicitly against the direct prohibition of dysfunctional behaviour: ‘I myself take a more libertarian line here: that is, I do not favour policies that would make unhealthy activities such as boxing, unsafe sex, football, and smoking illegal’ (2006: 171). Sen's theory is often interpreted in a similar way, despite his acceptance that governments may, at times, concentrate directly on functionings. Although Sen claims that our evaluative criteria can include achieved functionings, he has also suggested that this concession need not be seen as anything more than a practical expedient to be adopted in the absence of data about capabilities. We might use information about realized functionings only as a general indicator, ‘to try to surmise the capability enjoyed by a person’ (Sen Reference Sen1992: 111–12).Footnote 15
This third interpretation of the capability approach appears to concentrate exclusively on capabilities and not at all on functionings. But while there is a sense in which this is true, there is another in which it is not. It is true that the capability approach on this interpretation never concerns itself with the realization of valuable functionings. It is not true, however, that it wholly discards the political relevance of valuable functionings, for it retains a concern with their possibility rather than with other possibilities. What matters, on the third interpretation, is that people enjoy certain specific capabilities, not that they enjoy capability as such, and in order to identify such specific capabilities we shall still need to refer to a list of independently specified functionings. The value of freedom therefore remains dependent upon its content – on what it is the freedom to do or become – and that content depends, in turn, on the specification of a set of ends worth realizing. As a result, even the third interpretation of the capability approach leaves a residual space for paternalism.
One way of illustrating this residual space for paternalism is by comparing alternative ways of compensating for capability deficits. Consider the case of Arthur, who walks with a limp and will only be able to walk normally if he is given an operation he cannot at present afford. The capability approach, on the third interpretation, will presumably favour offering Arthur the operation. He will not be forced to undergo the operation; he can take it or leave it. Suppose, however, that giving Arthur the cash equivalent of the operation would give him additional freedoms without removing any of the freedoms implied by the offer of the operation, providing as it does the freedom to undergo the operation plus many other freedoms that involve spending the money in alternative ways. It does not seem to follow from the third interpretation of the capability approach that, in such circumstances, we must favour giving Arthur this greater freedom by giving him the cash rather than offering him the operation. Rather, on the third interpretation, the capability approach permits us to give Arthur the more restricted in-kind benefit, and is indeed compatible with the judgement that many of the additional freedoms supplied by the cash benefit themselves have disvalue.
Of course, many liberals consider it just, on various grounds, to refuse to substitute in-kind benefits with cash benefits. My point here is simply that the capability approach, even on the third interpretation under consideration, permits such a refusal on paternalist grounds. Moreover, those mainstream liberals who are happy to treat Arthur paternalistically might be less happy with some of the other paternalist possibilities brought to light by the above illustration. If the reason for removing the freedom to spend the compensation for a capability deficit on a sports car rather than a hip operation is simply that the freedom to make this choice adds no value to a person's life, or perhaps even that it adds disvalue, why not say the same of the freedom to become a couch potato, to eat unhealthily, or to engage in various sorts of non-social or anti-social behaviour that liberals would normally tolerate? The capability approach will prescribe each of the latter freedoms only to the extent that each of them can be shown to have value, intrinsically or extrinsically, in a way that derives from the nature of the thing it is the freedom to do or become – either because that thing is among the functionings on the list or because it represents an important particular alternative to one or more such functionings. Capability theorists have spent much time constructing lists of valuable functionings that people should be free to achieve, and free not to achieve; they have said very little about the particular content that freedoms-to-dysfunction must have in order to be valuable. Although most capability theorists are probably inclined against the curtailment of such freedoms (I have cited Nussbaum to this effect), the capability approach itself does not appear to supply us with the theoretical resources needed to justify such an inclination.
4. THE CONTENT-INDEPENDENT VALUE OF FREEDOM
We have seen that Sen ascribes intrinsic value to freedom. However, ascribing intrinsic value to freedom is not the same as ascribing value to freedom as such, where the latter means ascribing value to freedom independently of the nature of the specific things that it is the freedom to achieve. In the literature on freedom, the kind of value referred to by the latter expression has been called ‘content-independent’ or ‘non-specific’ value (Carter Reference Carter1999: chs 2, 5; Van Hees and Wissenberg Reference Van Hees and Wissenberg1999; Van Hees Reference Van Hees2000; Kramer Reference Kramer2003: ch. 2 sec. 4, ch. 5; Carter Reference Carter, de Wijze, Kramer and Carter2009). In what follows, I shall first explain how content-independent (or non-specific) value differs from intrinsic value. I shall then argue that the capability approach does not itself supply us with reasons against incorporating a recognition of freedom's content-independent value.
4.1 Intrinsic value and content-independent value
The distinction between content-independent (or non-specific) value and content-dependent (or specific) value is orthogonal to the distinction between intrinsic value and extrinsic value. The freedom to do or become a specific thing, considered as the freedom to do or become that specific thing, can itself have either intrinsic, instrumental or contributory value. Thus, intrinsic value, no less than instrumental or contributory value, can be content-dependent value. Capability theorists recognize that freedom has these three kinds of content-dependent value. But freedom can also be valuable as such – that is, in a way that is independent of its content – either intrinsically, instrumentally or in a contributory sense. For example, freedom might be said to have content-independent instrumental value because we lack knowledge about the specific nature of the ends to which we believe freedom to be a means: we might believe that freedom is a means to the realization of our individual values, or to the realization of societal progress (be it economic, social or moral progress), while also recognizing our inability to know our own future values or the forms that progress might take. J.S. Mill's insistence on the importance of ‘experiments in living’ can be interpreted along these lines (Carter Reference Carter1999: ch. 2). Or again, and less contingently, freedom might be said to have content-independent contributory value because it is a constitutive part of some wider individual good – for example, autonomy, or the Rawlsian ‘social bases of self-respect’ – given the nature of persons as setters of ends and the consequent inappropriateness – indeed, the disrespectfulness – of identifying any of their ends for them. Capability theorists tend to deny, explicitly or implicitly, that freedom has this content-independent kind of value. As Sen has put it, ‘[w]hile it is important to see that freedom can have a value that is additional to the value of the actual achievements, the requirement of [content-] independence calls for a “dissociation” that would be hard to justify’ (Sen Reference Sen1996: 110).
The capability theorist's tendency to place value on freedom only in a content-dependent way, and never in a content-independent way, is a feature that remains constant across the three interpretations of the capability approach. That tendency concerns not so much the relation between freedoms and realized functionings as the relation between freedoms and the hypothetical functionings that they are the freedom to achieve. The way the capability theorist conceives of the former relation creates space for paternalism only on the first two interpretations of the capability approach. The way the capability theorist conceives of the latter relation creates space for paternalism on all three interpretations, for on all three interpretations capability theorists are committed only to providing a certain set of particular freedoms. We have assumed that these particular freedoms include the freedom not to function well. But the freedom not to function well is a very general freedom-type, and is no less present when the range of freedoms-to-dysfunction is narrow than when it is broad. What capability theorists seem to lack is a reason for preferring more freedom to less freedom, in particular in cases where what would serve to expand a person's overall level of freedom is a set of additional freedoms-to-dysfunction. Indeed, if even the intrinsic value of freedom depends on its content, it is difficult to see why one should deny that increases in the freedom to dysfunction often have content-dependent disvalue, all things considered.
4.2. Why not content-independent value?
The anti-paternalist move can be completed by advocating a focus not just on capabilities rather than functionings but, further, on capability as such rather than, or in addition to, specific capabilities. The latter focus would give us a reason for giving an agent a greater freedom-to-dysfunction rather than a lesser one, compatibly with that agent's not compromising the rightful freedom of others. Why do capability theorists resist this further move? In the light of the foregoing analysis, two possible explanations suggest themselves.
(i) Anti-fetishism
A first explanation might be that the capability theorist feels, consciously or unconsciously, that the thesis of the content-independent value of freedom should be resisted on anti-fetishist grounds. We have seen how the anti-fetishist move is in tension with the anti-paternalist move: the anti-paternalist move tells us to leave the ends of life out of the picture, while the anti-fetishist move tells us not to do so. The anti-fetishist tendency indeed draws capability theorists away from the third interpretation of the capability approach and back toward the second or the first, and so helps to explain the tendency of some such theorists to oscillate between the three interpretations: to the extent that specific freedoms are recognized as having not only intrinsic value but also extrinsic value – that is, value as means to, and/or as non-causally necessary but insufficient conditions of, specific valuable functionings – even the third interpretation of the capability approach seems to be open to the charge of fetishism. A fortiori, it might be said, the anti-fetishist tendency motivates capability theorists in resisting any move beyond the third interpretation, further ‘dissociating’ the value of capability from that of functioning.
Let us assume that the capability theorist is anti-paternalist enough to embrace the third interpretation. Is anti-fetishism now a good reason for not going beyond the third interpretation in the direction of a recognition of the content-independent value of freedom? The following two points suggest that it is not.
First, anti-fetishism opposes the representation of merely extrinsic value as intrinsic value, and we have seen that the distinction between intrinsic and extrinsic value is orthogonal to the distinction between content-independent and content-dependent value. There is no reason to consider the ratio of intrinsic to extrinsic value to be any more biased toward the extrinsic component in the case of the content-independent value of freedom than in the case of its content-dependent value. Therefore, anti-fetishism should not lead one to object to the content-independent value of freedom any more strongly than it should lead one to object to the capability approach on the third interpretation.
Second, there is a sense in which a normative theory that recognizes the content-independent value of freedom is less open to the charge of fetishism than is the capability approach on the third interpretation. To the extent that the extrinsic value of freedom is of the content-independent kind, it is simply inappropriate to assess the value of particular freedoms in terms of the rates at which they convert into the realization of particular individual ends. The whole idea behind the content-independent value of freedom – including its instrumental and contributory content-independent value – is that the value of freedom is independent of any such ends. In this sense, the notion of content-independent value places a conceptual limit on the scope of anti-fetishism: to the extent that the extrinsic value of freedom is recognized to be of the content-independent kind, there are no particular individual ends to be taken into account, and therefore no relevant conversion rates from freedom into those particular individual ends.
(ii) Opposition to a purely content-independent account of freedom's value
A second possible source of the capability theorist's resistance to the thesis of the content-independent value of freedom is the thought that it would be a mistake to completely divorce the value of freedom from the nature of its content. This thought may be correct, but it does not contradict the assumption that freedom has enough content-independent value to eliminate the paternalist possibilities outlined above. In other words, in order to complete the anti-paternalist move, it is not necessary to eliminate all reference to specific capabilities and conceive of the currency of distributive justice exclusively in terms of freedom as such; rather, the anti-paternalist move can be completed either by rejecting all reference to specific capabilities, or by combining an interest in specific capabilities with an interest in freedom as such. Let us call these two theoretical alternatives, respectively, the purely content-independent account of freedom's value and the mixed account of freedom's value.
For those who adopt the purely content-independent account, so eliminating all reference to specific capabilities, the only kind of individual good that is politically relevant is freedom, conceived and measured in an evaluatively neutral way.Footnote 16 Such a stance implies a move in the direction of resourcism,Footnote 17 and reflects an interpretation of liberalism that Isaiah Berlin seemed to have had in mind when he claimed that ‘[m]ost modern liberals, at their most consistent, want a situation in which as many individuals as possible can realize as many of their ends as possible, without assessment of the value of those ends as such, save in so far as they may frustrate the purposes of others’ (Berlin Reference Berlin2002: 199, my emphasis). Here, the completion of the anti-paternalist move is at its most obvious.
A more moderate way to complete the anti-paternalist move, however, would be by adopting a particular kind of mixed account of freedom's value. On this particular kind of mixed account, while our interest in the content-independent value of freedom leads us to favour freedom as such (rather than specific freedoms to function well) as the relevant good to which individuals have claims, we do not necessarily consider two people with equal amounts of neutral freedom to be equally well off (in the politically relevant sense), for we nevertheless consider ‘being well off’ (in the politically relevant sense) to be a function both of neutral freedom and of the presence of specific capabilities. Decisions about how to distribute neutral freedom should therefore be partly determined by a concern with specific capabilities to function well.
To illustrate, consider how the mixed account would deal with the case of Arthur, introduced at the end of section 3.3. The mixed account states that, given our recognition of the non-specific value of freedom, and in the absence of any non-paternalist reasons for favouring benefits in kind over benefits in cash, Arthur should receive his compensation in cash. But it also states that our concern for the capability to achieve basic bodily movements explains why we give the cash to Arthur in particular rather than – say – to Betty, a person with normally functioning legs who, owing to other kinds of personal limitation not suffered by Arthur, lacks the skills necessary to drive a sports car at very high speeds, or the strength and weight necessary to be a successful boxer, or the psychological ability to engage in various sorts of anti-social behaviour that liberals would normally tolerate. If we aim to compensate people for specific capability deficits, treating people equally may involve departures from an equal distribution of neutral freedom. If Arthur and Betty are identical in all respects except that Arthur (unlike Betty) has a capability deficit in terms of the ability to walk and Betty (unlike Arthur) lacks the ability to engage in some particular form of non-functional or dysfunctional behaviour, then we should give greater neutral freedom to Arthur than to Betty as a means of compensating Arthur's deficit in terms of one of the specific capabilities on the list.
The difference between this mixed account and the purely content-independent account represented by the Berlin quote is not one of degrees of paternalism, for both alternatives avoid treating any individuals paternalistically. Rather, the difference lies in the degrees to which the two accounts are neutral between different individuals’ conceptions of the good. But that is another story.Footnote 18 I have been concerned, in the present context, not with the distinction between perfectionism and neutrality but with that between paternalism and anti-paternalism. From the point of view of some liberals, neutrality might represent a reason for moving beyond the mixed account of freedom's value and embracing the purely content-independent account. In the present context, however, the relevant question is whether the capability theorist has any good reason for resisting moving beyond the third interpretation of her normative theory and embracing the mixed account of freedom's value. Neutrality does not constitute such a reason, from the point of view of the capability theorist, for any perfectionism present in the mixed account is similarly present in the capability approach on any of the three interpretations considered earlier.
5. CONCLUSION
The role of the value of freedom in the capability approach has turned out to be less effective than one might have supposed in placing limits on the scope for paternalism. Capability theorists who consider themselves liberals can respond to this fact in one of two ways. First, they might choose to hold fast to one of the three interpretations of the capability approach set out above. In this case, they should devote more time either to justifying the kinds of paternalist intervention that are permitted by the capability approach, or to characterizing and explaining the ways in which the capability approach is subject to the normative constraints of an independent liberal theory of the right (as opposed to a theory of the good that incorporates the value of freedom). Alternatively, capability theorists might choose to move beyond the capability approach as it is normally understood, and recognize the value of capability as such. I have suggested that, to the extent that capability theorists have good reasons for embracing the third interpretation of the capability approach, they lack good reasons for rejecting this additional move.
