Introduction

Deacon (2007) has recently proposed that Kittay (1999) offers Social and Public Policy a new theoretical approach. By arguing that Rawlsian liberalism fails to capture the density of human relationships and interdependent obligations, her work has influenced his attempts to combine an ethic of care with welfare conditionality (Deacon 2005). I wonder though whether Deacon has paid enough attention to the deeper implications of Kittay’s intervention.

Nussbaum (2006, pp. 217–219), for instance, while agreeing with much that Kittay (2006, pp. 559–563) says, is nevertheless concerned that she is too dismissive of liberalism. There is indeed a branch of care ethics which rejects liberal conceptions of justice and autonomy as inherently masculinist (Fitzpatrick 2003, pp. 116–118). Nussbaum’s preference is therefore for a social philosophy that will encompass care and offer a robust critique of liberalism but without losing what is indispensable within the latter. To what extent should we adopt Nussbaum’s approach? Should we accept her critiques of Rawlsian liberalism, for instance? Would this provide social and public policy with richer theoretical foundations?

This article will explore Nussbaum’s recent attempt to theorise care, and associated ideas such as dignity and interdependency, while preserving and renewing what she sees as the best elements of the liberal tradition. I propose that various limitations in Nussbaum’s position suggest that we should not rush to ditch Rawlsian liberalism or its moral philosophy, contractualism. I close by speculating that something closer to Scanlon’s philosophical ethics is where social and public policy theorists should look when framing concepts of justice, autonomy and care.

Nussbaum’s Critique

Nussbaum (1986) is arguably the most brilliant of those aiming to relocate liberalism upon Aristotelian foundations, her objective being to preserve what is best about liberalism while detaching it from those premises, concepts and methods which she regards as outdated. It is valuable, she says, in its commitment to individuals’ freedom but less so when it conceives of freedom in atomistic, economistic and anthropomorphic terms. For Nussbaum these flaws are articulated by, and attributable to, the social contract tradition of thought, or the idea that society is a product of individuals leaving a state of nature to enter cooperative systems designed to render the mutual advantage of all. At its crudest, contractualism interprets humans as ‘bargainers’ who enter into contracts with others for instrumentalist gains. This is condemned by many as offering too threadbare an account of the dense webs of social and emotional belonging within which humans are entwined.

Frontiers of Justice is the culmination of Nussbaum’s engagement with Rawls. As the person who renovated and revived contractualism Nussbaum depicts him as a reminder of all that is valuable within it and liberalism, and so a necessary corrective to those Aristotelians who might revoke both (MacIntyre 1990), but also someone who adopts many of its worst features. Nussbaum criticises Rawls, and through him contractualism per se, for three principal reasons.

Firstly, Nussbaum alleges that Rawls operates with too narrow a conception of what it means to be human. Derived from Hume’s notion that questions of justice only occur between certain ‘extremes’, the inhabitants of the original position are assumed by Rawls to occupy a ‘normal range’ (Hume 1998, pp. 86–89; Rawls 1972, pp. 126–130). What this does, she claims, is to exclude from view those who do not lie within that range. If we are assumed to be self-interested ‘contractors’ then those who cannot (or are thought unable to) contract with others are excluded from the original position and its considerations of justice. Nussbaum objects in particular to what this means for mentally disabled people. Where Rawls proposed that their needs and interests can only be addressed after society’s basic social and political structures are in place, Nussbaum contends that this would exclude them from considerations of justice and that, instead, they must be at the centre of such considerations because any principles of justice must be sensitive to differences and diversities (Rawls 1993, p. 20; Nussbaum 2006, pp. 108–154).

Secondly, contractualism overestimates the importance of reason, she argues. Although Rawls (2000, pp. 283–290) rejects Kant’s (1996, pp. 215–222) metaphysics he nevertheless adopts a similar notion of reason as abstract and disembodied. The original position plays a role in Rawls’ philosophy not dissimilar to that played by the noumena in Kant’s; in each case moral and political reasoning is assumed to proceed from a space that transcends natural and social contingencies and heteronomies. The veil of ignorance strips us of our particularities, of the characteristics through which we habitually make judgements about the world and form relationships with others. Nussbaum (2001, pp. 242–243, 264–270) alleges that Rawls is here being perfectly consistent with the contractualist tradition in its tendency to downgrade the role which emotions, feelings, benevolence, love and sentiments play in human affairs, thus misunderstanding much about human motivation and behaviour.

Finally, to treat mutual advantage as the principal purpose of social cooperation, to even treat society as a cooperative system, is to neglect the ex ante status of empathy and belonging (Nussbaum 2006, pp. 57–62). In short, where contractualists believe that we arrive at principles of justice through cooperation, for Nussbaum we can only be led to cooperate because of an antecedent sense of sympathy and goodness. The instrumentalist ‘proceduralism’ favoured by contractualists detaches us from each other, purporting to show what pre-social atoms have to gain from social bargaining; whereas Nussbaum’s ‘outcome-oriented’ justice weaves philosophical procedures into the social outcomes we already agree are good. What we do for and with others cannot be reduced to what we gain from them.

Nussbaum (2003, pp. 41–42; 2006, pp. 69–92) proposes that a capabilities approach can retain what is best about contractualism while avoiding the above deficiencies. That approach begins with the dignity of every human, respect for whom means recognising and valuing the plurality of activities through which they come to realise themselves. This means making room for those too often pushed to the margins of ‘normality’, such as disabled people—though Nussbaum does not limit dignity to humans. Wellbeing and quality of life cannot be measured by a single index of resources, she says, but only through a diverse array of capabilities by which resources may be converted into a functioning and flourishing life (Nussbaum and Sen 1993). Nussbaum’s ‘capabilities’ are intended to be universal descriptors that permit cross-cultural variations. Among them are being able to: live a life of normal length; possess bodily health and integrity; cultivate and express imagination and thought; form emotional attachments; form and pursue a conception of the good; interact with and be respected by others; relate to the natural environment; play and enjoy; have some control over one’s political and material circumstances. Benevolence and compassion, rather than abstract rationality and impartiality, are the means by which we reach out to others, the good of whom is already part of my good and not a result of a successfully completed bargain or exchange. Reciprocity is not reducible to contractual advantage.

There are important intersections here with some recent policy and welfare debates (Williams 2002; Lister 2002; Lewis and Giullari 2005). Nussbaum emphasizes that dependency is an inherent human condition rather than a divergence from a norm of independence, so that we should not reify reason and autonomy. Reason is always contextualised by care, emotion and compassion; autonomy is always woven within interdependencies as a relation of connections rather than of separation. Hence care, benevolence, respect and solidarity more accurately capture the fact and value of this interdependency than the contractualist impulse to construct society’s principles and structures from scratch.

It may seem churlish, therefore, to criticise Nussbaum when she is seeking not to reject contractualism per se but effect a rapprochement between its best features and the capabilities alternative she prefers. And yet there are several respects in which she underestimates some of contractualism’s strengths.

Critiquing Nussbaum

There are three key points which should be made.

Right

Firstly, Nussbaum insists that contractualists confuse those who frame the rules of society with those for whom those rules are framed. If contractualism only allows contracts between those who are roughly equal vis-à-vis one another, i.e. lie within the ‘normal range’, then all it can do is treat alleged ‘non-equals’ as objects of charity. For Nussbaum this is to unfairly categorise and exclude from considerations of justice those who are equally deserving of dignity. This is why she objects to Rawls’s exclusion of mentally disabled people from the original position, for instance.

Nussbaum (2006, pp. 135–139) notes two possible solutions. We could (1) conceive of the parties in the original position as trustees for those who cannot fully participate, or (2) abandon the device of the original position. Nussbaum doubts that the first solution is desirable. Not because the idea of ‘trusteeship’ is without merit but because even here contractualism must continue to distinguish between rational/reasonable participants and their dependants, which implies the latter are still not being treated as ‘equal subjects of justice’. In preferring the second solution, then, Nussbaum (2006, pp. 146–151) imagines that at best a contract approach could supplement a ‘rich and comprehensive benevolence’, for only this can provide a substantial, ‘multivalued account of the good’.

A pure rationalism and proceduralism does indeed betray the defects Nussbaum alleges. Without an antecedent idea of what is good, rational processes are unlikely to yield much that is meaningful since if there is nothing ‘outside of’ reason then reason will have nothing to work on other than itself. For Kant (1996, pp. 532–533; Fitzpatrick 2008), the right consists of universal maxims but also of ends and objects through which rational obedience becomes manifest. That said, reason is not dependent upon any particular set of ends and objects, as determination of the latter will always be subject to ‘imperfect’ translations of reason into action. Therefore, rational processes may need the good at some level, though for Kantians like Rawls what is needed is merely a ‘thin’ account of the good, one whose elaboration depends substantially upon rational analysis rather than latching too quickly onto notions of benevolence, instinct and emotion that could lead us astray.

Thus, unless it is grounded in reason, any social consensus risks embodying oppressive characteristics. Nussbaum (2006, p. 158), for instance, believes that state-of-nature thinking ignores the extent to which we already possess shared ends and mutual recognition, but the contractualist observes that without some process of reason-based agreement these may come to embody crude and narrow values. Conservatives, for instance, will want to root ends and recognitions in traditional sources of authority, an unreflective adherence that rational universalists challenge. If our ‘shared ends’ make room for such conservative values then they may be too diffuse to be meaningful; if they do not, then associating ends so closely with what is shared, rather than what ought rationally to be shared, loads too much onto an antecedent sense of belonging.Footnote 1 Nussbaum’s capabilities are presented as universals, but unless anchored in more impartial methods of justification their content and their political implications may remain hollow and indeterminate (Scanlon 1993, pp. 190–191).

Rawls’ project, arguably, was to incorporate such thin accounts of the good into contractualism. He acknowledged that the good should not be neglected, if only because philosophers cannot separate themselves from their own societies (Rawls 1993, pp. 174–176; 2001, pp. 153–154). Yet it is precisely in recognition of this fact that many contractualists have sought not to ignore ‘the good’ but to annex it to ‘the right’, i.e. to impartial methods of rational justification. We may be embedded in our social environments but the goal of contractualists is to theorise the critical and imaginative leaps that must be made if those environments are to be subjected to balanced, analytical scrutiny. Rawls’ notion of ‘reflective equilibrium’ is meant to capture this to-and-fro of principles and considered judgements via a process of critical reason. The right may require at least a thin account of the good to get going, but any substantive account of the good changes as our developing, reason-based judgements of it change. That we do not reason in a vacuum is a long way from asserting that we must leapfrog into dense accounts of the good. Yet Nussbaum (2006, pp. 162–164) jumps over this ‘middle way’ too quickly by treating the right and the good as ‘thoroughly intertwined’.

For example, she recognises what, for political liberals, constitutes this thin account—‘respect for persons’—and is happy to associate her capabilities approach with it (Nussbaum 2004, pp. 59–63, 328–330, 339, 341–345). However, having rejected the comprehensive liberalism of Mill, Nussbaum then proceeds to associate this political conception with what might be called a comprehensive view of the person, one that speaks of ‘tensions and difficulties inherent to humanity’ (my italics). She here defines the ‘inner aspect’ of capabilities as being ‘prepared to engage in the form of functioning in question’ through education, healthcare and emotional support. This is an ambivalent expression. ‘Prepared’ can imply the capacity to engage but it can also refer to motivations and a willingness to function (‘being prepared to…’). Equipping people with the former is consistent with a political conception of the person; equipping them with the latter may not be, as Nussbaum had earlier seemed to acknowledge. It is therefore not clear why Mill should be criticised for being ‘comprehensive’ when Nussbaum’s account of agency sets off eagerly down the same road. Humans may be as material, mortal and needy as Nussbaum alleges; yet if we allow such notions to ‘thicken’ (or to ‘essentialise’) then we may end up imposing upon them an inherent meaning, instead of respecting individuals’ rights to work out the meaning, significance and implications of such properties for themselves.

To non-philosophers the differences may look pedantic. Nussbaum wishes to start from benevolence and proceed to reason while Rawls wants to start with reason and proceed to benevolence. But there are crucial differences here. Nussbaum’s notion of agency invokes the benevolent/caring agent—the agent whose needs lead her to recognise the needs of others. Yet benevolence, care and need cannot be indiscriminate: I am bound to care for my mother more than a pet, or for people today compared to those 1,000 years from now. If rationality can sometimes be unsentimental, sentiment is either unhelpful or simply conservative without some form of orientation to the ideals of impartiality and objectivity. For instance, Nussbaum denotes Dworkin’s legal philosophy as one that treats reasons as separate from emotions and passions; but in fact having allied emotions to moral judgements Dworkin’s (1977, p. 250) point is simply that the latter offers a means of justification for the former, and not vice versa (Nussbaum 2004, pp. 5, 352 no. 14). Debate filtered largely, if not entirely, through an account of the good involves asking what benevolent people agree to, which risks leaving the determination of ‘benevolent’ entirely to the contexts to which those people belong. Nussbaum risks equating what is (actually) preferred to what is (rationally) preferable.

I am not suggesting that we go as far as Feinberg (1986, pp. 58–62) for whom no benevolently-motivated, paternalistic intervention is warranted (unless the aim is to judge whether a person’s choice is truly voluntary). This is due to well documented problems with the weight he gives to voluntariness (Brock 1988). But even if we prefer a ‘balancing strategy’, as some critics of Feinberg have argued, such that some genuinely voluntary actions should be prevented, this balance may nevertheless be tipped in favour of personal sovereignty (Shafer-Landau 2005; Malm 2005). This implies that we treat self-determination as being lexically prior to any intervention. You are permitted to intervene when my actions contradict my conception of what is good for me—with the caveat just noted—but not when you evoke abstract notions of ‘the good’. Contra Feinberg, this is not to subordinate self-determination to a person’s own good but merely to claim that the latter can sometimes outweigh the former; thus, Feinberg (1986, pp. 71–81) might permit you to sell yourself into slavery if and only if such slavery is non-exploitative, whereas I would not permit this even where slavery took such a benign form. Contra Nussbaum, it would imply that what we care for, before we care for anything else, is the individual’s rights and powers of self-determination.

All of this suggests we should look again at option (1) where the parties in the original position are trustees for those who cannot fully participate. We cannot value others without also asking difficult questions about the distance which separates us from them. We may formally acknowledge the needs and rights of someone living 1,000 years from now, but since we cannot communicate with them we can only make rational estimates as trustees of their future welfare (Fitzpatrick 2003, pp. 131–137). And if this is true of future generations then presumably it is also true of other species and, yes, of some human beings, e.g. foetuses and infants, those in a coma or persistent vegetative state (PVS), and some categories of mentally disabled people. Parties in a contractualist situation can strive for a greater understanding of differences and dependencies without imagining that all subjects can or should be equal as deliberative subjects of justice. (I return to this point below.) Nussbaum’s stress upon benevolence and care might serve to reduce the distance between agents, but it cannot eliminate distance itself.

Wellbeing

Secondly, Nussbaum (2006, pp. 165–168) acknowledges Rawls’ intention is to provide a scale of well-being (primary goods, including income and wealth) across which comparative measurements can be made and redistributions take place. As laudable as that is, though, Nussbaum complains that it distorts the wider range of factors that account for wellbeing. We could try expanding Rawls’ list of primary goods, of course, but for Nussbaum this is pointless unless we recognise that wellbeing is ultimately a complex matrix of interactions and not a one-dimensional index. The basic argument, of course, derives from Sen (1992, pp. 109–112) and captures the notion that income/wealth are imperfect proxies for wellbeing because different people will be able to translate those resources into different levels of well-being dependent upon their diverse opportunities and capacities. This is an increasingly popular idea within Social and Public Policy.

I have indicated elsewhere, however, that there are difficulties with ‘over-pluralizing’ wellbeing (Fitzpatrick 2003, pp. 42–46). There is value in drawing attention to social contexts, unless and until doing so displaces the central role that material goods, and the relational distributions of material goods, play in determining social inequalities and individuals’ opportunities. This is partly because we can never entirely accommodate the full range of social factors that contribute to welfare into our matrix of wellbeing. It may be that, all other things being equal, my day is ruined compared to yours because I slipped in the bath this morning; does this mean our matrix should allow for bathroom hazards (or the avoidance thereof)? How far can and should we complicate definitions and measurements of well-being? We surely have to stop somewhere if we are not to over-complicate things beyond usefulness. Nussbaum (2006, pp. 190–191) arguably admits as much when she grants the value of applying a single list of entitlements to people with impairments.

Sen and Nussbaum would probably reply that this diversity is precisely why plural measures go further in capturing wellbeing than singular measures; why addressing capabilities is superior to an exclusive focus on material resources and primary goods. This is persuasive to some extent. We ought to resist focusing exclusively on the material and, indeed, such warnings have long emanated from ethical egalitarians in the idealist tradition. Yet one of the grandfathers of ethical idealism, Tawney (2004, pp. 33–35, 84–87), observed repeatedly that wealth and property constitute the foundations of self-esteem and that a ‘functional society’ would be dependent upon their fair distribution. Given that Rawls’ list of primary goods includes not only income and wealth but more ‘social’ relations like opportunities, membership and esteem it surely makes sense to regard him as lying within this tradition too.

Yet the capabilities approach risks treating Rawls as if he thinks well-being can be measured, and society organised, mechanically (Nussbaum 2003, pp. 50–51, 53):

Sen and I both argue that Rawls’s theory would be better able to give an account of the relevant social equalities and inequalities if the list of primary goods were formulated as a list of capabilities rather than as a list of things.

A woman may be as well off as her husband in terms of income and wealth, and yet unable to function well in the workplace, because of burdens of caregiving at home.

But there is no evidence that Rawls conceives of income and wealth merely as ‘things’; given that his theory of justice is concerned with socioeconomic relations of power and opportunity it makes more sense to interpret income and wealth as such. Whether giving more income/wealth to x produces less injustice depends upon the overall patterns of resource distribution. The woman’s plight vis-à-vis her husband is partly due to the inadequate resourcing of childcare in the labour market and workplace, i.e. she is not “as well off” as her husband after all. Exclusive focus on material goods like income and wealth cannot offer comprehensive accounts of justice, freedom and wellbeing (a woman with less income/wealth but with a partner who shares caring responsibilities may be better off in crucial respects than Nussbaum’s woman). Yet, in characterising them merely as ‘things’ Nussbaum ignores the extent to which Rawls acknowledge this, as well as the point made by Tawney: that while it depends upon an ethical philosophy, an equal society must also be founded upon the just distribution of distinct, material goods like wealth and property. In order to represent capabilities as multifaceted and dynamic, Nussbaum characterises Rawls’ primary goods as more static and ‘simple’ than they really are.Footnote 2

For what the capabilities approach risks displacing is the centrality within Rawls’ theory of undeserved circumstances, of the material conditions which shape socioeconomic structures and over which individuals qua individuals possess limited control. Nowhere in Frontiers of Justice can I find reference to capital or capitalism, for instance. Class is mentioned but without a grounding in an ideological critique which establishes possible lines of socioeconomic cause and consequence. There are obviously big questions here of how we should continue to theorise class and capital, given that injustices and oppressions cascade back and forth along multiple paths, yet simple invocations of human rights and dignity sound directionless without at least broad answers to them (Feldman and Gellert 2006). For instance, Nussbaum (2006, pp. 315–324) criticises multinationals for exploiting their workers but it is not clear how far she imagines reforms of the socioeconomic system must go for this to be rectified—as opposed to merely encouraging multinationals to recognise human dignity. True, Rawls hardly indulged in ideological prescriptions for socioeconomic reform but he did indicate that capitalism was not necessarily compatible with justice as fairness and supported a property-owning democracy; and while he emphasised minimum thresholds the difference principle also directs our attention, more than Nussbaum’s (2006, pp. 293–295) approach, beyond the social minimum to the importance of inequalities in material distribution (Rawls 1972, pp. 270–274; Parfit 2001). So if income/wealth cannot be a full proxy for wellbeing, the latter may still need ‘approximates’ for which the redistribution and democratic control of income/wealth remains a leading candidate.

Trustees

Finally, Nussbaum (2006, pp. 118–119) complains that contractualism invites the parties into the contracting situation based upon their productive capacities, such that social cooperation is interpreted by contractualists primarily as a matter of making productive contributions. But if, she proposes, we define people in terms of their right to be treated with dignity then such frontloading of productive criteria seems limiting. If there are many humans who should be valued regardless of the measurable contributions they make then how much more so of animals, for instance.

Nussbaum is correct to identify the habitual links many contractualists make between ‘contract’ and ‘productivism’, with all that this implies for subsequent conceptions of citizenship. Rawls (2001, p. 179) offers an eight-hour working week as a means of distinguishing between those who are legitimately part of the ‘least well-off’ and those who are not. If, indeed, contractualism cannot make room for activities, social relations, lives and lifestyles that are not necessarily quantifiable along a productivist spectrum then that may well be a decisive criticism for the reasons Nussbaum cites. One solution is to effect a partial quantification of informal activities which are contributive but not inherently productivist (such as carework), valuing them in terms analogous to the formal economy without entirely losing their informality (White 2008). Another is to give much greater value to the non-quantifiable and the non-productivist, though at the risk of also valuing the non-contributive (in the form of the free-rider) (Fitzpatrick 2005). Either way, perhaps contractualism does not necessarily force us to equate ‘contributive’ with ‘productive’, it being possible to imagine cooperative arrangements where making a productive contribution is less central to social membership.

Yet for Nussbaum (2006, pp. 333–335, 349–352) this would still miss the point. Contributions, agreement and cooperation simply do not relate, she observes, to the lives of those with severe mental disabilities, animals or future generations. Contractualism, she would say, confuses ‘contract’ with ‘justice’ and so, by treating anything outside the contract as an object of charity, makes too stark a distinction between charity and justice.

But Nussbaum, it can be claimed, merely inverts the contractualist logic she identifies. It is indeed difficult to imagine what rational agreements with people experiencing severe mental disabilities, with animals or with future generations could resemble. As such, Rawls excludes them from the contracting situation while, in order to include them, Nussbaum would abandon the contracting situation. Yet is it possible to retain the contract without the severe division between justice and charity that Nussbaum alleges contractualism embodies? It is indeed simplistic to interpret beings as either equal subjects of justice or as objects of charity. But why should this rule out reference to cooperative systems? Is it perhaps possible to be in the cooperative system but not of it? This would mean making individuals with severe mental disabilities, foetuses, coma/PVS patients, animals and future generations occupants of a system of just cooperation without the pretence that they can shape it except through our trusteeship. That trusteeship would make them partners but partners who are nevertheless dependent upon the privileged positions we occupy in relation to them. Since asymmetries of power towards such groups are unavoidable the best we can do is to combine that fact with a normative stance which is appropriately sensitive. Nussbaum (2006, pp. 361–362) herself effects this equilibrium when acknowledging sentience as the basic threshold for membership within the community of justice.

So, the capabilities-oriented question ‘how should we treat those groups with dignity?’ is not dissimilar to the contractualist question ‘how would we want to be treated if we belonged to those groups?’ The difference is that the latter arguably allows the imaginative leap to be made at a distance critical and objective enough to balance conflicting needs and interests, rather than through the discretions of emotion and sentiment.Footnote 3 We may be moral equals with all or some of those groups’ members, but we are not and can never be deliberative equals and it would betray the special responsibilities we have towards them to propose otherwise. Imagining society as a cooperative scheme does not, then, commit us to proposing that every member must always agree to each and every decision made; it merely enjoins us to have deliberative institutions and reflective processes through which all of the relevant parties can be actually or potentially engaged.

In sum, Nussbaum (1) rushes too quickly past contractualism’s thin accounts of the good, (2) over-pluralises the notion of wellbeing, and (3) underestimates the value of trusteeship. These three criticisms are not intended to challenge Nussbaum’s basic project, as indicated earlier; yet, at the centre of contemporary moral and political philosophy lies the question of how to balance reason and care when theorising about justice. Should we tip the balance in favour of moving mainly from reason to care (contractualism) or vice versa (capabilities)? A nuanced distinction perhaps—but because they represent fundamental philosophical divisions, e.g. between Kantianism and Aristotelianism, we cannot simply split the differences by trying to weigh the scales equally. If the above counterarguments to Nussbaum hold, but her philosophical orientation is nonetheless worth accommodating, how should we proceed?

Scanlon

Can we preserve the best parts of Nussbaum’s critique (her rejection of extra-social bargaining and instrumentalist advantage, of excessive rationalism, her promotion of dignity and benevolence) within a reformed contractualist framework? The most influential post-Rawlsian contractualist, Thomas Scanlon, offers a possible means of doing so. This is because, in bypassing Rawls’ original position, he makes less of an appeal to self-interest and integrates rationality and reasonableness to an extent arguably not available through Rawls’ reflective equilibrium. He therefore anticipates the above critiques without abandoning contractualism. What follows is not meant to be a complete or systematic account of how we might combine contractualism and capabilities. Instead, its intent is to offer the beginnings of a sketch that can be filled out later.

Scanlon (1998, p. 4; Matravers 2003) regards judgements of right and wrong as deriving from principles that ‘could not reasonably be rejected’ by others who are similarly motivated. Those principles that could reasonably be rejected are those that would cause the person serious hardship and where there are alternative principles that would not impose comparable burdens on others (Scanlon 1998, p. 196).Footnote 4 Immediately, therefore, we are no longer dealing with the mutual advantage of atomised, self-interested bargainers, but with a shared understanding of reasonableness that makes people social and moral neighbours before issues of advantage and bargaining arise.

Nevertheless, Scanlon (1998, p. 196) is a contractualist because he is concerned with principles formulated from rational, hypothetical agreement among agents.Footnote 5 It is trivially true, says Scanlon (1998, pp. 37–55; Hume 1969, pp. 460–461), that people have desires and urges to act in certain ways, but this does not thereby mean that reasons are the slaves of the passions for only reasons can supply motives for any actions which follow. Motives for performing action x or y do not come from deciding which one is associated with the strongest desire, but because a ‘framework of maxims’ that gives either x or y greater meaning depends on the projects and strategies around which we have decided to shape our lives. In simpler terms, reason enables desires to be ranked, contextualised and either expressed or restrained. Wellbeing therefore comes through success in achieving one’s rational aims, and not merely through the fulfilment of informed desires; though well-being is itself of secondary consideration since in determining our goals and plans it is not irrational to choose life p which involves less well-being than life q (Scanlon 1998, pp. 123–133). It is values (right and wrong, justice) that are important rather than weighing one type of life against another in respect of their net wellbeing.

There is a potential circularity in Scanlon’s argument if by ‘reasonable’ we denote those people who agree with certain principles and if by ‘principle’ we mean that with which reasonable people would agree. Additionally, what if (1) others are not similarly motivated, or (2) they are, but the meaning of ‘reasonableness’ is too contested to be describable as ‘shared’? Scanlon’s (1998, Ch. 8) response involves a concentric model of moral justifiability. The model’s central domain contains judgements of right and wrong that pertain everywhere and so are morally universal. The second domain is similar but relates judgements more to social conditions, conditions that vary in their universality such that some principles will encompass all societies and others will be more culturally-specific. This domain seems to make room for the kind of context-sensitive judgements that give particularistic dimensions to the moral universalism of the central one (and see no. 49). The final layer articulates the extent to which morality can and should embrace a plurality of values and conduct; it is therefore more of a democratic, deliberative space that drives and articulates changes in values and attitudes of the social contexts.

The actual motivation of people—point (1)—matters less than a moral universalism, the baseline set of reasonable principles to which they should agree according to impartial methods of justification. In response to point (2) Scanlon acknowledges the value of contestability since the second and particularly the third domains are more indeterminate and thus open-ended than the central domain. Scanlon is thus not necessarily assuming that the meaning of ‘reasonableness’ and ‘sharing’ will be immediately obvious or invariable. So, although there is something of a circularity in Scanlon’s (1998, pp. 356–360) philosophy his concentric model, by encompassing the universal and the contextual, also suggests that ‘reasonable’ is far from being the rather conservative signifier, assuming a simple social consensus, that it may at first appear to be.

To summarise, Scanlon seems to embrace several regulative ideals. The first involves ‘impartial, objective justification’, based upon what people could not reasonably reject, but one that is not necessarily impersonal. This is because decisions about reasonable rejection ultimately take the form of dialogues between the relevant agents themselves. This idea has been developed recently in Darwall’s (2006a, b) ‘second-person standpoint’ which derives moral obligation from interpersonal claims of respect that can be weighed against one another. The second is that of reason-directing action where (contra Utilitarians) we do not treat desires as paramount and (contra Aristotelians) we retain a critical distance between reason and action, and between the self and its ends (MacIntyre 1982, pp. 214–220). Thirdly, reason is itself contextualised by reasonableness but the latter stretches across the three domains of a concentric model that encompasses both universalism and contextualism. By being interpersonal, reason-directed and contextually universal, Scanlon’s contractualism does not attempt to provide once-and-for-all foundations to moral philosophy.

What we have here, then, is a large degree of concordance between Scanlon and Nussbaum (2006, pp. 67–68). Scanlon, too, is hostile to the idea that moral principles are constructed either through self-interest or, in the case of Rawls, by individuals imaging what would best serve their self-interest once the veil of ignorance is removed. And while he defends rationality against a Humean account of motivation, his notion of reasonableness encompasses both what it means to be human, i.e. to be worthy of respect and dignity, and social contexts.

But Scanlon’s is still a contractualist liberalism for three reasons. (Most of what follows has been aired above already and so I will be brief—though this is less true of the third point which requires a lengthier exposition if we are to be clear about Scanlon’s specific contribution.)

Firstly, he accommodates a thin theory of the good in that rationality is dependent upon some prior value. For Scanlon this value is one of reasonableness and so respect for those in whom we recognise a shared capacity for reasonableness. This is a political value in that it offers a basis and procedure for agreement. A ‘thick’ conception of the good is avoided because the content of agreement is always open-ended, consensual and deliberative. But nor does this political conception collapse into mere formalism because the concentric model allows that value to encompass a moral universalism, social contextualism and open-ended pluralism (the three respective domains). So the value of reasonableness can still be subjected to impartial, objective (but not necessarily impersonal) methods of justification that bring ‘the right’ firmly back into the picture; and the concentric model echoes Rawls’ method of reflective equilibrium, while acknowledging that qualities such as reasonableness will never be finally determinate (Schroeter 2004). By comparison, Nussbaum’s comprehensive account of the person, of benevolence and her list of capabilities risks overestimating our pre-rational frames of reference and so the attendant basis for social consensus also.

Secondly, Scanlon makes room for trusteeship. Trusteeship, he says, may not be appropriate in the case of non-human animalsFootnote 6 but it is for those who cannot develop, or have not yet developed, the capacity to make judgements, including future generations, foetuses and infants, coma/PVS patients, and those with severe mental disabilities. The simple fact of their being human means they are accorded the same status as all other humans; but their limited capacities means that in deciding about principles of moral justification we must consider what they could not reasonably reject on their behalf (Scalon 1998, pp. 179–187). They are perhaps equal subjects of justice in that they are equally deserving of respect (though this is less clear in the case of distant future generations and some classes of non-human animals), but they are not deliberative equals. It is important to remember that no practical differences are likely to follow from these differences. A Scanlon-run society would not treat severely disabled people, etc. with any more or less consideration than a Nussbaum-run one. But, as noted above, the real differences between a broadly Kantian philosophy and a broadly Aristotelian one do present us with real choices when it comes to many ethical and political questions.

Finally, Scanlon (2003b, pp. 210–211, 217) does not treat wellbeing as paramount and draws attention to systematic disadvantages.Footnote 7 What a great deal of post-Rawlsian, egalitarian liberalism has done is to debate how to weigh choice against circumstance (conditions over which individuals have little or no control, such as accidents of birth, coercive social structures or sheer luck) (Knight 2005; Barry 2006). Too far in one direction and we overestimate the scope of individuals’ freedom; too far in the other and we underestimate the powers and responsibilities of agents. There are those who object to this debate, seeing it as undermining notions of moral equality (Anderson 1999). Yet if we avoid it, if we opt for something closer to the capabilities approach which simply foregrounds dignity, then we risk losing a distinction crucial to the moral assessments we make of others and the political assessments we make of societies. We perhaps respect people more when we acknowledge their status as actors against a background of constraints. Can we therefore presume in favour of human dignity while being sensitive to the choice/circumstance categories?

Scanlon (1998, pp. 251–266) offers two key distinctions regarding responsibility. The first is between a ‘Value of Choice account’ (VoCa) and a ‘Forfeiture View’. In the VoCa what matters is the value of the opportunity to choose which is offered by the conditions within which an individual is placed. If those conditions are sufficiently good then the agent can have no valid cause for complaint. The Forfeiture View is concerned more with the outcomes of actions resulting from a conscious decision that the agent could have chosen not to make. So where the VoCa directs attention to background conditions the Forfeiture View deals with the fact of, and thus personal responsibility for, choice. If Andrew and Lester choose to eat and drink heavily then there is an extent to which they forfeit the right to complain when their health suffers. But imagine that Andrew grew up in a deprived working-class home and Lester in an affluent, middle-class one. Since people in deprived circumstances experience stress and since people often cope with stress by indulging in unhealthy activities then, because he is the victim of unjust background conditions for which others bear responsibility, we should judge Andrew less harshly than Lester if their eating and drinking affects their health adversely. The VoCa, the quality of the opportunities available, revises our assessments of personal responsibility (the Forfeiture View).Footnote 8

It is common for people to emphasise the Forfeiture View as there are few circumstances where absolutely no freedom of choice exists. Yet Scanlon insists that we need to readjust our focus to appreciate not the fact of choice but the quality of the alternatives on offer. A bank-teller with a gun to his head could decide not to hand over the cash yet none of us would condemn him should he act otherwise.

Yet, such extreme examples aside, we are often reluctant to look beyond the fact of choice, one consequence of which is that we confuse different settings. The Forfeiture View encourages the belief that the costs and benefits of an action should always flow back to the agent who performed it, since the agent could always have chosen otherwise. In policy terms the Forfeiture View can mislead us, then. Because I could always flip burgers instead of claiming benefits then continuing to do the latter must mean I am workshy. This confusion of settings is also apparent when someone objects to taxation, as if ‘their’ earnings are unrelated to luck and circumstance. In short, the Forfeiture View foregrounds as individual matters what are actually questions relating to background social conditions.

The VoCa therefore incorporates a second distinction (Scanlon 1998, pp. 279–294). By ‘attributive responsibility’ is meant those actions which can be attributed to an agent as the basis of moral appraisal; ‘substantive responsibility’ refers to what people are required to do for each other. The bank-teller is responsible for handing over the cash in an attributive sense (he could have refused) but not in a substantive, blameworthy sense. Yet if we are persuaded by that example it is reasonable to apply the distinction to other scenarios: my refusal to flip burgers may involve a more complex set of decisions and events than those involved in the bank-teller’s decision but the basic underlying distinction still applies. I may be open to criticism, says Scanlon, but that does not mean I deserve to lose time and/or income in the form of welfare penalties if I am already bearing the costs of background conditions which are unfair and unjust.

Scanlon’s distinction therefore enables us to see the errors of conservatives in ignoring substantive responsibility, without falling into the mistake of neglecting agents’ attributive responsibility even when circumstances remain socially unjust. What makes his a non-presumptive philosophy of duty is Scanlon’s wariness towards the principle of desert and to the idea of imposing social sanctions on those judged to have acted incorrectly. He thus embraces an ethic of dignity concerned to secure just social conditions, the goods of which cannot be completely withdrawn whenever someone acts foolishly or immorally.

A Way Forward for Social and Public Policy?

A reformed, Scanlon-inspired contractualist liberalism means not giving up on the centrality of rational/reasonable agreement when debating social principles and reforms. Yes, we must not fetishize impartial procedures nor treat reason as a disembodied abstraction, but so long as devices like the original position are treated heuristically (as means for critical and analytical reflection) then we can expand their purview to encompass relations of trusteeship for those whom we respect as subjects of justice even though they are not deliberative equals. Though interdependency is a fact worth recognising and a value worth preserving there are those towards whom we inevitably occupy asymmetrical, privileged positions of power. By incorporating an ethic of trusteeship into the original position, or similar device, then those asymmetries can inspire relations of care and responsibility rather than of subordination. Nor need such embeddedness in social interdependencies dislodge the liberal virtues of autonomy and self-determination.

This would mean continuing to attend to the justice (or otherwise) of society’s basic structures in terms of undeserved dis/advantages and therefore to the circumstance/choice distinction. Income and wealth cannot substitute for all the variables relevant to unjust inequalities, but unless we are to displace the socioeconomic as a key axis of social interaction then the redistribution and democratic control of material resources and goods remain effective ‘approximates’ for wellbeing (Scanlon 1993, pp. 197–198). And even when people choose unwisely the VoCa does not presume it is a good thing for that person to suffer harm or loss as punishments for their poor choice. Applying a contractualist philosophy does not necessarily mean viewing society as a political and economic bargain where those who break the contract can be excluded from its benefits. That there are those who prefer not to flip burgers should direct our attention to the justice of social background conditions and to the question of substantive responsibilities. By contrast, welfare systems which prioritise constraints, sanctions and punishments treat claimants as only one step up from criminals. Indeed, Scanlon’s philosophy of duty allied to Nussbaum’s philosophy of dignity highlights the extent to which the contemporary politics of social and welfare obligation confuses attribution, blame and punishment.

As such, for Scanlon something similar to Nussbaum’s dignity does indeed seem to be a baseline for social organisation. We should therefore bypass those who would make mutual advantage central to socio-moral philosophy, and we should not necessarily measure contributions in productivist terms, but we can still conceive of society as a broad cooperative enterprise whose rationale is to invite deliberative participation and the freedom to reflect upon and join a diversity of cooperative schemes. Respect and reasonableness might both contribute to thin accounts of the good, but without some mechanism of impartial, hypothetical agreement among deliberative subjects we risk avoiding political and ideological critiques about social background conditions.

Conclusion

The capabilities approach offers considerable challenges to those traditions which have dominated modern social thought. In this article I have reviewed the criticisms directed towards one of those traditions by Nussbaum and acknowledged there is much we can draw upon in continuing to elaborate alternatives (like an ethic of care) to the market-based, consumerist orthodoxies which unfortunately dominate contemporary thinking. Yet there are also potential problems with the capabilities approach. We should therefore be careful before ditching the Rawlsian school of thought. This is not quite Nussbaum’s intention; even so, as I have argued, she underestimates the resources it has to offer. The purpose of the paper is not to endorse a Rawlsian over a capabilities approach, but to argue that the former cannot be dismissed easily, especially if the deficiencies of the latter are to be addressed. By rooting ourselves in a revised contractualism, perhaps something akin to Scanlon’s moral philosophy, we may have a more fertile ground upon which to build synergies with the capabilities approach. I have not sought a systematic overview of what form that synergy might take, merely the preliminary work of arguing that if we reject contractualist liberalism we may be depriving social and public policy of the resources it needs to incorporate and apply notions of reason, care, dignity, autonomy and justice for a world which needs them more than ever.