1. THE SITE OF EGALITARIAN JUSTICE

Rawls's difference principle states that economic inequalities are acceptable if and only if they serve to maximize the prospects of the worst-off members of society (I shall, at any rate, assume this interpretation of the principle). Inequalities are permitted if the work-incentives they provide to the most productive members of society are sufficient to allow the state (by means of taxes and benefits) to improve the prospects of the worst-off relative to their prospects under a regime of strict equality (Rawls Reference Rawls1999: ch. 2). Moreover, in judging between different unequal distributions, a government should always prefer the distribution in which those with least have as much as possible. Cohen's response to this incentives-based argument is as follows: if we were to move from a system containing such income inequality to another system containing the same total income distributed more equally, the worst-off members of society would be made even better-off than they are under the regime of inequality-generating market incentives (Cohen Reference Cohen2000, Reference Cohen2008).

Now it is of course standardly assumed, and Rawls duly assumes, that there are inequalities that cannot be eliminated without reducing the total amount of distributable income, because the required increase in taxation of higher incomes would result in lower incentive levels for the more productive members of society. Given this fact about human motivation, optimal taxation levels require inequality-generating incentives. But, Cohen says, this is not to deny that if the more productive members were to take Rawls's difference principle seriously – if their sense of justice really told them that, given the moral arbitrariness of talent differentials, justice involved maximizing the gains for those who have least – they would accept a higher taxation rate without reducing their own productivity levels. They would live, in other words, by an ‘egalitarian market ethos’. In failing to condemn the self-interested market behaviour behind the incentives argument, Rawls is being inconsistent.Footnote ²

In Rescuing Justice and Equality, Cohen deepened his critique of Rawls by denying the latter's claim to have described ‘first principles’ of justice (Cohen Reference Cohen2008: ch. 6). According to Cohen, instead of identifying first principles of justice, Rawls identified ‘rules of regulation’ – rules that might be justified instrumentally as ways to approximate the realization of first principles of justice given certain contingent facts about particular human societies. Rules of regulation – like the rule permitting inequalities that benefit the worst-off – are fact-dependent, whereas first principles of justice must be fact-independent. On the basis of our first principles of egalitarian justice, we ought to condemn the selfish tendencies that make the inegalitarian Rawlsian rule necessary. It follows that Rawls has not really described the nature of justice, but a second-best scenario – a kind of compromise with justice that we might make, and indeed often do make, in the light of certain ‘feasibility constraints’, including facts about unjust behaviour on the part of others.

Cohen's critique of Rawls has produced a growing literature, much of it aimed at defending the view that the concept of justice applies specifically to the institutional basic structure of society, understood as a set of publicly recognized, legally enforced rules (see, for example, Williams Reference Williams1998, Reference Williams and Feltham2009; Pogge Reference Pogge2000, Tan Reference Tan2004). I shall not discuss these counterarguments here, beyond noting that Cohen himself remained unconvinced by them and that his critique of the ‘basic structure objection’ still appears to create discomfort in the Rawlsian camp. This is not to say that Cohen's replies to his critics are necessarily decisive. On the other hand, I believe that Rawlsians would do well to consider the alternative argument that I shall put forward in the next three sections, an argument that takes as its starting point the concept of equality – more specifically, that of basic equality.

2. THE PROBLEM OF BASIC EQUALITY

In this section I shall explain why basic equality is an important premise in egalitarian normative theories. In section 3, I shall put forward a first account of basic equality which I have already defended elsewhere (Carter Reference Carter2011). In section 4, I shall build on this first account in order to provide a second possible account of basic equality.

The need to establish basic equality represents a challenge for any theorist of justice who advocates egalitarian principles as a matter of individual entitlement. It therefore represents a challenge for, among others, Cohen and Rawls.Footnote ³ I use the term ‘egalitarian principle’ in a broad sense, taking it to cover any principle that (a) prescribes equal amounts of something, and/or (b) prescribes a set of duties the validity of which is held to derive from a commitment to treating all relevant individuals as equals – for example, a set of duties held to be valid because they would be chosen by people who are equal in terms of moral standing and who follow a certain decision procedure.Footnote ⁴ Obvious examples of principles covered by condition (a) are ‘equality of liberty’, ‘equality of welfare’ and ‘equality of income’. Less obvious examples (because they equalize a more fundamental and less tangible good) are ‘equality of respect’ and ‘equality of concern’. An example of a principle covered by condition (b) is Rawls's difference principle, which, while not prescribing equality of anything, prescribes a distribution of income and wealth that is justified by reference to people's equal moral standing in the social contract. I shall call an ‘egalitarian duty’ any duty to act on an egalitarian principle.

Egalitarian principles (as defined above) presuppose the existence of some morally relevant feature of the relevant set of individuals such that those individuals are entitled to be treated as equals (in terms of that feature). Advocates of such principles therefore need to be able to say what that feature of those individuals is. Once we have specified what makes people equals (in a morally relevant sense), we can then go on to ask what sort of equal treatment they are due – that is, what treating them as equals (in the morally relevant sense) actually involves.

Egalitarian principles (as defined above) are fact-dependent in the following sense: such principles depend for their normative validity on the fact that individuals are equal in some sense, together with a higher-order hypothetical prescription of the form: ‘If individuals are equal in such-and-such a sense, treat them equally in such-and-such a sense’. This hypothetical prescription is the source of the moral relevance of the fact, and itself derives in part from the still-higher-order Aristotelian principle enjoining us to treat equals equally (and unequals unequally) together with normative claims about the relevance of certain kinds of facts to certain ways of treating people. ‘Treat people as equals’ cannot be the bottom line for egalitarians (that is, an ultimate principle in Cohen's sense), in the way, for example, that the norm of utility maximization can be the bottom line for utilitarians (‘This is what's valuable: maximize it!’).

One reason for egalitarians’ neglect of the nature of basic equality might well be the undeniable difficulty of identifying a property that is both possessed equally by all of the relevant individuals and able to ground the kind of treatment favoured by egalitarians. Consider people's physical and mental abilities, their susceptibility to pleasure or pain, their degree of vulnerability, their strength of will, or their capacity to be autonomous or rational or morally virtuous. Each of these natural properties of persons is possessed in different degrees by individuals whom the egalitarian would normally like to include within the ambit of her theory. If possession of one or more of these features were to provide the grounds for people's entitlements to certain goods, it would be appropriate to distribute those goods unequally.

Kantian egalitarians see people's basic equality as depending on each person's power to set ends and to act on norms. As long as this power is understood as an empirical capacity, however, it depends on at least some of the variable empirical capacities just mentioned, and so is similarly possessed unequally. On the other hand, contemporary egalitarians like Rawls and Cohen are unwilling to base our equality on the existence of a transcendental power that somehow prescinds from empirical phenomena (Williams Reference Williams, Laslett and Runciman1962).

Those who defend distributive equality only within state borders (call these theorists ‘state-based egalitarians’) seek to ground egalitarian justice not in a natural property of individuals but in the special relation between the state and the citizen, or in that between individual citizens within a state. Their main concern is with the scope of distributive justice rather than with its site. However, their arguments merit consideration in the present context because the account of basic equality that I shall provide in section 4 will itself be state-based.

Regardless of their degree of success in restricting the scope of egalitarian justice, existing state-based arguments for distributive equality tend to presuppose, rather than to supply, a solution to the problem of the basis of equality. Probably the best known of these arguments, due to Thomas Nagel (Reference Nagel2005), appeals to citizens’ ‘joint authorship’ of the state-enforced norms to which they are subject in a liberal democracy – a joint authorship that gives individual citizens a right to be treated as equals within the confines of the state. This argument assumes, rather than justifying, the equal standing of citizens posited by democratic theory. Citizens are not, from a purely social scientific perspective, equally authors of the state-enforced norms to which they are subject. Some individuals play an important role in shaping the content of those norms, others play a less significant role, and yet others provide no input at all. One might say that citizens are all equally authors in the sense of having equal moral standing, and/or equal legal rights, in the process by which those norms are constructed. But this moral or legal sense of ‘equal authorship’ presupposes a kind of basic equality that has yet to be rendered explicit. As individuals subject to the norms of the state, we certainly have ‘standing to ask why we should accept them’ (Nagel Reference Nagel2005: 129). But why should that standing be equal standing?

A similar argument, put forward by Michael Blake (Reference Blake2001), appeals to the need to justify state coercive power. Blake's argument runs as follows: citizens of the same state are all subject to its coercive power; coercion violates autonomy, and therefore needs to be justified to all the autonomous beings that are subject to it; and a state-enforced distribution will be justifiable to all such autonomous beings only if it has the hypothetical consent of the worst off (Reference Blake2001: 283). Hence the exclusive relevance of Rawls's difference principle to the domestic context. Whether or not Blake succeeds in thus restricting the scope of economic equality, he clearly assumes, rather than justifying, the moral equality of the hypothetically consenting autonomous beings. The need to justify this assumption takes us back to the search for an equally possessed property of autonomous beings. That property is not autonomy itself, nor is it the capacity for autonomy, for both of these properties (if understood as empirical properties) are possessed unequally.

In yet another state-based argument for egalitarian justice, Andrea Sangiovanni (Reference Sangiovanni2007) has appealed to the reciprocity that exists between citizens in terms of the mutual provision of a class of collective goods that can be provided only within the institutional framework of a state. Sangiovanni suggests that the requirement to treat citizens as equals is generated not by the coercive power of the state but by the reciprocity between citizens of the same state. It is difficult, however, to see in what sense reciprocity among citizens itself amounts to, or generates, a morally relevant kind of equality of citizens. Citizens reciprocate in virtue of the fact that each contributes as well as benefitting. They contribute, however, to unequal degrees. In answer to this objection Sangiovanni points out that, without the institutional framework provided through the cooperation and contributions of others, the greater contributions of the more talented and skilled members of society would come to nothing (Reference Sangiovanni2007: 25–26). This is true, but it does not show that all contributors ought, as such, to be treated as equals. Suppose that we decide to buy a house together. The house costs £300 000. I contribute £280 000 and you contribute £20 000. We then rent out the house. Without your contribution we would not have been able to buy the house, and I would therefore not have counted as a contributor to its purchase. Does this fact entail a duty on my part to share the rent equally between us?

Basic equality, then, remains a suppressed and unjustified premise in existing state-based arguments for distributive equality no less than in non-state-based or cosmopolitan arguments. To remedy this lacuna, such arguments might be supplemented with the account that I shall present below, in section 3, of the basic equality deriving from the respect persons owe to each other as such.

3. EQUALITY IN THE EYES OF INDIVIDUALS: OPACITY RESPECT AND THE RAWLSIAN RANGE PROPERTY

We could solve the problem of basic equality if we could identify a morally relevant binary property of individuals, for binary properties, unlike scalar ones (of which empirical human capacities all appear to be examples), are possessed equally by all those who do possess them.

Rawls identified a binary property of persons by introducing the notion of a ‘range property’ (Rawls Reference Rawls1999: §77). A range property is the property of falling within a certain range, where the range in question is specified in terms of possession of some other, scalar property or properties. In Rawls's theory, the morally relevant range property is called ‘moral personality’, and is possessed in virtue of one's having the capacity for a conception of the good and the capacity for a sense of justice. As Rawls admits, these two basic agential capacities, or ‘two moral powers’, are scalar empirical properties of individuals: different individuals possess them in different degrees. Nevertheless, in virtue of the fact of normal persons’ basic agential capacities falling within a certain range (that is, from a given minimum upwards), each individual within that range is in equal possession of the binary property of moral personality. Thus, the basis of our equality is the range property of moral personality.

Range properties certainly appear to be the only human properties that can qualify as both empirical and binary. However, we have yet to show why a range property should be given any moral weight vis-à-vis the subvening scalar properties in virtue of which it is said to exist. Why indeed should we posit the range property at all? Why not concentrate directly on the scalar properties? One answer that we cannot give is ‘Because we want to treat people as equals’, for that answer would deny us the very basis for treatment as equals that the notion of basic equality is supposed to supply.

Elsewhere I have argued that a plausible way to complete the Rawlsian account of basic equality is by thinking of the moral relevance of the range property as emerging in the context of certain kinds of relations between agents – in particular, relations in which a certain kind of respect is the appropriate attitude for one agent to show toward another. It is the appropriateness of this kind of respect that makes it appropriate to focus on the range property.

The relevant kind of relation, in the present context, is that between persons considered simply as persons – that is, the relation in which the appropriateness of a person's deliberation and behaviour depends on the mere fact of the others being moral agents. Within this particular relation, I have suggested, liberal deontology imposes on us a duty of opacity respect (Carter Reference Carter2011).

What is opacity respect, and how does it motivate our focus on the range property of moral personality? There is an intuitive sense in which respecting a person as a moral agent involves adopting a perspective that is external to that person, taking her agential capacities as given and refusing to consider relevant, in one's practical deliberations, certain evaluations of her abilities to make rational decisions, to form coherent plans, to form and act on a sense of justice, and so on. Respect involves maintaining a proper distance.Footnote ⁵ Thus, respecting a person can plausibly be thought to involve treating her as opaque, in the sense of refusing to ‘look inside’ her by analysing and evaluating her various component parts and instead taking into account only the outward appearance she presents to us as an integral and competent agent. Her possessing that outward appearance depends, of course, on her meeting certain minimum standards in terms of basic agential capacities, standards which we tend to take for granted when faced with normal outward human behaviour. We look to see if the minimal standard is met, but the attitude of respect (toward those who do meet that standard) also tells us to ignore, in our practical deliberations, particular levels above the minimum. While perceiving the levels of such capacities (above the minimum) is often unavoidable, we treat persons as opaque by turning a blind eye. To recognize a minimum threshold (in virtue of which someone qualifies as a moral agent) while also turning a blind eye to degrees above that threshold is to recognize possession of a range property.Footnote ⁶ Therefore, if the attitude of opacity respect is an independent moral requirement in relations between persons, it constitutes an independent reason to focus on the Rawlsian range property.

Why should opacity respect be an appropriate attitude to show toward persons considered simply as such? To consider another person simply as a person is to consider her as a being in possession of certain basic agential capacities. Beings in possession of these capacities possess dignity, in the Kantian sense of the term. Part of what is due to them in virtue of this agential dignity is, I have suggested elsewhere, another kind of dignity that is not itself constitutive of agency but can be conferred on agents (or indeed denied them). I have called this second kind of dignity ‘outward dignity’ (Carter Reference Carter2011: §V). Possession of outward dignity indicates a certain invulnerability to assessments of one's agential capacities. In our relations with others considered as bare agents, we ‘clothe’ those others in outward dignity by adopting an external point of view and refusing to expose them to certain kinds of evaluation, in the same way as we would clothe a naked person in order to preserve their outward dignity in a physical sense (refusing to expose them to inappropriate assessments of their physiques). In doing so, we show opacity respect toward those agents. Thus, opacity respect involves a conferral of outward dignity on persons considered as bare agents.

My argument for the basic equality of moral persons therefore runs as follows: moral persons should be regarded as equals insofar as their moral personality is considered a range property; moral personality is appropriately considered a range property in virtue of persons’ opacity in terms of the various capacities upon which moral personality supervenes; it is morally appropriate to regard persons as opaque in virtue of a certain kind of respect – opacity respect – which one person appropriately shows toward another when relating simply as one person to another.

Now, the requirement of opacity respect applies only insofar as one's treatment of a person is appropriately motivated by the simple fact of her being a person. Opacity respect is respect for persons considered simply as persons. This fact serves to circumscribe the moral relevance of basic equality in personal relationships. I am not required to treat other individuals only as equals, even though I am always required to treat them as equals among other things, given that persons always relate to each other as persons among other things.

In order to clarify this point, it will be useful to stipulate a distinction between interpersonal relations, on the one hand, and interpersonal relationships, on the other. Interpersonal relations, as I shall use the term, obtain between properties of persons, and the properties that interest us here are those that persons possess in virtue of their social positions or roles. Examples of such relations are those between father and son, between teacher and pupil, and between employer and employee. Another important relation, in the present context, is that between persons considered simply as such. I shall use the term ‘relationship’, on the other hand, to refer to the complete set of interactions between two or more concrete individuals occurring in light of their various relations. Thus, I might relate to another person as a son to a father, as a student to a teacher, as one citizen to another, and as one person to another. My relationship with that person will then consist in a series of interactions that occur, often simultaneously, in the light of the above relations. The way in which two or more individuals ought to conduct their relationship will depend on the forms of conduct appropriate to their various relations: different relations determine the appropriateness of different attitudes and forms of behaviour. In light of our relation as teacher to student, my deliberations about how to treat you should take account of your nature as a student; in light of our relating as one person to another, my deliberations about how to treat you should take account of your nature simply as a person.

Thus, individuals relate to each other in many ways other than as bare moral agents. In many of these other ways of relating to others, it is appropriate to assess, and act on, the degrees to which those others possess various scalar properties. Moreover, many of these scalar properties are human capacities. For example, teachers commonly make and act on assessments of their pupils’ abilities to follow more or less complex lines of reasoning. Similarly, when employers assign jobs, they typically assess the talents and skills of their prospective employees, and act accordingly. Again, when a prospective employee is considering which jobs to apply for, she may well weigh up her own talents and skills against those of her potential competitors on the labour market.

Cohen himself must admit that such assessments are appropriately made in the context of the labour market. Regardless of whether a person's aim is to maximize or increase her own income or, as Cohen would prefer, to maximize or increase productivity for the benefit of the least talented members of society, and regardless of whether that person is an employer or a prospective employee, that person will need, in the light of the relevant aim, to assess the talents and skills of various workers in order to promote their most efficient use. Indeed, if we assume a moral requirement to act on Cohen's egalitarian market ethos, such assessments will be morally required; otherwise, they will at most be prudentially required, and morally permitted.

This fact of relating both as equals and as unequals need not lead to any practical contradictions, as long as the capacities in terms of which individuals legitimately assess one another (and in virtue of which they relate as unequals) are conceptually distinct from, and do not themselves include, the basic agential capacities upon which the range property of moral personality is taken to supervene (and in virtue of which they relate as equals). Individuals may well be aware of empirical correlations between the possession of certain talents or skills prized by the market – for example, mathematical intelligence, artistic ability or mechanical efficiency – and degrees (above the minimum threshold) of the basic agential capacity to form and pursue a rational and coherent life plan. However, these empirical generalizations are insufficient to identify levels of basic agential capacities in particular cases.

What do individuals owe to each other in virtue of their equality as moral persons? They owe each other a certain set of ‘natural duties’ – duties that are owed to others simply in virtue of their moral personhood.Footnote ⁷ How demanding are these natural duties? On one possible account, the relevant set of duties is quite restricted, including only duties ‘[to help] another when he is in need or jeopardy’, ‘not to harm or injure another’ and ‘not to cause unnecessary suffering’ (Rawls Reference Rawls1999: 98).Footnote ⁸ On a slightly less restrictive account, the natural duties might also include a duty to respect the claim of each other person to a certain set of liberties similar to the set of ‘basic equal liberties’ which, at the institutional level, gains legal protection under Rawls's first principle of justice (see Hart Reference Hart1955).Footnote ⁹ These various natural duties count as egalitarian either inasmuch as they are duties to respect or promote equality of something (such as basic liberties) or inasmuch as their validity derives, in a Rawlsian way, from an agreement between moral persons with equal standing: natural duties ‘obtain between all as equal moral persons’ (Rawls Reference Rawls1999: 99).

Cohen's interpretation of the set of natural egalitarian duties owed by one individual to another is much more expansive than on the above account. We have seen that, for Cohen no less than for Rawls, individuals appropriately interact both as equals and as unequals. For Cohen, nevertheless, justice requires individuals to aim, at least ceteris paribus, for a maximal equal distribution of certain material benefits. Since the egalitarian nature of this individual duty must derive from some sort of basic equality, he presumably sees the duty to equalize such material benefits as deriving from the equality of individuals considered simply as moral persons. If this account of Cohen's view is correct, then we must conclude that Cohen allows individuals’ natural egalitarian duties to take over much of the space that a more liberal theorist would assign to non-egalitarian individual duties and permissions, denying or overriding the moral relevance of a number of relations between unequals in determining such duties and permissions. For Cohen, we must suppose, the duty-generating force of our nature as equals somehow trumps or outweighs much (though not all) of what liberal theorists see as the duty-or-permission-generating force of our nature as unequals.

I shall not attempt here to defend a restrictive account of the natural egalitarian duties generated by the basic equality of persons considered simply as such. Rather, I shall assume a restrictive account and ask whether it can be made consistent with an expansive account of institutional egalitarian duties. A positive answer to the latter question is sufficient to rebut Cohen's charge of inconsistency, even though a fuller defence of the Rawlsian position would certainly require further discussion of the restrictive account.

4. EQUALITY IN THE EYES OF THE STATE: OPACITY RESPECT AND A SECOND RANGE PROPERTY

Let us turn, then, to the egalitarian duties of political institutions. I shall concentrate here on the set of political institutions constituting the state. Relying on the existence of the state in order to ground equality inevitably gives rise to questions of scope. Where should state boundaries be drawn? What are the duties of states toward non-citizens? What are the duties of citizens toward non-citizens? I want to bracket these questions of scope in the present context. The question I am addressing here is a more limited one, and can be formulated as follows. For any historically given state with jurisdiction over a given set of citizens, how, if at all, does the sense in which that state must regard and treat those citizens as equals differ from the sense in which those citizens must regard and treat each other as equals?

From the liberal point of view, the state, no less than the individual, must forbear evaluating persons’ basic agential capacities (above the minimum threshold). However, the liberal state's duties of evaluative forbearance do not end here. There are, indeed, many other kinds of evaluation that liberals tend to view as inappropriate when carried out by the state. For example, liberals consider it inappropriate for the state to carry out evaluations of an individual's degree of attractiveness to the opposite sex, her abilities as a charmer or as a social entrepreneur, and, except to the extent that the state is itself an employer, the talents and skills she brings to the market place: ‘how dare the state pass judgement on its citizens’ worth as workers and lovers!’ (Anderson, Reference Anderson1999: 305).

This difference, from a liberal point of view, between acceptable assessments of individuals by other individuals, on the one hand, and by the state, on the other, can be reformulated in the following way: the state owes citizens a stronger form of opacity respect than that which individuals owe to each other. How are we to explain and justify this difference? We should begin by recalling some standard Rawlsian premises: that society is a system of social cooperation, that justice concerns the proper distribution of the benefits and burdens of that social cooperation, and that the tasks of the state include the coercive enforcement of that distribution. To these premises we should add the ‘fact of pluralism’ and the consequent view that the state should abstain from evaluating citizens’ capacities, conduct or situation in terms of any but the thinnest conception of the good. From these points it follows that the stability of the social contract, by appeal to which the system of social cooperation is regulated under the sway of coercive state power, depends on the ascription to individuals of a particularly robust form of outward dignity vis-à-vis the state. The conferral of legitimacy upon the state by appeal to a social contract depends on the state's conferring upon citizens this robust form of outward dignity – which is to say, it depends on the state's showing the corresponding form of opacity respect for citizens. The state may not assess my utility ‘to society’, nor may it assess the worth of my talents and skills from the point of view of one or more of the various private actors in the labour market. Similarly, although individuals vary in their abilities to understand economic and foreign policy, to vote strategically and to exercise political influence, the idea that their political rights should depend on an assessment of these abilities would be considered an affront to their dignity as democratic citizens (Ottonelli Reference Ottonelli2012). Instead, the parties to the social contract (who enjoy equal moral standing as derived in the previous section) agree that political institutions must be blind to degrees of such talents, skills and political abilities (above a bare minimum). Rawls's veil of ignorance, which rules out knowledge of many factors beyond the basic agential capacities, can be thought of as modelling this sense of respectful evaluative forbearance on the part of a liberal state.

Thus, the liberal ideal of equal citizenship (as opposed to that of equal personhood) can be derived from the liberal ideal of the opacity respect on the part of the state for persons considered as citizens. The liberal state is required to abstain from evaluating not only the degree (above the minimum threshold) of an individual's basic agential capacities, but also the degree (above certain minimum thresholds) of a number of less basic capacities. As a result, the respectful state must view the persons under its jurisdiction as possessors of properties that supervene on ranges of those capacities – that is, range properties. In virtue of these additional range properties, the state must regard those persons as equals in ways in which those persons are not required to regard and treat each other as equals.

The particular range property that interests us in the present context is the range property supervening on the talents and skills that the individual brings to the market place. Call this range property ‘being a potential contributor to society’. As long as an individual citizen has some talents and skills that are potentially valuable to others, she is a potential contributor to society (we might assume that all moral persons possess such minimal talents and skills).Footnote ¹⁰ The state is aware that different citizens have different talents and skills (and might indeed promote the development of talents and skills in a general and indiscriminate way), but is required, as a matter of respect for each particular citizen, to avoid assessments of particular individuals’ talents and skills. As a result, in considering each individual citizen as a bearer of talents and skills, the state appropriately views her as falling within the relevant range, and therefore as possessing the range property of being a potential contributor to society.

We need, then, to distinguish two kinds of opacity respect: first, there is the opacity respect that is owed by individual persons to each other (in virtue of their basic agential capacities). Call this ‘person-to-person opacity respect’, or, if you like, ‘horizontal opacity respect’. Second, there is the more demanding kind of opacity respect owed by states to citizens (or, more literally, by agents of the state to individual citizens). Call this ‘state-to-citizen opacity respect’, or, if you like, ‘vertical opacity respect’.Footnote ¹¹ The two kinds of opacity respect confer moral salience on different range properties, and these different range properties in turn serve to constitute different kinds of basic equality among citizens. Call these two kinds of basic equality, respectively, ‘moral basic equality’ and ‘political basic equality’.

What egalitarian duties should be assigned to the state if we assume political basic equality as the relevant starting point? We have said that justice concerns the proper distribution of the benefits and burdens of social cooperation, and that the tasks of the state include the enforcement of that distribution. Among the benefits of social cooperation are income and wealth, hence the inclusion of those benefits in the list of Rawlsian social primary goods. Which distribution of income and wealth should the state enforce? To the extent that citizens are basically equal in terms of the characteristics in virtue of which they are entitled to enjoy the benefits of social cooperation (characteristics which include the range property of ‘being a potential contributor to society’), they should receive those benefits equally, and the state should act to counter any inegalitarian effects of any non-entitling (that is, ‘morally arbitrary’) characteristics of citizens, to the extent that this can be done without violating state-to-citizen opacity respect. In particular, the state should refuse to recognize, in its practical deliberations, any reason why a particular person's talents and skills should result in her enjoying more or less primary goods than are enjoyed by others. We know, as a matter of empirical generalization (and without evaluating the talents and skills of any particular individual), that those with the lowest levels of talent and skill tend to be those with least income and wealth. Therefore, countering the effects of talent-and-skill differentials on income-and-wealth distribution will tend in practice to consist in implementing a sustainable system of taxes and benefits that make the worst-off (in terms of income and wealth) as well-off as possible.Footnote ¹²

Rawlsians should concede to Cohen that the difference principle is a second-best principle. If a government is as egalitarian as it would be on the above account, it should regret the fact that the market propensities of some individuals preclude even more egalitarian levels of taxation (Cohen Reference Cohen2008: 379). Self-interested behaviour on the part of the more productive members of society represents a ‘feasibility constraint’, the absence of which would allow the government to do even better in its egalitarian endeavour to increase the income and wealth of the worst-off. However, this concession need not worry Rawlsians, either substantively or methodologically.

The concession need not worry Rawlsians substantively, because the government's legitimately regretting the existence of the feasibility constraints need not entail the moral condemnation of the individuals whose behaviour constitutes those feasibility constraints. Consider, first, the point of view of the government, and recall that we are here assuming a restrictive account of individual egalitarian duties. In the light of the twofold account of basic equality expounded above, the government will recognize both the sense in which the state's strongly egalitarian duties follow from the basic equality of individuals as appropriately viewed through the eyes of the state, and the sense in which individuals’ more limited egalitarian duties follow from the basic equality of individuals as appropriately viewed through the eyes of individuals. As a result, despite its moral regret at not being able to do more for the worst-off, it will not hold that the individuals responsible for the constraints on its egalitarian endeavour have been failing in any of their egalitarian duties. Consider now the point of view of the better-off individual, again assuming a restrictive account of her egalitarian duties. In the light of the twofold account of basic equality, this individual will develop a ‘dualist’ sense of egalitarian justice. Her sense of egalitarian justice concerning the behaviour of the state will indeed tell her that the government would succeed in pursuing its egalitarian duties to an even greater degree if she, the individual, were to choose to work at the same level of productivity for a lower salary. But this sense of egalitarian justice concerning the behaviour of the state is one that she acquires by adopting the moral point of view of the state, thus viewing herself and others as fellow citizens as opposed to employers or employees or teachers or students. Adopting the moral point of view of the state, we have seen, does not entail placing any moral blame on the better-off individual for failing to work at the same level of productivity for a lower salary. It might, on the other hand, be taken to imply that the individual's working at the same level of productivity for a lower salary would be supererogatory in egalitarian terms (provided that it does not conflict with any of her moral duties, and is therefore morally permitted). In other words, we might say that such personal behaviour is valuable in overall egalitarian terms, but above and beyond the call of egalitarian duty.

And neither need the above concession constitute a methodological worry for Rawlsians, for it will not imply a fundamental change in the Rawlsian position on the issue of fact-dependence versus fact-independence. The difference principle is indeed a fact-dependent principle, but then, as we saw in section 2, all egalitarian principles are fact-dependent principles. The difference principle is simply located a little further down the line than are egalitarian principles that prescribe the equalization of something: in the case of the difference principle, the relevant facts include not only facts about equally possessed properties of individuals (basic equalities), but also facts about how individuals act, or are likely to act, in the light of the duties and permissions entailed by the various ways in which they appropriately regard each other as unequals.Footnote ¹³

5. A DIAGNOSIS OF THE POWER OF COHEN'S CRITIQUE

The twofold account of basic equality supplied in the previous two sections can help Rawlsians to escape Cohen's charge of inconsistency. However, Rawls himself does not provide such an account. Rather, he stipulates only one kind of basic equality enjoyed in virtue of our possession of the single range property of moral personality.Footnote ¹⁴ Much of the plausibility of Cohen's critique of Rawls depends, I submit, on this last feature of Rawls's account – a feature that seems to be preserved, implicitly, in most existing state-based accounts of the grounds for distributive equality.

Consider the following objection put forward by Cohen.

Before developing a twofold account of basic equality, I found this objection deeply disturbing from a Rawlsian point of view. Much of its force, I now think, depends on the fact that Rawls apparently derives his expansive account of the state's egalitarian duties from what I have called moral basic equality. This kind of basic equality ought, we feel, to exist in the eyes of individuals no less than in those of the state. If, then, the moral basic equality of persons is sufficient to generate, via contractualist reasoning, an expansive account of the egalitarian duties of the state, why should it not be sufficient to generate, via that same reasoning, an expansive account of the egalitarian duties of individuals? How can a single basis of equality be said to generate different sets of egalitarian duties for different agents?

One way to answer the above query would be by denying that individuals have any egalitarian duties at all. According to this answer, moral basic equality generates only one set of egalitarian duties of one agent: the strongly egalitarian duties of the state. Duties of individuals, on the other hand, are not egalitarian, for individuals need not regard or treat each other as equals. However, denying the existence of any egalitarian duties of individuals becomes counterintuitive once we take into consideration, under pressure from Cohen, individual duties to respect the basic liberties. It seems difficult to deny Cohen's assumption that fully just individuals will continue to respect each others’ equal right to the basic liberties where the state fails to enforce them. If, however, we recognize such an individual egalitarian duty, we must be presupposing a sense in which individuals appropriately regard each other as equals. And if this sense is the same sense as that in which the state must regard individuals as equals, it is difficult to see why it should not generate an individual duty to act on the difference principle.

The twofold account of basic equality offered in the previous two sections provides Rawlsians with a way out of this dilemma. Given the equality of individuals in each others’ eyes (that is, their moral basic equality), individuals have a natural duty to respect, or even to promote, the equal basic liberties to which they are all entitled. Given the equality of individuals in the eyes of the state (that is, their political basic equality, or basic equality as citizens), individuals have a right that the state act on the difference principle without themselves having a natural duty to do so (even though they have a duty to promote just behaviour on the part of the state). Individuals can develop a dualist sense of egalitarian justice thanks to this twofold account of the basic equality from which egalitarian duties arise.

Consider now the state-based arguments for distributive equality discussed at the end of section 2. As I suggested, those arguments can be supplemented with the arguments I provided in section 3 in favour of the moral relevance of the Rawlsian range property of moral personality. Perhaps these arguments for distributive equality within the borders of a state will then be seen to be both plausible and complete. It is not clear, however, that they will permit a Rawlsian account of the proper site of the difference principle in addition to a Rawlsian account of its proper scope.

If one says, for example, that equal moral persons are owed a special kind of equal treatment within the confines of the state in virtue of their subjection to coercive state power, it is difficult to see why such equal treatment should not be among the duties of citizens. ‘Shared citizenship . . . gives rise to a concern with relative deprivation that is absent in the international realm’ (Blake Reference Blake2001: 260). This may well be. But why, Cohen will ask, should that ‘concern’ not be shared by the individuals who partake in that ‘shared citizenship’? If one talented citizen is materially advantaged under a particular state-enforced distribution of primary goods, and if all citizens have equal moral standing, the talented citizen will presumably owe the less advantaged citizens a justification for her choice of that distribution rather than of the distribution that she could have brought about by working at the same level of productivity for a lower salary. Her freedom of contract and occupational choice gave her the freedom to choose which of these two distributions the state should enforce. She should now explain why she chose the distribution that left one or more badly-off people worse-off than under the alternative distribution. (If, on the other hand, we allow the citizen to appeal to her greater talents in justification of the unequal distribution, why should it be inappropriate for the state to do so in justification of even greater inequalities)?

Similarly, if one claims to derive an expansive account of the state's egalitarian duties from the interdependence of equal moral agents who reciprocate in contributing to a set of collective goods available within a particular state (Sangiovanni Reference Sangiovanni2007), it is difficult to see why one should not endorse an expansive account of the egalitarian duties of those reciprocating agents.

The twofold account of basic equality offered in the previous two sections allows the Rawlsian to escape this egalitarian implication for individual behaviour. Starting from a particularly strong kind of outward dignity on the part of citizens vis-à-vis the state, this account posits a range property (‘being a potential contributor’, or perhaps of ‘being a reciprocator’) that should exist in the eyes of the state but not in those of individuals. From this site-specific range property it derives a site-specific basic equality of individuals, from which we can in turn derive strongly egalitarian principles for the state without deriving similar principles for individuals.