Liberalism, Political, and Comprehensive

Jeremy Waldron. Handbook of Political Theory. Editor: Gerald F Gaus & Chandran Kukathas. Sage Publication. 2004.

The Background Difficulty

The modern distinction between ‘political’ and ‘comprehensive’ versions of liberalism arises in connection with a serious problem about the basis of justification for liberal principles in a pluralistic society. The problem arises as follows.

Liberals envisage a tolerant, inclusive society, populated by people adhering to a variety of belief systems. Many modern societies in which liberalism flourishes as a political ideal already have this character: they are religiously pluralist and multicultural societies, in which heritages and ideals of all sorts rub shoulders with one another and compete for adherents, and in which communities of faith and tradition share quarters with groups committed to radical exploration of new ways of living, thinking, and being. Societies of this sort of course face the same challenges that all human societies face, and they must deal with the questions of justice and order that affect human societies generally. How are property and the economy to be structured? What is the extent of each person’s responsibility for the fate of others and for the social fabric as a whole? How are structures of freedom and responsibility, mutual forbearance and mutual aid, coordination and co-operation, power and participation to be defined? Those are questions for every society.

But a pluralist society also faces an additional agenda. Where different faiths and cultures rub shoulders, there is likely to be friction and offence: one group’s worship or festivities might seem like a reproach or an attack on another group, and as values and philosophies compete in the marketplace of ideas, the competition will often seem disrespectful as each creed tries to discredit its opponents and gain adherents for itself. It is not easy to define the duty of mutual toleration under these circumstances, or to sustain the distinction between harm and offence that a pluralistic regime requires. And that is not the only distinction that pluralism threatens. The line between public and private, between issues of policy and social welfare on the one hand and individual ethics and religious or cultural observance on the other, is always going to be an issue. Certain cultures and religions in a pluralistic society may aspire to be a society unto themselves. A religion, for example, may have its own values with a distinctive bearing on the problems of social life, and it may impose quite particular obligations (for example, dietary laws or rules of religious observance) on its members, which may or may not be compatible with the society’s broader social arrangements. To make law and policy for a pluralistic society is thus a greater challenge than for a society that is religiously and culturally homogeneous. The latter just needs to settle on a single set of answers and enforce them. But the former has to deal with the fact that its members are already firmly wedded to disparate answers. The various answers may be incommensurable; but even if they are mutually intelligible, they may not present themselves simply as rival political opinions about how to solve the problems faced by the larger society. Instead they may present themselves as claims of identity, demanding accommodation as a matter of justice, or as a matter of respect for the persons whose religious and cultural allegiances they represent.

Now these distinctive difficulties associated with pluralism are not in themselves the problem which I mentioned in the opening lines of this chapter. Indeed, the challenge that I have just outlined is one to which liberals respond gamely and enthusiastically. Since the rise of religious toleration in the West, liberal political philosophy has made a speciality of arguing about structures of order, justice, and liberty for pluralistic societies; it has made a speciality of arguing about the distinctions between harm and offence and between public and private. Though liberals disagree with one another on many of these issues, they pride themselves on their willingness to confront these difficulties and deal with them honestly, straight on, without wishing them away. There is a whole heritage of reflection on the values and principles that may be used to define a fair social and political order under conditions of freedom and diversity. We find it in the canon of liberal theory—in the work of John Locke, Immanuel Kant, the French philosophes, the Federalists, John Stuart Mill, as well as the more problematic contributions of Hobbes, Rousseau, Hegel, the utilitarians, and the new liberals who combined with the canon some of the insights of socialism. There may not be consensus among these various liberal thinkers, but there is a wealth of resources to draw on.

But here is the difficulty. The ideas that we draw on in order to elaborate and defend liberal principles and liberal solutions to the problems outlined above are often ideas associated with particular philosophical traditions. The sanctity of life and bodily integrity, the importance of autonomy, consent and individuals’s control of their own destiny, the concern we are supposed to have for each other’s self-development, the inherent value supposedly attaching to the satisfaction of an individual’s preference, the respect accorded to ethical and spiritual thinking at the level of the individual mind and conscience, the key position occupied by reason and rationality, and the principle of equality which associates these values with all human beings, whatever their sex, race, age, or social position—these are artefacts of a particular tradition or cluster of traditions that have grown up in our civilization. Many of us find them compelling. But we cannot be under any illusion that they are features of every culture or tradition that we expect to find represented in a modern pluralistic society. They are features of some world views and not others. So: by elaborating and defending liberal principles and liberal solutions to the problems of social life on this sort of basis, we seem to be taking sides in the midst of cultural and ethical plurality. We seem to be picking and choosing among the variety of ethical, philosophical and religious traditions in the world, privileging some as foundational and marginalizing others.

An Example: Locke on Toleration

An example will illustrate the difficulty. John Locke’s Letter Concerning Toleration (1983) is plainly one of the foundational documents in the early modern liberal tradition of religious freedom and religious plurality. But part of the Lockean defence of religious toleration is built up on religious foundations: ‘The toleration of those that differ from others in matters of religion,’ says Locke, ‘is so agreeable to the Gospel of Jesus Christ, that it seems monstrous for men to be so blind as not to perceive the necessity and advantage of it in so clear a light’ (1983: 25). And he goes on to argue that the essence of religion is sincere belief, that Jesus used persuasion rather than force to win adherents to the Gospel, that he did not endow his apostles with earthly power, and that he did not need to because religious heterodoxy by one person is not harmful to the soul or spiritual health of another. Now these are Christian arguments, and there is no reason why they would be convincing to everyone for whom Locke envisaged toleration (let alone for everyone for whom we envisage toleration). True, Locke advanced other more pragmatic arguments for toleration that were not dependent on Christian conceptions in this way. But he evidently thought the Christian foundations were important, and it is arguable that his case would be incomplete or vulnerable to fairly straightforward objections if it were not defended in this way.

So Locke is put to his choice. Either he defends liberal toleration on a basis that is rooted in a particular world view (or a narrow class of world views), and risks diminishing its appeal as something held in common among a variety of faiths. Or he seeks a more pragmatic defence that can appeal to people who start from a genuine variety of religious and ethical assumptions; but it may have to be a shallower and philosophically less sophisticated defence. The second approach may not say, in defence of toleration, everything that John Locke thinks it important to say: maybe what is really important about toleration isits agreeability to the Gospel of Jesus Christ. Still this second kind of defence—if it can be concocted—will have the advantage of being more widely appealing. Now I don’st think Locke ever in fact faced up to this choice. The politics of his audience was not such that he needed to: though he proposed toleration for Muslims and Jews, and other non-Christians, the success of his proposal did not depend on the acceptability of his argument in their eyes. The mainstream audience he addressed was a Christian audience, and so he could afford to develop a Christian argument. (Indeed the mainstream audience he addressed was a Protestant audience, so he could afford to flirt with grounds for toleration that were unacceptable to Roman Catholics.) In our day, however, the politics of toleration are different: for us, a politically sensible defence of the principle of toleration has to command the allegiance of people of many different faiths (and some with no faith at all). If it cannot be articulated in a way that makes sense to them, it cannot be articulated as part of the shared, public order of the society. And so we face a dilemma. A liberal theory of toleration should try to say what is important about toleration, why it matters, and how values and ideals that seem to point in the other direction can be rebutted. Can this be done in a way that is acceptable to everyone, in a way that does not alienate the adherents of some faith or some philosophy in society? Maybe not. Maybe you cannot see what is really important about toleration except from a perspective that invokes particular values and particular philosophical conceptions. If that is right, then either we opt for a shallow theory of toleration with a broad appeal to diverse groups in society, or in the interest of developing a deeper theory of toleration, we face up to the fact that our liberal commitments will be seen as rooted in the values and conceptions of some particular philosophical outlook.

Defining ‘Political’ and ‘Comprehensive’ Liberalism

These two approaches have come to be associated with the labels ‘political liberalism’ and ‘comprehensive liberalism.’ Though ‘political liberalism’ is now associated particularly with the recent work of John Rawls (1993), both labels should be understood in the first instance as referring to types of position.

This is partly because ‘liberalism’ itself is not the name of a determinate set of social and political commitments. There are certain core positions and the various schools of liberal thought may have a family resemblance to one another. But in many areas they offer rival conceptions of the values that are characteristically associated with liberalism, like liberty, equality, democracy, toleration, and the rule of law. Two political liberals may therefore be distinguished from one another by their different positions and their different conceptions. But what they will have in common—as political liberals—is their insistence on a distinction between the principles and ideals that (in their respective views) define a liberal order for society, and the deeper values and commitments associated with particular philosophical outlooks. The political liberal insists that the articulation and defence of a given set of liberal commitments for a society should not depend on any particular theory of what gives value or meaning to a human life. A comprehensive liberal denies this. He maintains that it is impossible adequately to defend or elaborate liberal commitments except by invoking the deeper values and commitments associated with some overall or ‘comprehensive’ philosophy.

There may also be a second layer of difference among political liberals. Whether or not the substance of their liberal commitment is the same, two political liberals may differ in the justificatory strategies they adopt as political liberals. One may emphasize the idea of an ‘overlapping consensus’ a variety of justificatory paths from disparate philosophical premises to a plateau of liberal principles. (This is Rawls’s view, which we shall examine shortly.) Another may opt for a ‘lowest common denominator’ approach, emphasizing justificatory premises that all members of a pluralist society may be presumed to accept, whatever the differences in their ethics or world view. And the phrase ‘may be presumed to accept’ may be glossed in various ways, ranging from the idea of universally accessible reasons and reasoning to some fairly aggressive account of basic human interests, like the survivalist account developed by Hobbes (1991).

Obviously there are important differences, also, among comprehensive liberals. A first layer of difference is the same as for political liberals. Two comprehensive liberals may have different liberal commitments: one may be a left liberal and the other a libertarian liberal. A second layer of difference has to do with the content of the comprehensive outlooks on which their liberal commitments are based. John Locke’s Christian foundations are not the same as Immanuel Kant’s (1991) theory of autonomy, and none of those is the same as the hedonistic foundation of Jeremy Bentham’s (1982) utilitarianism. But they all have this in common: they relate liberal commitments in political philosophy to some vision or conception of what matters in life and of the human person and its place in the world. And they are united, too, in their negative conviction that the political liberal cannot complete the assignment that he has taken up. Liberalism, says the comprehensive liberal, is a robust position in political philosophy, a position whose moral partisanship reaches deep into the foundations of our conceptions of person, freedom, and value.

The general difficulty that I have outlined, and the two kinds of response to it, have not always been a staple of discussion in liberal political philosophy. Many of the canonical figures in the liberal tradition are unapologetically ‘comprehensive’ liberals in the sense that their conceptions of social order and their elaborations and defences of freedom and equality are patently rooted in a deep and extensive vision of the human person and of the ethical significance of human interaction in society.

I have already mentioned John Locke’s argument for toleration. Locke’s more general theory of politics (his theory of inalienable natural rights, his critique of slavery, his contractarianism, his argument against absolutism) also has a straightforwardly comprehensive character. It rests on the axiom that men are ‘all the workmanship of one omnipotent, and infinitely wise maker; all the servants of one sovereign master’—God—‘sent into the world by his order, and about his business,’ and that ‘they are his property, whose workmanship they are, made to last during his, not one another’s pleasure’ (Locke, 1988: 271). One might also mention the political philosophy of Immanuel Kant. Kant’s basic principle of right—‘An action is right if it can coexist with everyone’s freedom in accordance with a universal law’ (1991: 56)—is presented as part of a metaphysical system that makes rational sense of our ability to distinguish those aspects of force and constraint that are a necessary part of any social order from those that are condemned by the value we attribute to freedom. We proceed in political life as though this distinction were of tremendous importance, and Kant’s position is that we cannot explain it without a philosophical theory of the relation between reason, universality, and the harmonization of human wills. And one might cite, too, the work of John Stuart Mill in this regard. A superficial reading of Mill seems to indicate that he was proposing only a political principle: ‘the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others’ (1956: 13). But Mill’s commitment turns out to depend on a particular vision of human flourishing, in which (as Wilhelm von Humboldt put it, in a passage Mill cited), ‘the end of man… is the highest and most harmonious development of his powers to a complete and consistent whole,’ and therefore the object ‘toward which every human being must ceaselessly direct his efforts… is the individuality of power of development’ (1956: 69).

It does not seem to have occurred to Locke, Kant, and Mill that these foundational positions would pose a problem for the politics of liberalism in a society whose members disagreed about the existence of God, the nature of reason, and the destiny of the human individual. They just took it for granted that liberalism required a philosophical foundation of this kind, and that their task as political philosophers was to articulate that foundation, convince (as Mill put it) ‘the intelligent part of the public… to see its value’ (1956: 90), and if necessary argue, as Locke argued in his discussion of atheism (1983: 51), that those who could not subscribe to these foundational positions might have to be regarded as dangerous by the government of a liberal society.

The Principle of Liberal Neutrality

The difficulty I have raised came to the fore in discussions of ‘liberal neutrality’ in the 1970s and 1980s. A number of theorists attempted to sum up the essence of liberalism in terms of a principle requiring the state to refrain from taking sides on disputed ethical and religious questions. Thus Ronald Dworkin suggested the liberal commitment to treating people as equals meant that

political decisions must be, so far as possible, independent of any particular conception of the good life or of what gives value to life. Since the citizens of a society differ in their conceptions, the government does not treat them as equals if it prefers one conception to another, either because the officials believe that one is intrinsically superior, or because one is held by the more numerous or more powerful group. (1985: 191)

Dworkin did not suppose that neutrality was a general moral requirement, one that everyone should strive to satisfy. Neutrality was proposed as a principle of specifically political morality. It is not wrong for someone to favour a particular conception of what gives value to life, but it is wrong for him to do so in his capacity as a legislator or as a judge. It is not wrong for a church or a firm to pursue some particular spiritual or ethical religion, but it is wrong for the state to do so (Larmore, 1987: 45). The idea had a lot in common with American constitutional doctrines of state action: the First Amendment makes it unconstitutional for the state or the law to favour a religion and the Fourteenth Amendment makes it unconstitutional for the state or the law to discriminate, but neither provision is read as prohibiting religious choice or even racial discrimination by individuals, firms, churches, or clubs (except in cases where their private actions can plausibly be imputed to the state).

Liberal neutrality may be seen as a generalization of religious toleration into the realm of ethical choice generally. The liberal state was no longer required merely to be neutral as between religions; it had to be neutral also between almost all aspects of its citizens’ conceptions of the good, whether these were spiritual or secular (the only exceptions being conceptions of the good that themselves denied liberal principles). But therein lay the position’s vulnerability. So long as liberalism was read as a principle about religious neutrality, its defence could be rooted in moral ideas. Once it expanded into the ethical realm, it was not clear what it could rest on. It couldn’st be based on scepticism about values, for it seemed to represent a particular commitment in the realm of value (Dworkin, 1985: 203). Liberal theorists scrambled to define a distinction within the realm of values between political morality (e.g. moral principles of justice and right, like the neutrality principle itself), on which the state was permitted to act, and ethics (and perhaps the rest of morality besides justice and right), on which it was not permitted to act (Waldron, 1993: 156-63). But it was always a fine line, and the wider world tended to blur the distinction between ethics and morality and talk generally about the liberal commitment to value neutrality. Once the position was characterized in that way, the defenders of neutrality faced a dilemma: either they left their principle undefended, or they faced accusations of non-neutrality on their own side in respect of the values which they adduced to justify it.

A similar dilemma confronted those who tried to use neutrality as a meta-principle of political justification. Bruce Ackerman (1980) developed a theory of justice in the form of a contactarian dialogue, for which it was laid down as a ground rule that no reason (adduced in conversation to justify any particular distribution of power) ‘is a good reason if it requires the power holder to assert… that his conception of the good is better than that asserted by any of his fellow citizens’ (1980: 11). Now, why should this be the ground rule? Ackerman said that there were several ways to justify the neutrality principle: it could be justified by reference to the epistemic value of experiments in ethics, or the intrinsic importance of autonomy, or scepticism about ethics, or about the ability of power-holders to reach accurate conclusions about the good (1980: 11-12). The liberal state need not side with any of these justifications in particular. It only needs an assurance that everyone can reach neutrality by at least one of these routes.

Could this strategy work? It might, but only if we were certain that the different paths to neutrality did not make a difference to the meaning or character of the destination. But this seems unlikely. Moral principles are characteristically dependent for their interpretation on some understanding of the point or purpose for which they are imposed. Change the purpose and you provide a different basis for interpreting the principle. So far as neutrality is concerned, one of the main interpretive difficulties concerns the issue of intention: does neutrality forbid only political action motivated by a non-neutral intention or does it forbid also action, however motivated, which is non-neutral in its effects? It turns out that some of Ackerman’s paths to neutrality favour the intentionalist interpretation while one, at least, favours the consequentialist interpretation: scepticism about a power-holder’s ethical abilities should inhibit only his deliberate attempts to favour one conception of the good. The value of ethical diversity, on the other hand, should make us pause whenever state action actually has a detrimental impact on some conceptions of the good, whether this is intended or not. Ackerman’s ‘overlapping consensus’ is really a recipe for a disordered society, as citizens follow their different paths to an interpretive quarrel and find no common basis to resolve it (see Waldron, 1993: 151-3).

Rawls’s Political Liberalism

When it was first published in 1971, John Rawls’s book A Theory of Justice seemed to present itself as a set of more or less universal claims: it was supposed to tell us what justice was and what it required in any society which faced what Rawls called ‘the circumstances of justice’—moderate scarcity, mutual disinterest of individuals in one another’s ends, and so on (1971: 126). Under these circumstances, Rawls seemed to be implying, it was appropriate for people to use the idea of the ‘Original Position’—decision behind a ‘veil of ignorance’—as a way of figuring out appropriate principles of justice. And he argued that anyone selecting principles from that perspective would adopt strong principles of equal basic liberty, equal opportunity, and a social framework oriented to the well-being of members of the worst-off group. He seemed prepared to argue for these conclusions and defend them against rival conceptions (like Nozick, 1974) as a conception which could command the support of anyone interested in the subject.

Rawls’s withdrawal from Comprehensive Theory

Through the 1980s, however, Rawls began to offer a more modest characterization than he had in 1971:

[W]e are not trying to find a conception of justice suitable for all societies regardless of their particular social or historical circumstances… We look to ourselves and to our future, and reflect upon our disputes since, let’s say, the Declaration of Independence. How far the conclusions we reach are of interest in a wider context is a separate question… What justifies a conception of justice is not its being true to an order antecedent to and given to us, but its congruence with our deeper understanding of ourselves and our aspirations, and our realization that, given our history and the traditions embedded in our public life, it is the most reasonable doctrine for us. (1980: 518-19)

This resonated with a theme emphasized by Walzer (1983) that a well-ordered society is a society true to its own understandings or, if it is to be reproached as unjust, it has to be reproached as having fallen away from values that already have a purchase in the life and practice of its members.

That amounted to a withdrawal from moral universalism in one direction: Rawlsian justice was not a theory for all societies, but a theory for societies like the United States. But it then required us to focus on some of the particular characteristics of societies like the United States, and the most prominent of these—apart from their prosperity and their traditions of political stability—was their religious and cultural diversity. Ethical and religious heterogeneity were no longer to be regarded as a feature that societies governed by justice might or might not have, or might have at one period but not at another. It was to be seen instead as a permanent feature of the societies, one that could not be expected soon to pass away.

By the beginning of the 1990s Rawls had become convinced that his approach in A Theory of Justice was disqualified generally on this ground. Though it contained, in the principle governing basic liberties, a robust defence of toleration and mutual accommodation, it grounded that approach in a particular vision of the human individual—‘a thin theory of the good’ (1971: 395-9)—according to which individuals have a fundamental interest in forming and following a rational plan of life which enables them to realize and exercise the whole range of their individual capacities. An individual’s active engagement with this task, Rawls suggested, is the basis of his or her self-respect. Some commentators (e.g. Barry, 1995) have expressed doubts about Rawls’s self-criticism that the adoption of this ‘thin theory’ means that A Theory of Justice was rooted in a particular comprehensive conception. But it is pretty clear that large parts of Rawlsian justice would not work without this thin theory of the good and of the importance of self-respect. The thin theory of the good and the notion of self-respect are implicated in the non-negotiable status that Rawls accords to freedom of conscience, for example, as well as in the general doctrine of the priority of liberty, the doctrine of the priority of opportunity, and his argument to the effect that citizens in a well-ordered society will not be motivated by material envy. Someone who did not regard self-respect as so important, or did not associate it so tightly with individual self-development or the active pursuit of value, might well come up with other conclusions on any or all of these fronts. Moreover, the individualism of Rawls’s thin theory drew criticism from communitarian philosophers, who repudiated the implicit assumption that individual plans of life are chosen by persons unencumbered by prior commitments and allegiances. Those who thought of themselves as essentially members of a particular family or community or people might find it hard to accept a theory of justice oriented at foundational level to the well-being of persons conceived as liberated from all such attachments (Sandel, 1982).

Rawls’s Strategy

‘[H]ow is it possible,’ Rawls asked, ‘for there to exist over time a just and stable society of free and equal citizens who remain profoundly divided by reasonable religious, philosophical, and moral doctrines?’ (1993: 4). In the introduction to Political Liberalism, he argued that this could no longer be achieved by convincing everyone of the ethical and philosophical premises on which a comprehensive liberal theory of justice might be founded. Instead Rawlsian justice would now have to be presented as something that could command support from a variety of ethical perspectives.

The first task in this new Rawlsian agenda was to characterize the array of comprehensive views which needed to be taken into account in our thinking about justice: must a conception of justice be accessible from literally every standpoint that we find represented in society, or are we entitled to ignore or marginalize some as crazy or unreasonable? A second task was to define the appropriate relation between a conception of justice and the various comprehensive doctrines that the political liberal was required to take seriously: should we think of the conception of justice as a modus vivendi, or should it be related more robustly to the relevant comprehensive doctrines either by way of minimal shared premises or by way of overlapping consensus? A third task—and this was the substance of the new approach—would involve the detailed reformulation of a conception of justice in a way that was responsive to these specifications: how many of the substantive principles and doctrines of A Theory of Justice would survive this new approach? Finally, what would be the implications of this new approach so far as the actual politics of a liberal society were concerned?

All parts of this agenda have proved challenging and controversial, and there has been considerable debate about the ability of Rawlsians and other political liberals to carry through their programme on these four fronts. In what follows I shall outline Rawls’s views and explain some of the difficulties they face.

Reasonable Diversity

The starting point of Rawls’s new approach was the sheer ‘diversity of opposing and irreconcilable religious, philosophical and moral doctrines’ (1993: 3-4) flourishing in modern society. Rawls described this diversity as a social fact—a permanent feature of modern society. Some who would agree with him about that might nevertheless regard such diversity as a pathology endemic to the modern or postmodern condition. But Rawls takes no such approach. Human life engages multiple values and it is natural that people will disagree about how to balance or prioritize them. What’s more their different positions, perspectives and experiences in life will give them different bases from which to make these delicate judgements. Together factors like these make disagreement in good faith not only possible but predictable.

However, not all dissensus is regarded as reasonable. Some positions, he says, are just crazy and irrational, and those do not in themselves present a compelling case for accommodation. He seems to think that a society will have to exclude aims that are unreasonable in this epistemic sense: ‘In their case the problem is to contain them so that they do not undermine the unity and justice of the society’ (1993: xvii).

Unfortunately, as Rawls uses it, the term ‘reasonable’ is ambiguous as between this and another use. Sometimes he uses it in the sense of something that represents a fair use of human reason under modern circumstances. Other times, he uses a more moralized definition: persons are reasonable when

they are ready to propose principles and standards as fair terms of cooperation and to abide by them willingly, given the assurance that others will likewise do so… By contrast, people are unreasonable… when they plan to engage in cooperative schemes but are unwilling to honour, or even to propose, except as a necessary public pretence, any general principles or standards for specifying fair terms of cooperation. They are ready to violate such terms as suits their interests when circumstances allow. (1993: 49-50)

What is crucial on this second definition of reasonableness is one’s willingness to share a world with those who accept other conceptions of the good or the holy, one’s willingness to submit one’s own convictions along with theirs to the governance of neutral principles. Plainly the two senses of reasonableness come apart. There are views which are tolerant but nevertheless irrational, and—this is the greater problem—there are views which are reasonable in the sense of not crazy, but whose orientation to social accommodation with other views is problematic. Militant Islam might provide an example of a comprehensive conception whose claims are (arguably) reasonable in the first (epistemic) sense, but quite unreasonable in the second sense of willingness to live in accommodation with others.

Overlapping Consensus

Assuming we can define, even roughly, the set of conceptions of the good that must be accommodated in the approach that we take to justice and the justification of the basic structure of a liberal society, what is the relation supposed to be between the set of reasonable conceptions and an acceptable theory of justice?

One possibility is to insist on something like a unanimity requirement: i.e. we could say that no theory of justice is acceptable if members of a given conception of the good are inclined to repudiate it. But this is much too strong, and in a way that misconceives the nature of the difficulty that political liberalism addresses. The problem is not that theories of justice are controversial; the critical reaction that led Rawls to modify the approach he took to the subject of justice was not that people (like Nozick, 1974, for example) disagreed with his principles on justice-related grounds. The problem was that some people would have a particular kind of difficulty with his theory, based on the terms in which it was formulated and the approach it took (for the purposes of justice) to the question of what matters in social life. The key, then, is to insist that an acceptable theory of justice, T, must be such that, among whatever reasons there are for rejecting T or disagreeing with T, none turn on T ‘s commitment to a particular conception of value or other comprehensive philosophical conception. Obviously, of course, this is a threshold test only: T may be acceptable in this sense, but still unacceptable overall as a theory of justice. But this would be for justice-related reasons, not because of T ‘s complicity with a particular comprehensive conception.

And there are further questions about how this threshold test should be understood. One possibility is that T represents an acceptable modus vivendi for the adherents of the various comprehensive conceptions{C 1, C2,…, Cn}. Like a treaty that puts an end to conflict between previously hostile powers, T may be presented as the best that C1 can hope for in the way of a theory of justice given that it has to coexist with C 2,…, Cn, and the best that C2 can hope for given that it has to coexist with C1, C 3,… C, and so on. Rawls, however, regards this as unsatisfactory as a basis for a conception of justice. It leaves Tvulnerable to demographic changes or other changes in the balance of power between rival comprehensive conceptions, a vulnerability that is quite at odds with the steadfast moral force that we usually associate with justice (1993: 148).

Instead Rawls develops the idea that T should represent an overlapping moral consensus among {C 1, C2,…, Cn}. By this he means that Tcould be made acceptable on moral grounds to the adherents of C 1, and acceptable on moral grounds to the adherents of C 2, and so on. The grounds of course would not be the same in each case. But still the adherents of each comprehensive conception would adhere to T for moral reasons. Thus, for example, the proposition that religious toleration is required as a matter of justice may be affirmed by Christians on Lockean grounds having to do with each person’s individualized responsibility to God for his own religious beliefs, by secular Lockeans on the grounds of unamenability of belief to coercion, by Kantians on the grounds of the high ethical importance accorded to autonomy, by followers of John Stuart Mill on the basis of the importance of individuality and the free interplay of ideas, and so on. There are different routes to the toleration principle, and adherents of the various conceptions of the good in the society embed toleration in their own overall philosophy in different ways. Still each accepts toleration for moral reasons, of the right sort of weight and stringency, and each knows that all the others do this as well.

Whether this actually works is an issue we considered when we discussed Ackerman’s approach to neutrality. The idea of overlapping consensus assumes that there can be many routes to the same destination. Geographically the metaphor is plausible enough, but when the destination is a set of moral principles, and ‘routes’ is read as reasons for the acceptance of those principles, then the matter is less clear. Unlike legal rules, moral propositions are not just formulas. A principle is perhaps best understood as a normative proposition together with the reasons that are properly adduced in its support. On either of these accounts, the principle of toleration arrived at by the Christian route is different from the principle of toleration arrived at by Mill’s route. And this is a difference that may matter, for a theory of justice is not only supposed to provide a set of slogans for a society; it is also supposed to guide the members of that society through the disputes that may break out concerning how these slogans are to be understood and applied.

Justice in Political Liberalism

There are many sources of disagreement about justice, and in the previous section I tried to emphasize that the rivalry among comprehensive conceptions does not account for all of them. Accordingly we should not understand the strategy of the political liberal as a strategy of attempting to suppress all basis for disagreement about justice. Political liberals should think about justice as a topic that naturally evokes disagreement even when the influence of rival comprehensive conceptions is left out of account. The fact that one major source of dissensus is removed should not lead us to assume—what many political theorists mistakenly assume about rights—that what is just and unjust can be determined in some realm of principle that is beyond politics, some arena of philosophical argument where political procedures like voting will not be necessary. Like individual rights, justice remains an intensely contested issue, and though the contestation may be diminished it is not eliminated by the strategies that the political liberal proposes.

A further question is whether political liberalism actually defines a terrain on which disagreements about justice can actually be played out. It is hard to tell from Rawls’s later work, for very little of it is concerned with detailed issues of social and economic justice, or with controversies of the level and ferocity that the earlier book evoked. But I suspect the answer is ‘No.’ Social justice, after all, raises concerns that can hardly be dealt with by the strategy of vagueness or evasion associated with overlapping consensus—putting about a set of anodyne formulas that can mean all things to all people. A theory of social justice has hard, critical work to do, on Rawls’s original account: it has to settle complex questions about freedom, equality, desert, and opportunity, and it has to hold its own against rival conceptions (against Nozickian historical entitlement, for example, or against utilitarian or efficiency-based approaches). The actual examples of overlapping consensus for a pluralist society provided in Political Liberalism are laughably easy by comparison. Both Kantians and non-Kantians might favour democracy, Rawls says, and both Christians and secularists may well oppose slavery (1993: 122-5). The hard part comes when we try to establish an overlapping consensus among (say) Christian fundamentalists, Hindus, secular humanists, scientific determinists, and members of the dot-com generation on the definition of ‘equal opportunity,’ the use of economic incentives, and the distinction between liberty and the worth of liberty.

Just to give a taste of the difficulty, consider the problem of the relevance of desert to basic social entitlement. This was central in social justice discussions in the 1970s and 1980s, and various approaches to it informed people’s views about market success, the problem of the undeserving poor, and so on. Now, it was not hard to see that insistence on a strong theory of desert might mean that a theory of justice would have to buy into social and religious controversies about virtue. But it was much more difficult to know what to do with that point, or what would be a fair or a neutral way to move on from it. Does one simply reject desert in this sphere (and the whole view of the person that goes with desert), or does one try to develop a thin theory of desert, or to modify the assumptions, e.g. about freedom and background responsibility for character, that deserving is sometimes thought to presuppose? Can we imagine an overlapping consensus on problems like that between (say) the Protestant work ethic, the notion of apostolic poverty, and ideas of the fundamental solidarity of community? It is easy to despair of answering questions like this under the conditions that Rawls’s later work has emphasized.

Public Reason

So far our discussion has been pitched at a rather abstract philosophical level: on what sort of basis is it appropriate to construct a theory of justice? But Rawls is also interested in exploring the consequences of his political liberalism for real-world political arguments. A theory of justice, on his account, is not just some set of esoteric formulas; it is supposed to be something public, something shared among the citizens as a common point of reference for their debates about the allocation of rights and responsibilities. So political liberalism also has implications for what this sharing a conception of justice amounts to. We do not share a conception of justice, Rawls argues, and our society is not well ordered by his lights if, when there is argument about the allocation of rights and responsibilities, conceptions of value are invoked which not everybody shares.

Rawls believes this point about public reason is pretty obvious when we think about the way in which some of our political institutions are supposed to operate. For example, the justices of the US Supreme Court cannot invoke ‘their own personal morality, nor the ideals and virtues of morality generally’ (1993: 236). They must view these as irrelevant to the issues they decide, not just because they are unconnected to the texts and doctrines the courts are supposed to interpret, but because it would be disrespectful to justify the exercise of public power on grounds that they knew many citizens would quite reasonably—be unable to endorse. Rawls generalizes this to apply also to the other arms of government in their public deliberations and acts and decisions. But he does not leave it there. He also believes it applies to the citizens, who are after all exercising a modicum of public power when they decide how to vote or when they bring pressure to bear on the government or its agencies. At least when they are addressing the fundamentals of justice, Rawls believes that citizens must search their consciences to ensure that they are not voting one way or another under the auspices of principles that they know their fellow citizens cannot accept, and—as a matter of basic civility—when they make arguments in the public realm, they must address these arguguments not just to their co-religionists or those who share their values, but to all their citizens conceived as participants in the just ordering of a society dedicated to the accommodation of all who hold reasonable views about what makes life worth living. So a left liberal like me may not say, for example, to a Social Darwinian that even the feeblest person is entitled to our compassion because he is created in the image of God. I must find some way of putting my point about equality that can be affirmed even by people who do not share my religious convictions. Equally a Christian conservative may not justify laws restricting abortion on the grounds that foetuses have souls, since this too is rooted in a comprehensive conception he cannot expect others to share.

Interestingly some of the discussion in Political Liberalism of the abortion example showed how difficult it is to apply this stricture in practice. In a footnote to the original edition Rawls inferred, from the fact that anti-abortion laws usually rest on controversial religious grounds, that liberty in this regard was required (1993: 243n). But he quickly had to concede that that was a mistake (1996: lv), for three reasons. First, we are not entitled to assume liberty in such an area as the default position, any more than we are entitled to conclude that foetuses do not have souls from the fact that political liberalism is unable to countenance religious arguments to the effect that they do. Second, although there might be good neutral arguments for a right to choose abortion in the first trimester, we must not assume that there are no contrary arguments or no way of opposing abortion rights that does not run foul of the strictures of political liberalism. Many opponents of abortion will insist that their arguments for protecting human foetuses are continuous with arguments (that they insist any theory of justice must acknowledge) for protecting all human life, particularly in its most vulnerable forms. They do not accept that a political liberal can casually cut off debate about this and still have a strong doctrine of human dignity and human equality to deploy in other areas of justice where such doctrines are indispensable. Third, the fact that a religious doctrine may not be appealed to in order to justify restrictions on abortion doesn’st mean that such doctrines are altogether beyond the pale. Rawls came close to implying that doctrines which might have this consequence are ipso facto unreasonable, and so need not even be considered as constituents in the overlapping consensus on which principles of justice are to be based (1993: 243n). It would surely be catastrophic for Rawls’s theory if say all or even most religious conceptions were to be excluded from the realm of the reasonable. But if they are not so excluded, then (as we saw in the earlier section on overlapping consensus) there must be a way of reaching the public doctrine of respect for human life from the premises of these religious accounts. And it is not at all clear that that way can be charted if the end-point is supposed to be a position that the religious adherent cannot but regard as offensively flawed and inconsistent.

The issue illustrates how quickly these arguments can turn into a debate about the whole viability of the political liberal’s approach. On the one hand, the political liberal says that the only doctrines of human dignity and human equality we are entitled to deploy are doctrines that are elaborated independently of any comprehensive conception. On the other hand, the political liberal knows that these doctrines have work to do in a theory of justice that requires a considerable amount of moral weight. The doctrine of human dignity and equality deployed in a theory of justice must be able to resist in more or less the manner of a moral absolute various pragmatic considerations that might tempt us to sacrifice or neglect the interests of a few weak and vulnerable persons for the sake of the convenience or prosperity of the wealthy or powerful. Justice has to be able to stand up to that, and its constitutive doctrines have to have what it takes to do that heavy moral lifting. Many of the comprehensive conceptions that political liberals want to exclude from the public realm address themselves to exactly this issue: they explain in ethical or transcendent terms why exactly it is that the few weak and vulnerable may not be sacrificed in this way. The political liberal proposes to do this work without help from any such conception, but in a way which nevertheless retains their allegiance in overlapping consensus. It is, I think, a tall order, and the tendency of latter-day Rawlsians to shy away from hard issues of justice like their masters’ own equivocations on the issue of abortion—provides ample reason to think that that tall order cannot be filled.

Liberalism and Unanimity

One of the impulses behind the move to political liberalism was the suggestion by some liberal theorists (e.g. Waldron, 1993: 43ff) that social arrangements must be not only justified in the abstract, but justifiable to each and every one of persons who have to live under those arrangements and who are (potentially) subject to the force that backs them up. This suggestion is bound up with the basic liberal idea of government by consent, i.e. the principle that the exercise of power can be made legitimate only when those who are subject to it can accept the principles on which its exercise is based. Of course this suggestion might be taken more or less literally: we might talk about principles that everyone actually does accept or we might talk about principles that people would accept if they were well informed, thinking logically, and so on (see Gaus, 1996). Still, the thought was that if political justifications rested on values that derived from religious or ethical conceptions held by some citizens but repudiated by others, then they would not satisfy even the looser versions of this requirement of justifiability to all.

However, the dative element in this requirement that political justification be understood as justification to each and every individual—can be understood in more than one way. It may be understood as a requirement that the justificatation of political arrangements should be directed to the good or interests of each and every one who is subject to those arrangements. I shall call this the ‘interest-regarding’ interpretation. Or it may be understood as a requirement that the justification of a political decision be plausibly reckoned likely to persuade everyone who is subject to the arrangements. I shall call this the ‘premise-regarding’ interpretation, because it understands ‘justification to X’ as justification that seeks to hook up with premises to which X is already committed.

It’s pretty clear how the two interpretations may come apart. Suppose I justify the criminalization of prostitution on the ground that this is good for the souls of the prostitutes. A prostitute who is an atheist may find my justification unacceptable, perhaps even unintelligible. Still it does purport to address her interests (only not as she understands them). On the other hand, taking the premise-regarding interpretation, I might justify a law against prostitution on grounds that are perfectly intelligible to the prostitute, but they might be grounds that take no account whatever of her interests. I may say, for example, that it is in the interests of the majority of decent citizens that prostitution be banned. The prostitute may understand this argument, and if she were submissive and demoralized enough she might even accept this justification. Still the justification would fail my test on the interest-regarding interpretation, because it would not be addressed to her interests.

Clearly Rawls’s political liberalism assumes what I have called the ‘premise-regarding’ interpretation of the requirement that political justification must be justification to each and every individual. It may accept the interest-regarding interpretation as well, for the two are not mutually exclusive. Rawls indicates, in his later work, that political liberalism will still have room for the idea of social contract and choice of principles of justice in an original position (1993: 304-10), and these are conceptions which model the idea that principles are acceptable only if they advance the interests of all. A political liberal, however, will have to articulate these ideas carefully so that they do not embody—for example, in the assumptions they make about the motivation of parties in the original position—anything associated with particular conceptions of the good.

However it is also important to see that interest-regarding interpretation of justifiability to all can be maintained even if the premise-regarding interpretation is given up. Our earlier case of the do-gooder’s concern for the soul of the prostitute provides one crude example. But even if one felt uneasy about this example, the uneasiness may not be best explained by political liberalism. Someone may reject the salvation-of-the-soul argument for the law against prostitution on the grounds that this makes no sense to the prostitute. But his basis for that may not be political liberalism. It may rather be that the alleged justification fails to connect in the appropriate way with the prostitute’s well-being. A case may be made on comprehensive grounds for insisting that a justification counts as connecting in the appropriate way with X’s interests only if it connects with X’s well-being as X understands it (or as X would understand it under moderately favourable conditions). And this condition might be argued for, not as a weak version of political liberalism, but on the basis of some affirmative account of what well-being is and of the importance of a person’s own conscious engagement with her well-being. For example, the terms of some comprehensive conception may be such that it is implausible to attach moral importance to X’s well-being and to insist that others respect it unless X herself already affirms it or could be thought to affirm it under suitable conditions. Liberals have always insisted on paying attention to how things actually are for the people they claim to respect, and they have been impatient with political proposals oriented to a person’s real self, where that self is impossibly distant from the person’s occurrent experience. But this is not on account of any metatheoretic requirement of neutrality of the sort maintained by political liberals; it is on the basis of an affirmative focus on the here and now, and on how things actually are for people in their own felt sense of what actually matters to them.

Now here’s the point. Some comprehensive conceptions will affirm the moral importance of people’s actual experience here and now, while others may sideline or denigrate it. Those that do affirm it will sit more naturally with, and in a way will generate and inspire, the moral and political commitments traditionally associated with liberalism. And that is what the comprehensive liberal wants to remind us of. Liberalism is based on certain ethical commitments, certain propositions about what matters and about the importance of certain kinds of respect for the lives, experiences, and liberty of ordinary men and women. It is not a neutral or nonchalant creed, and its commitments arguably cannot be articulated at a purely political level. Traditional liberals have said that human life, liberty and experience command respect, period, not just because of the way we configure our politics but because of what they are. And the same is true of many of the moral absolutes and the stringency and priority associated with justice and rights. These again are not just pragmatic matters established at a political level; instead their application in politics is derivative of deeper truths about the nature of the goods they protect and the moral concerns that they express.

So, although liberals seek a universal application for their principles in a world of many faiths and many philosophies, it may be a mistake to base that universal application on the shallowness of the liberal claims. If liberal positions are to be sustained, certain confrontations at the comprehensive level may be unavoidable. A willingness to face up to these issues and to explore these deeper foundations may be the price one has to pay for robust liberal convictions.