Rawls and Religion

An important objection to political liberalism is that it has no way to adjudicate conflicts between public and nonpublic reasons.¹ It cannot explain why citizens should adhere to the constraints of public reason when these are at odds with their comprehensive ethical or religious perspectives. In such circumstances, citizens may wonder why they must conform to public reason rather than giving priority to conflicting nonpublic values.

Consider the following schematic example: Suppose citizen C agrees with the broad outlines of the ideal of public reason; she attaches great value to a political system in which citizens justify the exercise of their collective power in terms that everyone can reasonably accept. Suppose further that C is faced with a choice between two mutually exclusive public policies, α and β. After thinking carefully about the issues, she concludes that only β can be supported by a reasonable balance of political values. Thus, she has a pro tanto, or political, justification for β. This means that, so far as public reason is concerned, β is politically justifiable.2 Now, under normal circumstances, C might be satisfied with this outcome. She accords significant weight to the value of giving others public reasons, and the fact that public reasoning reaches a conclusive determination is usually sufficient to settle the matter for her. In this case, however, C worries that her comprehensive view conflicts with her pro tanto justification for β. She believes that she has weighty nonpublic reasons to favor α, and these reasons conflict with the pro tanto political justification of β. To arrive at a full, or all-things-considered, justification of α or β, she must balance two competing sets of values. On one side is the set of political values that justifies β, as well as the political values that support the general practice of giving public reasons; on the other is the set of nonpublic values drawn from C’s comprehensive view. The question is: why should she give priority to the values of public reason?

Critics argue that even if the ideal of public reason is supported by weighty values such as fair social cooperation, mutual respect, and political autonomy, there are significant costs attached to “bracketing” or setting aside the values of one’s comprehensive doctrine.³ An adequate defense of public reason would have to evaluate such costs and show that they are insufficient to undermine the duty of civility, which requires citizens to give public reasons for their political decisions. Unfortunately, the objection continues, political liberalism cannot provide such a defense because it is constrained by its own agnosticism about the nature and truth of comprehensive doctrines. The only way to establish the priority of public reason is to weigh conflicting public and nonpublic values, and this political liberalism is either unwilling or unable to do.

It is worth exploring the possibility of using a form of nonpublic reason to help answer this objection. The basic strategy is to argue that C should respect the limits of public reason by demonstrating that the pro tanto justification of β is indeed fully justified according to her comprehensive view. The full justification of β is based, at least in part, on a nonpublic justification, according to which C is committed to supporting β because her comprehensive view, when properly understood, is compatible with and perhaps even justifies the practice of providing others with public reasons. John Rawls refers to this type of argument as “reasoning from conjecture.” As he describes it, “we reason from what we believe, or conjecture, may be other people’s basic doctrines, religious or philosophical, and seek to show them that, despite what they might think, they can still endorse a reasonable political conception of justice.”⁴

My purpose here is to elucidate and defend the idea of reasoning in this way. In particular, I hope to show that engagement with comprehensive ethical and religious views can play a limited but potentially significant role in defending a commitment to public reason. I do not argue that reasoning from conjecture provides a complete response to objections concerning the priority of public reason, only that such reasoning may open up a domain of nonpublic justification for promoting the ideal of public reason. Nor is my aim to do the first-order work of engaging in reasoning from conjecture.5 My broader purpose is to describe the structure of this form of reasoning (the first part of the chapter) and to show that, at least in principle, it can be defended against objections based on concerns about insincere argumentation (the second part), threats to cultural sovereignty (the third part), and challenges to the epistemic authority of those who reason from within comprehensive views other than their own (the fourth part). After responding to these three objections, in the fifth part of the chapter I discuss some limits to reasoning from conjecture, and then conclude.

We argue from what we believe, or conjecture, are other people’s basic doctrines, religious or secular, and try to show them that, despite what they might think, they can still endorse a reasonable political conception that can provide a basis for public reasons. The ideal of public reason is thereby strengthened. However, it is important that conjecture be sincere and not manipulative. We must openly explain our intentions and state that we do not assert the premises from which we argue, but that we proceed as we do to clear up what we take to be a misunderstanding on others’ part, and perhaps equally on ours.⁷

Three aspects of reasoning from conjecture (or conjecture, for short) are worth noting here. First, as its name suggests, reasoning from conjecture is a form of reasoning.8 It is not simply an attempt at rhetorical persuasion. The idea is to present others with arguments that give them good reasons, as evaluated from within their own comprehensive views, for supporting a reasonable political conception as the basis for public reasoning. Thus, reasoning from conjecture is a form of nonpublic justification. Its purpose is to generate rational agreement on a shared or public basis for political decision-making.

Another way of describing the aim of conjecture is in terms of Rawls’s distinction between pro tanto and full justification. As mentioned above, C has a pro tanto, or political, justification for policy β when that policy is supported by a reasonable balance of political values.⁹ But she does not yet have a full, or all-things-considered, justification for supporting β. If her comprehensive view contains nonpublic values that conflict with the pro tanto justification for β, then she may decide to give priority to those values. In that case, C has a pro tanto justification for β, but not a full justification for it. The aim of conjecture is to bridge the gap between pro tanto and full justification by showing that C’s comprehensive view supports, is consistent with, or at least is not in conflict with β.¹⁰

In general, political liberalism leaves it up to individual citizens to determine how pro tanto political justifications fit within their comprehensive doctrines.¹¹ But when citizens fail in this regard, or when they appear to be unable to reconcile the demands of a reasonable political conception with their comprehensive views, it may be possible to continue deliberating on the basis of conjectural arguments. If it can be shown that a comprehensive view is at least consistent with a reasonable political conception, and not in conflict with all such conceptions, then perhaps the weighty values that support providing others with public reasons—such as mutual respect, reciprocity, and political autonomy¹²—will be sufficient to establish a full justification for a reasonable political conception as the basis for making political decisions.

If it is asked why a full justification is required, or why it is not sufficient for political liberalism to provide a pro tanto justification for a political conception, the answer is that, without a full justification, citizens may believe they have nonpublic reasons to act in violation of principles acceptable under a reasonable political conception. Because it is undesirable, and perhaps unjust, for citizens to act in this way, reasoning from conjecture is justified as a means of showing that citizens may have nonpublic grounds for respecting the demands of public reason. In other words, conjecture aims at providing a nonpublic justification for being reasonable.13 Since political liberalism avoids specifying a complete account of why people ought to be reasonable, conjecture may facilitate the development of such accounts for particular comprehensive views, if and when they are needed.¹⁴

Second, those who engage in conjecture—call them “conjecturers,” for lack of a better term—offer arguments based on premises they do not accept. They argue from within comprehensive doctrines other than their own,¹⁵ for the purpose of justifying a reasonable political conception. For example, if A conjectures that C should endorse justice as fairness, what she says to C is, “You should agree with Rawls’s principles of justice because your comprehensive doctrine commits you to endorsing them. I don’t happen to agree with your comprehensive doctrine. As I see it, I am committed to justice as fairness for different reasons. But given what you believe, I think you should be committed to the same conception of justice.” Those who reason from conjecture must make clear that, as Rawls says, “we do not assert the premises from which we argue.” This is, in part, to reassure others of our sincerity. In making this clarification, however, we are also asking others to take our arguments seriously on the ground that these arguments are based on premises to which we believe they are committed.

Third, despite what its name might suggest, reasoning from conjecture need not be based on arguments that are merely tentative or suggestive. To some extent, Rawls trades on the uncertainty implied by the word “conjecture.” But this is not essential to the basic idea. Conjecturers may be quite certain that their claims are conclusively justified on the basis of particular comprehensive doctrines. The central feature of conjecture is not a matter of epistemic credibility, or a lack thereof. It is rather that those who advance conjectural arguments do not share a belief in their premises. Of course, in defense of Rawls’s terminology, it might be said that, in most cases, conjectural arguments really are conjectures, in the ordinary sense of the word.16 Unless conjecturers know a great deal about specific comprehensive views, their arguments will often be speculative and inconclusive. For that matter, even when they are deeply familiar with particular doctrines, their arguments may still be relatively uncertain. This may be simply the natural result of differences in judgment. How much of a problem this poses for reasoning from conjecture is a serious question, addressed below. For now, however, the important point is that conjecture is (or should be) defined by the kind of reasoning involved, and not by the degree of confidence expressed by that reasoning.

With these three points about reasoning from conjecture in place, I now want to consider three corresponding objections. The first is that conjecture is not a form of justification but rather an insincere mode of argument designed to manipulate people into accepting liberal political principles. The second objection is that, even if conjecture is sincere and nonmanipulative, it is an attempt to impose alien interpretations on particular comprehensive doctrines. According to this objection, internal criticism of a doctrine is legitimate only when it comes from people committed to that doctrine. The third objection is that conjecture lacks epistemic credibility. It presumes that conjecturers have a better understanding of comprehensive doctrines than those who affirm them. Since this is a dubious assumption, conjecture is unlikely to achieve the goal of generating rational agreement on a reasonable political conception. I shall argue that none of these objections is decisive, although the third presents a challenge that citizens and public officials must take seriously if they are to engage in this form of nonpublic reasoning.

In this case, reasoning from conjecture is both insincere and manipulative. A argues that C should adopt a political position on the basis of a belief that is invalid even from within C’s perspective. She knows that C is not personally justified in accepting that N → β. So when A says that C should vote for β because C has good reason from her perspective to do so, she is being insincere. A is simply playing on beliefs that she knows C would have reason to reject from within her own comprehensive view. As Gerald Gaus rightly says about such a case, “This may do its pragmatic job, but it hardly counts as a justification. Justification gives way to mere rhetoric (and perhaps cynicism) when one ‘justifies’ a belief to another by citing reasons that one thinks are inadequate even from the perspective of the person to whom they are presented.”17

If we accept that justification is about giving others good reasons, even when the “goodness” of the reason is judged from within their perspectives, then the principle of disclosure is insufficient to guarantee the sincerity of conjecture. Since conjecture is a form of justification, and thus a matter of giving good reasons, conjecturers must do more than disclose that they do not believe the premises from which they argue. They must also disclose whether they believe their arguments are justifiable from within the comprehensive views of those with whom they are reasoning. We can call this the principle of full disclosure. It says that reasoning from conjecture is sincere and non-manipulative if, and only if, conjecturers (a) disclose that they do not believe the premises from which they argue, and (b) disclose whether they believe their arguments are justifiable from within the comprehensive views of their intended audience. Provided that conjecturers are open about their intentions, and about the conditional status of their arguments, it is difficult to see why their claims should give rise to accusations of insincerity and manipulation.

Here one might respond that conjectural arguments manipulate people simply because they appeal to premises that conjecturers themselves do not accept. Notice that, while this objection is framed in terms of reasoning from conjecture, it has much broader reach. The basic argument is that whenever we offer others reasons we do not believe, we are, in some sense, attempting to manipulate them. As Robert Audi asks, “Does it not smack of manipulation to give reasons that do not move me, in order to get others to do what I want? I use the reasons as psychological levers to produce belief on a basis that does not carry my own conviction.”18 Audi calls this “leveraging by reasons,” which he defines as the attempt “to move an audience to a view by noting one or more reasons there are for it from the audience’s point of view.”¹⁹

Importantly, however, Audi accepts that “leveraging” loses its manipulative force when those engaged in it openly explain that they do not believe the arguments they are presenting. Conceding that leveraging by reasons is not necessarily manipulative,20 Audi presents a further argument to defend his underlying intuition about the wrongness of giving others reasons we do not believe. This argument, which does not rest on claims about insincerity or deceit, is that offering others such reasons demonstrates a lack of respect for them. Audi offers this argument in the context of defending his principle of secular motivation. His claim is that it is illegitimate to give others non-motivational secular reasons—defined as secular reasons that do not move one—to justify coercive public policies. Audi writes that “there is a sense of ‘respect’ in which, if we manipulate others, we thereby show a lack of respect for them. But I also suggested that there is a lack of respect in offering, as adequate secular reasons, considerations that are not persuasive to oneself. Why should others be moved if I am not? Do they have lower standards?”²¹

Although Audi is concerned here with the special case of religious believers offering secular reasons they do not believe, a similar point might seem to apply in the case of conjecture. If we do not accept the reasons we offer to others, why should we expect them to accept those reasons? This objection might be carried even further by contrasting conjecture with the demands of public reason. Assuming that any plausible conception of public reason requires that political justifications be based on reasons we sincerely believe,²² why is it acceptable in nonpublic reasoning to offer justifications based on reasons we do not believe? Why is it that, in some situations, we demand that others give us reasons we share, but in other situations we do not?

The answer to these questions appeals to the fact of reasonable pluralism. This fact suggests that people may reasonably disagree about doctrines that address matters of fundamental importance, such as the meaning and value of human life. To recognize this fact is to see that people will often have different reasons for coming to similar moral and political conclusions. Thus, it is appropriate in nonpublic reasoning to assume that others may be motivated by reasons that are different from our own. By contrast, from the point of view of public reason, justification for state action based on nonpublic reasoning is a remedial or second-best solution, adopted only under special circumstances. To the extent possible, we ought to justify our collective decisions by appealing to public values that we share with others in virtue of our common citizenship.²³ But if some people hold comprehensive doctrines that seem to them to reject such values, even if only in particular cases, arguments that appeal to public values will not provide them with justifications they can accept from their own perspectives. Under these circumstances, we may²⁴ appeal to values drawn from their comprehensive doctrines to try to show them that they are mistaken. For political liberals, this is the only remaining form of justification available, short of appealing to considerations of prudence. Provided we are clear about our intentions, then, I see no reason to think that engaging in conjecture is insincere, manipulative, or otherwise disrespectful.

A second objection to reasoning from conjecture is that it manifests a form of cultural imperialism. This type of objection has been suggested by Abdullahi An-Na‘im, a Muslim legal scholar who has argued for a progressive account of Islamic reformation,25 which Rawls mentions as a “perfect example of overlapping consensus.”²⁶ In developing his views, however, An-Na‘im argues from within a tradition to which he is committed and indeed warns that Islamic arguments for liberal principles must be developed by Muslims and not by those who are outside the tradition.

For example, in a discussion of Islamic law and women’s rights, An-Na‘im argues that only Muslims may participate in deliberations about Islamic reformation.²⁷ He writes, “this is an internal debate for the Muslims to conduct and settle among themselves.”²⁸ An-Na‘im offers two main arguments for this view. The first is that criticisms made by outsiders are often perceived, and rejected, as forms of cultural imperialism. As he says, “to avoid even the appearance of dictation by outsiders, which is likely only to be counterproductive, the classification of certain cultural (legal or religious) norms, as archaic and oppressive, must be done by the members of the cultural or religious group themselves.”²⁹

Of course, resistance to external sources of criticism may give rise to practical reasons for refraining from conjecture. Engaging in it may be pointless or counterproductive. But that some might perceive conjecture as a form of alien imposition is not, in itself, a principled argument against presenting criticisms based on sources internal to their cultural or religious traditions. It may be the case that, when dealing with illiberal cultural or religious groups, conjecturers find that the only way to gain an audience is through cautious cooperation with internal social critics. And even this may be overly optimistic. Those who resist reform often attempt to suppress criticism from within.30 Under these circumstances, however, conjecture might be used to voice criticisms that would otherwise have been made by social critics. Whether this is possible or desirable is a question that can only be answered by looking at specific examples. The important point here is that resistance to conjecture by groups that stifle criticism, whether it originates internally or externally, does not represent a moral, or otherwise principled, objection to justifying, or at least attempting to justify, liberal principles on the basis of beliefs and values to which such groups are committed.

An-Na‘im’s second argument is a stronger claim for cultural or religious sovereignty. He suggests that every cultural or religious tradition should have the opportunity to determine its own understandings of, and justifications for, basic human rights, without “allowing any tradition to dictate to the others.”³¹ Yet this claim is not necessarily inconsistent with reasoning from conjecture. Insofar as conjecture is based on appeal to internal sources, it does not impose alien standards on cultural or religious traditions. Indeed, the very possibility of conjecture is premised on the idea that various traditions will have different ways of justifying moral and political principles, including those that call for the protection of human rights. Note, further, that the purpose of conjecture is, in part, to address cultural or religious groups that adhere to “archaic and oppressive” norms. As An-Na‘im himself recognizes, such groups “cannot be left to themselves completely.”³² They must be open to criticism, even on their own terms, unless they develop beliefs and practices consistent with respect for basic human rights and liberties.

It has to be admitted that a conclusive reply to this objection may not be available. Comprehensive doctrines may be so complex or esoteric that few people can claim to interpret them authoritatively. Some religions seem to fit this description. Believers might point out that nonbelievers have not been trained to read texts correctly, to draw on relevant analogies and metaphors, to reference apposite commentaries and interpretations, or, more generally, to frame moral arguments in terms acceptable within their religious traditions. One can imagine a Talmudist or an Islamic jurist (faqih) saying such things. Furthermore, some religious communities delegate the responsibility of answering ethical and political questions to specific authorities. The average member of the community may not be qualified to assert an opinion on such important matters. If ordinary members of a community cannot interpret with authority, how can outsiders, of all people, expect anyone to take them seriously?

Conjecturers can give one of at least three answers to questions about their epistemic authority. The first would be simply to concede the point by admitting that they are not qualified to make authoritative statements about how to interpret a comprehensive doctrine other than their own. As noted above, the word “conjecture” implies speculation, a proposition that has not been proved, or an educated guess. In the end, maybe Rawls is right to frame the matter in these terms. Perhaps tentative suggestions are all that conjecturers can hope to provide.

Those who are unsatisfied with this response, especially those committed to justifying a liberal political conception on the basis of a particular comprehensive doctrine, can undertake the task—and it may be an enormously difficult one—of becoming authorities on a particular doctrine. Obviously, they will not be authorities in the eyes of those who accord authority only to the leaders of their own communities. But if conjecturers are, or become, educated enough to argue intelligently from sources recognized by the leaders of such communities, perhaps they can make a plausible case for having their arguments taken seriously.34 The second answer, then, is that conjecturers who become students of particular comprehensive doctrines can say that they have learned what they can about those doctrines, that they have listened carefully to the moral and political claims made on the basis of them, and that they are committed to a process of sincere argumentation based on what they have learned. This reply may not always be convincing, but it is the most that can reasonably be expected of anyone not already committed to a particular comprehensive doctrine.

The third answer challenges the reasonableness of excluding an outsider, or a nonbeliever, solely on the basis of epistemic considerations. Recall that the purpose of conjecture is to try to reconcile conflicts between nonpublic and public values. If a person, or a community, asserts that nonpublic values should take priority in cases of conflict, then it is reasonable to ask for a justification for that view. Perhaps a justification can be given that will at least be intelligible, if not defensible, to outsiders. But if the proffered justification is not even intelligible, if an outsider cannot even hope to understand the beliefs or values on which that justification is based, then there are strong grounds for questioning whether it is an appropriate basis for justifying the exercise of political power.

As with all forms of moral justification, reasoning from conjecture must acknowledge certain limits. I shall focus here on two in particular, one dealing with comprehensive doctrines that continue to reject the priority of political values and the other with doctrines so fundamentally unreasonable that reasonable citizens should not attempt to argue from within them.

First, even if conjecturers can support their claims to arguing validly from what others believe—for example, by demonstrating expert knowledge of particular comprehensive doctrines—and even if the supporters of those doctrines accept their authority, nothing guarantees that those doctrines will contain the resources necessary to reconcile conflicts between public and nonpublic values. Thus, the main practical limitation to conjecture is that some doctrines may simply be opposed to certain political values or the manner in which they are applied. If no authoritative interpretation of a particular doctrine yields a commitment to the values of public reason, then conjecture has not resolved the conflict at hand. Furthermore, if reasoning from conjecture is to be sincere, those who engage in it must be prepared to admit this possibility. The principle of full disclosure requires that conjecturers disclose whether they believe their arguments can be justified from within a particular doctrine. If they do not believe this is possible, they should not pretend otherwise. To do so would be to muddle the aim of rational justification with that of mere, and perhaps even deceitful or manipulative, rhetorical persuasion. If conjecture is not to be mistaken for a form of propaganda, conjecturers should not sacrifice their claim to sincerity in moments of justificatory defeat. Better to acknowledge that some doctrines cannot justify the values of public reason than to cast aspersion on the larger project of justifying a commitment to liberal principles.

Second, some comprehensive doctrines may be based on beliefs or values that reasonable citizens cannot, in good conscience, take as starting points for justifying the values of public reason. For example, some doctrines are defined in essentially racist terms. It is, therefore, no surprise to find that those who espouse them typically reject the values of public reason.35 Yet, even if it were possible to offer conjectural arguments on the basis of morally bankrupt views,³⁶ it may seem perverse to engage in such reasoning. Indeed, insofar as reasoning from conjecture might confer some level of social or political legitimacy, there may be good reason to refrain from engaging with such views, except perhaps to discredit them. Rather than attempting to tailor moral arguments to doctrines that are completely unreasonable, it may be preferable to contain them as much as possible, and by whatever means are legally permissible according to a reasonable political conception of justice.³⁷

One might object that if we do not engage racist doctrines, or doctrines otherwise bent on domination, there is no purpose in reasoning from conjecture. After all, the need for conjecture arises because some citizens believe that they are justified in coercing others to act according to values that cannot be publicly justified. What else can this be but an attempt to dominate others through the illegitimate imposition of a particular comprehensive doctrine? We may decide not to engage such doctrines, but then what is there to conjecture about? To answer this question, it may help to draw a distinction between doctrines that are completely or fundamentally unreasonable and doctrines that contain unreasonable elements. Examples of the former include ideologies that define their politics in terms of racial, ethnic, or religious supremacy. The central beliefs and values of these doctrines may be so warped or morally corrupted that they are irredeemable from the point of view of public reason. By contrast, some doctrines may contain elements that conflict with the values of public reason. As a whole, these doctrines may be worthy of respect, even if they fail to recognize certain limitations imposed by reasonable conceptions of justice. Again, some religious doctrines may fall into this category. Although they are prima facie unreasonable, at least on some moral and political issues, internal criticism may reveal the potential for such doctrines to endorse views compatible with the demands of public reason. This possibility is, at any rate, what motivates the idea of reasoning from conjecture.

When nonpublic and public values conflict, the idea of public reason requires that citizens give priority to the latter. Yet, as critics have often pointed out, this priority cannot be justified according to political values alone. A pro tanto justification of a reasonable political conception must be confirmed by a full justification that draws on the beliefs and values of a particular comprehensive doctrine. In other words, political values must be endorsed by, or at least held to be compatible or congruent with, a broader ethical, philosophical, or religious perspective. Although political liberalism provides reasons for thinking that reasonable comprehensive doctrines will affirm a liberal political conception, it does not provide a conclusive argument for this view. Rather, it defends the possibility of an overlapping consensus as an article of liberal faith. Thus, when faced with a comprehensive doctrine that conflicts with the political values of a liberal political conception, or the values of public reason, political liberalism may seem to run out of answers. I have argued that reasoning from conjecture is a way of continuing the process of rational justification beyond the limits of public reason. It may be possible, in some cases, to appeal to beliefs and values within particular comprehensive doctrines for the purpose of showing that they can, indeed, affirm political decisions based on public reasons. Moreover, by adhering to certain precautions, this form of nonpublic reasoning can be conducted in a manner that is sincere and respectful, not culturally imperialist, and as epistemologically authoritative as anyone could reasonably expect. Conjecture may not ultimately succeed in demonstrating that particular doctrines should endorse a commitment to public reason. But that conclusion need not be assumed in advance of efforts to engage citizens on their own terms.

3. For this objection, see, for example, Michael Sandel, Liberalism and the Limits of Justice, 2nd ed. (Cambridge: Cambridge University Press, 1998), 196–202, 215–17; Stephen Mulhall and Adam Swift, Liberals and Communitarians, 2nd ed. (Cambridge: Blackwell, 1996), 228–31; and Christopher Eberle, “Basic Human Worth and Religious Restraint,” Philosophy and Social Criticism 35 (2009): 151–81.

5. For some recent examples of reasoning from conjecture, see Andrew March, Islam and Liberal Citizenship: The Search for Overlapping Consensus (Oxford: Oxford University Press, 2009); Lucas Swaine, “Demanding Deliberation: Political Liberalism and the Inclusion of Islam,” Journal of Islamic Law and Culture 11 (2009): 92–110; and Joshua Cohen, “Minimalism About Human Rights: The Most We Can Hope For?” Journal of Political Philosophy 12 (2004): 190–213.

10. Although Rawls describes conjecture as an attempt to show others they can endorse a reasonable political conception, in some cases conjecture might have more modest goals. It might be used to show that a comprehensive view is consistent with, or perhaps not in conflict with, a reasonable political conception (or with political claims derived from such a conception). In what follows, I shall not distinguish between these various forms of reconciling nonpublic and public reason. If conjecture can achieve any of them, it will have served an important purpose. I thank Tom Bailey for clarification on this point.

13. Within political liberalism, citizens are “reasonable” only if they accept (1) that political society is a fair system of social cooperation between free and equal people, and (2) that any liberal society will be marked by the fact of reasonable pluralism that results from the burdens of judgment. Accepting these two ideas, reasonable citizens recognize that the exercise of political power cannot be justified to others on the basis of any particular comprehensive doctrine. This recognition leads, in turn, to the demand for political justifications based on public reason. For discussion of these points, see Jonathan Quong, Liberalism Without Perfection (Oxford: Oxford University Press, 2011), 37–39, 166–67.

15. This condition might be arbitrary to some extent, and it could be relaxed. If comprehensive doctrines are broadly construed as traditions of thought, then chances are good that people can argue on the basis of premises they do not accept, even though these premises in some sense belong to or are associated with their own comprehensive doctrines. Nevertheless, I shall keep this restriction in place for now, if only because it simplifies matters and tends to bring out objections that might not be raised against those who argue from within their own ethical or religious traditions.

35. For the possibility that people who hold some racist views can still be reasonable citizens, see Erin Kelly and Lionel McPherson, “On Tolerating the Unreasonable,” Journal of Political Philosophy 9 (2001): 38–55; but cf. Jonathan Quong, “The Rights of Unreasonable Citizens,” Journal of Political Philosophy 12 (2004): 314–35.