As indicated in my Introduction, the minimal secularism I defend disaggregates the notion of religion and places reasonable disagreement about justice (not only the good) at the heart of its theory of public reason and exemptions. The aim is to defend a liberalism that takes a middle course between the antireligious proclivities of political liberalism, on the one hand, and the relativist pluralism of critics of secularism, on the other. My interlocutors in this symposium challenge this account of minimal secularism from two opposite perspectives. The first group (Micah Schwartzman, Lori Watson, Avia Pasternak) thinks my liberalism is too permissive towards religious claims The second group (Melissa Williams, Mark Storslee) argues that my liberalism is, on the contrary, too exclusive or ungenerous towards religious claims. In this reply, I attempt to defend minimal secularism from both charges, emphasizing the role that different conceptions of equal respect play in identifying the boundaries of reasonable disagreement about justice.

Micah Schwartzman pointedly asks about the political import of the modesty of Liberalism's Religion. I began thinking about its themes as the hopes triggered by the Arab spring and the early reforms of Erdogan in Turkey suggested the possibility of religious liberal states. At that time, scholars in various disciplines began exploring the contours of “multiple secularisms” as alternatives to the Western model of church-state separation. What was missing, I felt, was a systematic exploration of whether these alternative secularisms are compatible with liberalism. Can a state recognize a religion and remain recognizably liberal?

The context has changed since. With the rise of Islamism in the Middle East, the capture of the US Right by Christian movements, a worldwide backlash against women and gay rights, and the assertion of a defensive nationalism associated with authoritarian religion in many places from India to Israel to Poland, we observe an extreme polarization between religious conservatives and secular liberals. Many liberals are tempted to draw the conclusion that any political settlement that falls short of a strict separation between religion and state is illiberal. This, I think, is a mistake. Liberalism is compatible with a range of state-religion relations.

I identify and defend the parameters of this compatibility—what I call the limited state, the justifiable state, and the inclusive state. Together, they define the contours of a liberal “minimal secularism.” Even though the political space for minimal secularism is shrinking under the pressure of polarization, the tone of the book is hopeful rather than melancholic. We still have good reasons to hold on to that vision. It is a modest, but philosophically principled, vision.

Schwartzman asks whether, at a time of triumphant resurgence of authoritarian religion, it is a good time to be modest about liberalism. Political liberalism is a dualist theory: it identifies broad principles of liberal legitimacy (including those of minimal secularism) which can accommodate a range of reasonable conceptions of justice. A liberal state must accommodate reasonable disagreement about justice. Liberal-minded religious conservatives should not be treated on a par with illiberal religious fundamentalists. Modest liberalism, as a result, grants religious exemptions to some morally ambivalent claims (such as conservative religious demands, on which more below) and allows for procedural, democratic solutions to political conflict. Modesty, therefore, is not a pragmatic withdrawal, denoting a lack of confidence in what we take to be just. It is, rather, a substantive commitment that takes seriously pluralism and the burdens of judgment, and considers democratic politics as the forum in which most of those debates about justice are conducted—with interlocutors with whom we reasonably disagree: adversaries rather than enemies.

By contrast, a monist view of liberalism—which postulates that there is only one conception of justice that liberal states can legitimately pursue—encourages the view that all dissenters are ipso facto unreasonable and illiberal. Those dissenters are likely to become alienated and disdainful of what they see as the imperialist claims of liberals. Schwartzman asks: Why should we be worried about this alienation? Is this not similar to appeasement—a pragmatic and unprincipled betrayal of our principles for the sake of peace and tranquility?

A modest liberal, however, has principled, not prudential, reasons to worry about certain kinds of alienation. Sometimes, religious conservatives are rightfully alienated by a liberalism that assimilates all religious critics to unreasonable fundamentalists or fanatics. Schwartzman's appeasement charge, therefore, only works against those who do not embrace modest liberalism. Being principled, modest liberalism will worry about rightful forms of alienation. So the political and the philosophical questions are much more intertwined than Schwartzman presents them.

Lori Watson puts pressure on my defense of exemptions. She argues that “no citizen can justifiably claim a right to subordinate others so that they may live out their conception of the good.” I agree that religious exemptions that subordinate, or deny the basic rights of, others are never permissible. In particular, religious motivation cannot excuse unfair discrimination on grounds of gender, religion, race, or sexuality, as such discrimination flatly contradicts liberal principles of equal respect.

The problem, however, is that what exactly constitutes equal respect, and the rights that it grounds, is the subject of reasonable disagreement between liberals. (By contrast to Watson, I do not think that this crucial question can be dissolved or addressed by shifting to a different conception of integrity.) Do we have rights not to be exposed to offensive speech? Do we have rights that others treat us civilly? Do we have rights that all associations be structured according to egalitarian norms? Liberals reasonably disagree about the scope of the demands of equal respect. At one end are relational egalitarians, who argue that all social interactions should be permeated by an egalitarian ethos. At the opposite end stand what we may call orthodox political liberals, who argue that the principles of egalitarian justice only apply to the basic structure—institutions and laws. As long as the state guarantees distributive fairness, a liberal society protects extensive rights to personal and associational freedom. Like Watson, I am sympathetic to relational egalitarianism: it is a key commitment of the republicanism of nondomination that I defend elsewhere.Footnote ¹ But orthodox political liberals, who insist on a moral division of labor between basic institutions and individuals, also have a place in the broader liberal family. Disagreement between relational egalitarians and orthodox political liberals is a reasonable disagreement about liberal justice.

Granting exemptions is one way to recognize reasonable disagreement about liberal justice. Religious exemptions protect claims that are compatible with at least one reasonable conception of liberal justice—and therefore show appropriate respect to those who hold it. In particular, exemptions protect what I call “morally ambivalent claims.” These claims are not so morally abhorrent that they violate basic rights (such as cases of unfair discrimination mentioned above): morally abhorrent claims would be impermissible under any conception of liberal justice. Morally ambivalent claims are those whose rightfulness depends on adjudicating the contested question of the scope of relational equality. Excluding women from religious leadership, speaking against equality, following traditionalist norms of civility: these are morally ambivalent claims that qualify as legal “rights to do wrong” in liberal societies. My claim is not that such exemptions are required, but merely that they are permissible.

Avia Pasternak invites me to think harder about which exemptions are permissible in liberal society. Can it really be correct to say that a church will be exempt from antidiscriminatory law no matter how objectionable its religious doctrine is? I do not say that. As Pasternak recognizes, I place strict limits on the justification of the exemptions claimed. Churches can, for example, only discriminate if they have an open, transparent doctrine that provides theological grounding for the discrimination in question: they should not be left with arbitrary discretion in their employment practices. But does this imply that, as long as a church makes its doctrine public, even if the doctrine is morally abhorrent, it has discretion to discriminate? Pasternak is right that any collective exemption, just as any individual exemption, must meet a criterion of “thin acceptability”: it cannot be morally abhorrent.

She then raises a difficult case. Should groups motivated by a racist religious doctrine be allowed to refuse members on explicitly racist religious grounds? There are several complexities here, which I can only briefly allude to. First, we need to decide whether (nonviolent) racist groups in general have a place in liberal societies: whether they should be free to speak, organize, and demonstrate, but also to associate and dissociate according to their racist prejudices. Second, some religious groups define their own membership on ethnic or racial grounds, and do so without necessarily expressing racist attitudes. Ultra-Orthodox Jewish associations apply strict rules of matrilinearity. Black churches in the United States are also restrictive, as they were part of the struggle against racial segregation. Neither, I think, are racist in ways captured by the wrong of unfair discrimination. Third, most problematic are those religious groups whose exclusionary doctrine is specifically motivated by racist animus—such as white-supremacist churches in the United States. Should these be allowed to deny membership to African Americans?

This is a complex matter; here let me simply say two related things. First, whatever legal protection such religious groups should receive should be equivalent to the protections enjoyed by other racist groups. There is nothing special about religion, such that religious speech or association should provide its practitioners fewer or greater immunities and privileges. Second, whether racist groups are themselves permissible in a liberal society is, I think, an issue of reasonable disagreement between liberals. Orthodox political liberals would point out that freedoms of speech and association are there to protect objectionable speech and practices, provided no one's basic rights are infringed. Relational egalitarians, by contrast, would argue that the mere existence of such groups helps perpetuate racist attitudes of subordination and domination, and thereby undermines the equal civic standing of the members of targeted racial groups.

Now, it is quite possible that, in certain societies, there is no reasonable disagreement about the way in which the harms of racial segregation, combined with the continuing pervasiveness of racism and inequality, render certain expressive exclusions incompatible with political liberalism. The legacy of racial subordination in the United States is such that there is no justification for extending robust rights of free association to white supremacist groups—religious and nonreligious. To that extent, Pasternak is right to suggest that I have gone too quickly over the case of white-supremacist churches in the United States.

I claim that for laws to be justifiable, they must be supported by accessible reasons: reasons that citizens can understand and debate. Melissa Williams worries about the exclusion of those whose reasons are merely “intelligible” rather than “accessible.” She urges me to accept the normative status of intelligible reasons: reasons that are reasons only for the speaker. This is because the experience of vulnerable and marginalized groups can often be expressed only by appeal to such reasons. Which reason becomes accessible—which becomes part of the public currency of democratic debate—depends on structural positions of subordination and domination. The risk is that, by insisting that only accessible reasons can ground laws, we silence subordinated groups and perpetuate their subordination. I make two comments in response to this important worry.

First, there is no restraint on what ordinary-citizen, civil-society groups can say. As Williams concedes, because of my sociological and historical perspective on the dynamic of public reason, I suggest that we cannot be a priori prescriptive about which reason or argument is likely to become accessible. There is a temporality to the process of trial and error of democratic contestation, such that we only find out ex post which reason can become accessible. However, I am prescriptive in one sense: public officials and representatives have a responsibility, at time t, to present publicly accessible reasons when they seek to justify coercion on all: actual citizens should only be coerced by reasons that they can understand. Legislators should not justify laws by appeal to reasons that are merely intelligible to them.

The claim, however, is not that officials should not use intelligible reasons. Rather, it is that these cannot be the only reasons for coercion. Intelligible reasons can be used in support of, or to clarify the evolving meaning of, accessible reasons. Williams's comments, then, help me articulate the sense in which intelligible and accessible reasons are mutually supportive. Let me give a couple of examples.

Freedom of religion is an accessible reason. Yet to understand which claims properly fall under its scope, we need to hear intelligible accounts of how certain unfamiliar or alien practices—say, wearing an Islamic niqab, ingesting peyote—express the value of freedom of religion, in the experience of the practitioners themselves. Another accessible reason is the badness of sexual harassment. It can only be communicated via the experiences and stories of women who have suffered it. The #MeToo movement provided the intelligible stories that make the coercive enforcement of antiharassment legislation possible. In these cases, personal experiences generate reasons for others because they are shown to fall under the scope of broader accessible, public values.

We can now get back to Williams's example—a speech by Carol Mosely-Braun about the meaning of the Confederate flag. It is not only an intelligible argument. It is an intelligible argument that connects to an accessible reason. Some symbols are experienced by African Americans as graphic representations of white supremacy: this is a reason for others to act insofar as they are committed to the accessible values of equal civic status. What Mosely-Braun did in her speech is to connect intelligible with accessible arguments in just this way.

All of these are cases of complex interaction between personal and public, subjective and objective, particular experiences and general principles. My subjective, personal experience can give me reason to act, but it does not by itself give others reasons to act, particularly not via political coercion. In particular, my experience of alienation might generate intelligible reasons, but if these do not connect to good accessible reasons, it cannot give reasons for others to deploy state coercion. For example, the subjective alienation of racists from a liberal egalitarian state might be intelligible, but it has no weight because it does not connect to any public accessible value. The upshot, again, is that only rightful alienation, not alienation per se, generates normative reasons to act politically.

The related question whether religious experiences and arguments get a fair treatment in the law is raised by Mark Storslee. He points out that the concept of religion refers to a unique constellation of concerns—only it combines ethical conscience, cultural identity, associational freedom, vulnerability to discrimination, government incompetence, and so forth. Paradigmatic religious groups such as the Amish and Native Americans deserve special protection precisely because all those concerns are present. No secular group or commitment is as multidimensional; and we may conclude that religion is “more than the sum of its parts.”

I remain unconvinced that it is an objection to the disaggregation approach. I do not deny that religion often exhibits all these relevant features; nor do I deny that secular beliefs and practices will, at best, exhibit a few. The disaggregation approach in fact explains why the Amish and Native American claims deserve special protection: they tick all the boxes, so to speak. Furthermore, the disaggregation approach is compatible with keeping the semantics of “religion” in the law. My argument is not that religion should be dispensed with as a legal category, only that we must be clear about the plurality of concerns that it covers.

But it does not follow, I think, that religion is “more than the sum of its parts,” if this implies that, once the semantics of religion are invoked, the relevant claims fall into an opaque “black box” only to be approached with deference and reverence. This is because not everything that is claimed as religious in fact exhibits all, or indeed any, of the legally relevant concerns. The burden of my argument is precisely to show that religion should be subjected to a dual regime of special protection and disability only to the extent that it does exhibit one or several relevant features.

Let me draw on a few illustrations I develop in the book. Religious arguments should only be barred from public justification if they are publicly inaccessible (chapter 4). Government can dismiss corporate religious claims if these are made on behalf of commercial, nonidentificatory associations such as Hobby Lobby (chapter 5). Majority religions have less of a claim to exemptions than religious minorities, because they are not as vulnerable to discrimination (chapter 6). All these examples show, contra Storslee, that the relevant features of religion will not “always [be] intersecting and woven together.” In these cases, religion is, indeed, less than the sum of its parts; and an interpretive approach, such as the disaggregation approach, is preferable to a purely semantic approach.

I hope to have shown that minimal secularism has the resources to respond to these probing critiques. More work needs to be done, however, to identify the scope of reasonable disagreement about the boundaries both of religion and of liberalism. I am deeply grateful to my commentators for forcing me to think harder about these questions.