Akeel Bilgrami’s paper is very rich; I cannot speak to all its arguments. I focus on his principal concern: his desire to amend Charles Taylor’s definition of secularism so that the rights to free exercise of religion and evenhanded treatment are qualified by “the ideals that a polity seeks to achieve (ideals, often, though not always, enshrined in stated fundamental rights and other constitutional commitments).” When religions are inconsistent with these ideals, “there is a lexical ordering in which the political ideals are placed first.”  These ideals can, apparently, be anything, as long as they “do not get articulated in terms that mention religion or the opposition to religion.” What is important is that they exist and that they come first.

One might ask why it is necessary to include these ideals in the very definition of secularism. Bilgrami excludes other things; he does not require, for example, that the overall orientation of the state be liberal. Why shouldn’t a state that allows free exercise and that seeks to be evenhanded count as secular even if it doesn’t affirm a set of ideals that take lexical precedence? Why can’t the state’s toleration for religion be qualified by other values without those values being incorporated into the very definition of secular? Why can’t, in other words, religious toleration be balanced against goals that lie outside the domain of secularism—outside, that is, the state’s particular concern with religion. After all, if those values are sufficiently important to have lexical priority over religious beliefs, shouldn’t they have lexical priority over every position? Bilgrami doesn’t tell us why he wants to make them internal to secularism.

What is clear is that he wants the polity’s ideals to take lexical precedence over religious beliefs and practices. Why? Apparently because he doesn’t think a secular state should be held hostage to any religion’s vision of things. His argument is framed in direct response to Charles Taylor’s argument that a secular state should be neutral with respect to religion. Bilgrami wants the polity to be able to act against some religious precepts. He refutes relativism precisely so that judgements can be made that are not constrained by religious positions.

I share the view that polities must be able to make some decisions that are opposed on religious grounds—that they must be permitted to say, at the end of the day, or at least at the end of some days, that their position must prevail. But lexical ordering is the wrong way to think about this requirement. Lexical ordering suggests that the ideals must be decided first, before one turns to religion, and that once those ideals are decided, religious beliefs and practices must simply give way. Secularists have nothing to learn from religion in the elaboration of their ideals. The ideals are articulated unilaterally, framed in advance, and the role of believers is to comply either immediately or, if forcing compliance would be too coercive to be expedient, at some later stage.

Indeed, at no point does Bilgrami so much as hint that the society might itself change through dialogue with religious positions. On the contrary, dialogue is directed only towards making “a strongly held religious standpoint … change its mind.”  His anti-relativism “allows one to make no concession to a possible right or truth or correctness on the side of one’s opponent” (his emphasis). This is excessive.

Think, for example, of the attempt to ban the hijab from public institutions in the name of secularism. Bilgrami suggests that banning the hijab has nothing to do with secularism, apparently because he believes that the ban is about hostility to Islam, not political principle. But this is wrong. The banning of the hijab from public institutions may well be misconceived but it satisfies Bilgrami’s tests. In France, the best justification for the ban is a political ideal: the policy seeks to create spheres of citizenship in which ostentatious markers of differentiation are excluded so that, in these spheres, everyone can participate in full equality, simply as citizens. That ideal is misconceived. It asks certain citizens (and only certain ones) to leave behind crucial dimensions of their identities, so that their ability to participate in public life is impaired, not fostered. It prevents other citizens from encountering their fellow citizens’ beliefs, perpetuating prejudice and misunderstanding. And the policy has always been tainted by ethnocentrism. But even those lessons cannot be learned effectively if a strict lexical ordering is followed, the ideals are elaborated without reference to religion, and they are then simply imposed.

Take another example at a tangent from religion but that involves a similar encounter of incommensurable metaphysics: I work extensively on Indigenous issues. The constitutional orders of settler states have often foreclosed crucial dimensions of Indigenous social life: they have brought Indigenous lands within the public domain of the state and then regranted them for economic development and non-Indigenous settlement; they have eliminated Indigenous jurisdictions, centralizing authority in state institutions; and they have imposed a single law, displacing Indigenous modes of social order. These measures have all formed part of the states’ central ideals—founding new societies, engaging in a civilizing mission. But look at the cost of that imposition: not only has it devastated Indigenous communities and cultures, but it has cut settler states off from dialogue with, and learning from, Indigenous societies. One of the great tragedies of Indigenous/non-Indigenous relations has been the arrogance of colonists, who thought that their ways were so infinitely superior that, out of care for Indigenous people, it was appropriate to impose their ways, without waiting to hear what they might have to learn. But isn’t that exactly what we risk doing when we take the society’s declared ideals to be lexically prior to all religious beliefs, so that the latter should simply comply?

The whole purpose of placing the ideals first in a lexical ordering is to draw a line in the sand. But it insulates the imposers from insights that might be gained from more integral engagement. Bilgrami makes an eloquent plea for the capacity of believers’ commitments to evolve with time, as a result of history, so that it is wrong to think that they are locked within an inescapable relativism. But he applies the insight only to believers. Aren’t rationalists also historical beings, whose most cherished commitments evolve with time, in ways that are similarly a product of experience? If so, we should open ourselves to the possibility of change.

This militates in favor of an ethic of engagement that is very different from lexical ordering. At the very least we should understand the beliefs and practices that we are encountering, so that we know accurately how our principles interact with them. And we might stumble upon something unexpected in the encounter: market interactions tempered by ethical injunctions to an extent that we in the non-Islamic world have neglected; an idea of gender equality that is genuinely equal but that de-sexualizes the public space; or something else.

I don’t mean that all will be sweetness and light. I don’t even mean that the suggestions in the last paragraph are a complete representation of Islamic societies today. I simply want to open up the possibility that a lexical ordering forecloses. There will be times when imposition is justified, when the majority of the community finds that, even having heard the position of the minority, they simply cannot abide certain practices. But it is essential that, at the very least, we know what we are dealing with when we proceed to that step. I suspect that, often, the range of outcomes will be considerably broader than either imposition or blank-check permission. We may well find intermediate outcomes that allow for religious concerns to be satisfied (perhaps in adjusted forms) while respecting the ideals. Indeed, our ideals may not prove to be as univocal as we thought them to be. None of this stops our societies from being, in intelligible fashion, secular.

One of the things that knocks us off-stride in this debate is an exaggerated ideal of consent as the foundation of legitimacy. We often talk about societies as though they are only legitimate when we can agree to an outcome for our own substantive reasons—as indeed Bilgrami suggests early in his paper (although it seems clear, by the end, that the secular society’s ideals are not to be subjected to the minority’s consent). But even though societies may be most legitimate when that stipulation is fulfilled, there is too much disagreement in any society for it to be satisfied in real time. Generally, people acquiesce because they want to be governed by some order of right, and for that to happen in a society in which people disagree, they know that they have to accept some norms that they would not adopt were they left to their own devices. Of course it is better if they do substantively accept them, or at least that the final determination occurs within a tolerable substantive range—and the ethic of engagement suggested here is most likely to produce such an outcome. But in any society, there is significant room for the determination of a legal outcome even in the face of continued disagreement.

What matters most, then, are the ethical practices one adopts for maneuvering in the face of these provisional, never fully resolved or elaborated assertions of ideals. They are practices of deliberation and decision-making, where no one’s cherished commitments are relegated to the second stage of a lexical ordering, but where one seeks to understand and, if possible, respond to deeply held opinions. Sometimes the disagreement will be so acute as to render imposition inescapable. Sometimes it will be possible to live and let live. Often there will be room for adjustment by all parties – and that adjustment may even represent a normative advance for all parties, as they take into account insights drawn from a broader range of experience.

The abandonment of full-throated consent does mean that secularism’s pretensions to neutrality must be significantly qualified. They must focus on citizens’ equal participation in determining society’s goals rather than insisting that the outcomes be acceptable, necessarily, to all. Something like the following can be achieved: 1) the institutional separation of religious from governmental authority, so that all citizens have, in principle, an equal role in government; 2) the separation of religion from government on the symbolic plane, so that citizens have equal status in relation to the state; and 3) prohibition on the state imposing or discouraging religious beliefs or practices, for religious reasons. To this we should add 4) the state should seek to minimize impositions on beliefs or practices for non-religious reasons, in order to avoid constraints on citizens’ most cherished commitments without sufficient justification, but this will necessarily involve deliberation and judgement. (Taylor, in the essay Bilgrami criticizes, does begin by focusing on equal participation, although, regrettably, he slips at times into language that suggests neutrality of outcome.)

What can’t be achieved is the segregation of religious reasons from public discourse. The effects of comprehensive doctrines, religious and non-religious, are pervasive. They color one’s entire approach to the world. In acting as a citizen, one’s comprehensive commitments will inevitably be operative. Furthermore, there is no way to ensure that religious and non-religious reasons will be equally influential in public decision-making. Any decision involves a preference for some approaches over others; some people will inevitably feel more satisfied than others (as Bilgrami rightly notes).

Secularism is the attempt to preserve the independence of the state from identification with particular religious doctrines so that the state remains open to individuals of different beliefs. But that does not mean that it can be insulated from those beliefs, or that it can satisfy them equally in every decision, or that it must subject them as a matter of lexical ordering to “the ideals that the polity seeks to achieve.” It involves the managing of continual encounters between religiously and non-religiously motivated reasons. The solution involves an ethic of dialogue prior to decision, not peremptory imposition.