Most contemporary liberal political philosophers in this country are kindly disposed toward the idea of religious liberty. But ever since the Reagan era they have been anxious about the rise of the Religious Right, and correspondingly eager to contain the influence of religion over politics. Drawn to the notion that the state ought to be neutral with respect to any controversial conception of what constitutes a good life, they have spawned a number of theories that aim to recast what is politically salient about religion in neutral, nonreligious terms, such as “conscience” or “individual autonomy.” (A recent example of this effort is Brian Leiter’s book Why Tolerate Religion?, reviewed in these pages by William A. Galston, May 3, 2013.) Unhappily, this move toward abstraction discards everything that is specifically valuable about religion, even as it threatens to deprive us of the legal and political tools we need to deal with the problems that religious diversity generates.

For many years, the Canadian philosopher Charles Taylor has been a leading voice for a more historically and ethically grounded approach to the complexities of religious practice in secular society. In his most recent writings, however, Taylor—a Catholic who has written for Commonweal—has started to sound a lot like the advocates of neutrality. In an essay titled “Why We Need a Radical Redefinition of Secularism,” he argues that “all spiritual families must be heard” in the process of social self-determination. And in Secularism and Freedom of Conscience, a 2011 book co-authored with Jocelyn Maclure, he writes that the democratic state must “be neutral in relation to the different worldviews and conceptions of the good—secular, spiritual, and religious—with which citizens identify.” Anything else, he insists, would make some into “second-class citizens.”

These claims can be read two ways. One of these takes them as a call for a negotiated common ground, working from whatever commitments citizens happen to have—religious or otherwise—toward an outcome that will honor all those diverse commitments while adjudicating the inevitable conflicts among them. In present U.S. law, for instance, accommodation is given to ritual peyote use in Native American religious ceremonies but not to recreational use of hallucinogens, and nobody gets permission to use heroin. The other view is that Taylor has joined cause with those theorists who claim that the state ought to be neutral with respect to any conception of the good. Which view of Taylor’s thinking is more accurate?

Secularism and Freedom of Conscience usefully distinguishes primary from subordinate principles—“aims” from “operative modes.” Equal respect and freedom of conscience are the ends; neutrality, separation of church and state, and accommodation are “essential means toward the realization of properly moral ends.” Taylor and Maclure reject “rigid” forms of secularism which, by allowing greater restrictions on free exercise in the name of state neutrality, commit a “fetishism of means.” A prime example is France’s prohibition of headscarves, a policy that, in the name of secularism, has deepened the country’s religious divisions.

The fetishism of means is ubiquitous in discussions of American law as well, and so Taylor and Maclure’s intervention is welcome. Their understanding of the law’s core aims—nondiscrimination among comprehensive views and the accommodation of individual conscience—will be familiar to students of the First Amendment, where they are labeled “disestablishment of religion” and “free exercise.” As Taylor observes elsewhere, there is some tension between these principles. If government must be neutral toward religion, how can religious accommodation be permissible? It is not logically possible for the government both to be neutral between religion and nonreligion and to give religion special protection.

Many liberal political theorists—John Rawls, Joseph Raz, Martha Nussbaum, Christopher Eisgruber, and Lawrence Sager among others—have responded to this dilemma by trying to eliminate the notion of “religion” as any part of the justification for both practices. Like many of these writers, in addressing the free exercise of religion, Taylor and Maclure substitute “conscience” for “religion,” arguing that freedom of religion is “a subcategory of freedom of conscience.” Conscientious convictions are distinguished by the fact that they “occupy a preponderant role in a person’s life.” “Core beliefs” are those that “allow people to structure their moral identity and to exercise their faculty of judgment.” In this light, “the special legal status of religious beliefs is derived from the role they play in people’s moral lives rather than from an assessment of their intrinsic validity.” There is no good reason to single out religious in preference to nonreligious “comprehensive” views; what matters is “the intensity of the person’s commitment to a given conviction or practice.”

Intensity of commitment makes for a slippery hook on which to hang a theory of legal accommodation. Religion, after all, does not hold the same place in the lives of all religious people. A given individual may have no particular commitment to his religion at all—until, say, a crisis in middle age. If commitment is what matters, then what basis exists for protecting spiritual exploration by the merely curious; and what about proselytizing, which necessarily targets the uncommitted? Yet Taylor and Maclure go further, declaring that “it is not enough for the petitioner simply to affirm that he sincerely believes that a given conviction must be translated into action in a particular way; he must also explain why that conviction or value is intimately linked to his moral integrity and in what way he is attempting to respect it or to live up to it.” Given the difficulty even traditional religious claimants have in offering a persuasive account of how their theological views fit together, this seems like a recipe for punishing the inarticulate.

A second basis of accommodation, according to Taylor and Maclure—one whose relation to the first is not clear—is fairness. The trouble they see is that “certain public norms applying to all citizens are not neutral or impartial from a cultural or religious point of view.” For example, public holidays in Canada tend to be Christian in origin, ignoring Jewish or Muslim holidays. Accommodation is thus necessary “to reestablish equity within the terms of social cooperation.” “It is because certain laws or rules are not neutral,” Taylor and Maclure argue, “that accommodations are sometimes justified.” This reasoning will justify some religious accommodations, holidays being the obvious case. But not all mainstream laws reflect a religious background in the way our calendars reflect Christian holidays. Requiring members of the military not to wear hats (which burdens Orthodox Jews and Sikhs) or criminalizing the use of peyote (which burdens some Native Americans) reflects no particular religious view. (Both burdens were lifted by Congress after the Supreme Court refused to mandate accommodation.)

Of course, a legislature with a majority of Jews and Sikhs, or of Native Americans, would not have enacted those laws in the first place. But it is hard to turn this observation into a principle that does not entail anarchy. What would that principle be? That no law is valid unless it could have been enacted by a legislature dominated by the minority religion? Such a principle would give dispositive weight to the preferences of dissenters, whether or not the beliefs it privileged were “core.” Whenever a law is enacted, it reflects a normative orientation, one that won’t be shared by everyone (or else no law would have been necessary).

And then there is the problem of disestablishment. Critics worry that state endorsement of religious symbols turns some people into second-class citizens. Yet the rejection of all such symbols would be a radical move, given that most states in the world support some religions more than others. Taylor and Maclure do not propose dismantling all such symbols, such as the cross on Mount Royal in Montreal. Instead, they envision a state neutrality that is consistent with accepting “practices and symbols that have a heritage value rather than a regulatory function.” Such a “heritage” symbol merely attests to history, they argue, and “is thus compatible with secularism,” since “it is a reminder of the past rather than a sign of religious identification on the part of a public institution.” But this is a hard distinction to sustain. Even if a symbol genuinely is part of a nation’s “heritage,” that doesn’t make it nonreligious. Indeed, its significance as heritage is part of the problem. The Mount Royal cross identifies Canada with Christianity.

The Supreme Court has taken a different approach, in effect adopting a “grandfathering” rule by which old forms of “ceremonial deism,” such as “In God We Trust” on the currency and the reference to God in the Pledge of Allegiance, are deemed constitutionally permissible but newer ones are not. This may be a sensible compromise—this far, but no further—yet it clearly allows for the accommodation of religious symbols with religious meaning, and there is no point pretending otherwise. The United States has done a pretty good job of addressing religious diversity. But it has not dispensed with “religion” as a legally significant category.

It appears that Canada can’t do without religion as a category, either. Its Supreme Court has held that religious freedom protects even heretics: an accommodation claim need not be “in conformity with the position of religious officials.” But in its 1994 ruling defending the right of orthodox Jews to decorate condo balconies with sukkahs—cedar-branch structures commemorating the difficult conditions Jews faced after fleeing Egypt—the Court went on to say: “It is the religious or spiritual essence of an action, not any mandatory or perceived-as-mandatory nature of its observance, that attracts protection.” Taylor and Maclure claim that the criterion employed by the Court was not merely sincerity, arguing that the inclusion of other values within traditional categories such as religion “contributes nothing toward helping us distinguish them from personal preferences.” But that is not what the Court said. It specifically referred to an action’s “religious or spiritual essence” as the reason for protecting it. Taylor and Maclure enlist the Supreme Court of Canada as an ally, not noticing that the Court is playing a game very different from theirs.

The attempt to reduce religious liberty to a few principles has an abstract quality unusual for Taylor. His other work is far more context-sensitive. Sources of the Self (1989) depicted European liberalism as a distinctive set of moral aspirations that emerged out of a specific cultural context. A Secular Age (2007) traced the roots of modern secularism to Christian alienation from religious forms that were considered oppressive, thus helping to explain the militant idealism of many contemporary atheists. These and other historicizing analyses suggested the possibility of many modernities, many liberalisms—and many morally attractive secularisms. In Taylor’s vision, any society’s solution to the problem of religious diversity will result from context-specific negotiation. His 1992 essay “The Politics of Recognition” argued that “a society can be organized around a definition of the good life, without this being seen as a depreciation of those who do not personally share this definition,” and that such a society can be liberal if it respects fundamental liberties.

American religious neutrality is not neutrality toward all conceptions of the good. It proceeds from the premise that government may not take a position on contested religious questions—but that the state is permitted to favor religion-in-general. The American tradition of state neutrality treats religion as a good, not only at the level of practice, but also at the level of justification; it is based on the premise that religion can be corrupted, and its value compromised, if the state is permitted to manipulate it. That’s the answer to the free exercise/establishment dilemma: religion’s goodness can only be acknowledged at an exceedingly abstract level. (I elaborate this point in my book Defending American Religious Neutrality).

The practice of religious accommodation is not and cannot be a deduction from first principles. Under the recent state and federal Religious Freedom Restoration Acts, those who seek accommodation of their religious claims are entitled to individual review by a judge, and the question of whether those claims are trumped by sufficiently pressing state interests in any specific case is a matter of judgment. There are some rules: “The First Amendment...gives no one the right to insist that in pursuit of their own [religious] interests others must conform their conduct to his own religious necessities.” But mostly the work must be done—often it is done—by the judges’ sympathetic perception of what is at stake.

We can take a general lesson from our exploration of “conscience” as an unsuccessful substitute for “religion” in secularism’s effort to accommodate religious liberty. Even if we find a single morally attractive factor that justifies some religious-liberty claims, that factor won’t do as a general basis for all such claims. No single factor for singling out religion can succeed. Any invocation of any factor X—such as conscience—as a justification will logically entail substituting X for religion as a basis for special treatment, making “religion” disappear as a category of analysis. This substitution will be unsatisfactory because it will inevitably narrow what practices can be accommodated. There are many different versions of X on offer. Conscience is only one; commentators have also proposed individual autonomy, psychologically urgent needs, and many more. The same point applies to all of them. Any X will be an imperfect substitute for religion; and a theory of religious freedom that focuses on that X will not be able to say why religion, rather than X, should be the object of solicitude.

There are two ways around this difficulty. One is to say that these are not ends that the state can aim at directly, and that religion is a good proxy for them. This justifies some imprecision in the law. We want to give licenses to “safe drivers,” for example, but since safe drivers are not directly detectable, we use instead the somewhat commensurate category of “those who have passed a driving test.” But this proxy rationale doesn’t work for at least some of the substitutes for religion that are on offer—since the state can aim directly at accommodating conscience, say, or autonomy. So the other solution is to say that religion is an adequate proxy for multiple goods, some of which are not ones that can directly be aimed at. Each of those goods is, at least, more likely to be salient in religious than in nonreligious contexts. The fact that there is so much contestation among religions as to which of these goods is most salient is itself a reason for the state to remain vague about which of the goods associated with religion is most important. Because “religion” captures multiple goods, aiming at any one of them will be underinclusive. That is enough to justify singling out religion.

In sum, the answer to the problem of plurality is not to dispense with the category of religion, but rather to keep that category vague. And in fact this is how the United States has addressed the problem, becoming ever vaguer—productively so—over time. It’s a theoretically untidy solution, but it works. Charles Taylor was once more tolerant of theoretical untidiness than he is now. His move toward abstraction has obvious rhetorical advantages, but they come at a cost in reduced clarity about how to reach the very goals he would urge upon us.