The Yale Law Journal

December 2013

Why Protect Religious Freedom?

Michael McConnell

Why Tolerate Religion?



author. Richard and Frances Mallery Professor and Director of the Constitutional Law Center, Stanford Law School; Senior Fellow, Hoover Institution. The author wishes to thank William Baude, Nathan Chapman, Richard Epstein, Chad Flanders, Robert George, Luke Goodrich, Paul Harold, Joshua Hawley, Steffen Johnson, Burt Neuborne, Eric Rassbach, James Sonne, and Eugene Volokh for helpful comments on an earlier draft, and Spencer Churchill and Mark Storslee for invaluable research assistance.


Religious beliefs have always generated controversy. But religious freedom—the right of individuals and groups to form their own religious beliefs and to practice them to the extent consistent with the rights of others and with fundamental requirements of public order and the common good—has long been a bedrock value in the United States and other liberal nations. Religious freedom is one thing nearly all Americans, left and right, religious and secular, have been able to agree upon, perhaps because it protects all of us.1 Atheists are protected from imposition of prayer and Bible reading in state schools;2 churches are protected from interference with the hiring of ministers;3 religious minorities are protected from majoritarian legislation indifferent or hostile to their concerns.4 Progressive churches are protected when they oppose segregation or counsel draft resisters;5 traditionalist churches are protected when they oppose abortion or operate faith-based schools;6 nontraditional faith groups with unfamiliar worship practices are allowed to carry them out in peace.7 Because none of us can predict who will hold political power, all of us can sleep more soundly if we know that our religious freedom does not depend on election returns.

When the Supreme Court narrowed its interpretation of the Free Exercise Clause in 1990, in the so-called “Peyote Case,” Employment Division v. Smith,8 Congress passed the corrective Religious Freedom Restoration Act (RFRA)9 by unanimous vote in the House and a margin of 97-3 in the Senate.10 Supporters included the ACLU, the National Association of Evangelicals, People for the American Way, the American Jewish Congress, the Christian Legal Society, and virtually every other religious and civil liberties group.11 Recently, however, this consensus seems to be weakening—largely from fallout over culture-war issues such as abortion and the legal recognition of same-sex relationships. Many activists on these issues see religion as antagonistic to their interests, and are responding in kind. A new whiff of intolerance is in the air.12

University of Chicago law professor and legal philosopher Brian Leiter has entered the debate with his new book Why Tolerate Religion?13 His answer? Although we should not persecute religious believers, religion as such does not warrant any “special” legal solicitude such as that provided by the Religion Clauses of the First Amendment.14 “[T]here is no apparent moral reason why states should carve out special protections that encourage individuals to structure their lives around categorical demands that are insulated from the standards of evidence and reasoning we everywhere else expect to constitute constraints on judgment and action.”15 Leiter argues, moreover, that it would be consistent with “principled toleration” for the secular state to affirmatively discriminate against religious believers in access to public spaces, such as by barring student Bible clubs from meeting on public school property, even when every other form of student organization is free to meet.16 So long as religious believers retain the right to express their own beliefs (including wearing religious symbols and clothing), the regime may advocate a “Vision of the Good” that is “irreligious”17 and may selectively deny religious believers and religious speakers equal access to public resources and opportunities.

When it comes to accommodation of practices, and not just beliefs, Leiter argues that it would be impractical to accommodate all claims of conscience and “unfair” and “arbitrar[y]” to single out claims that are grounded in religious belief.18 So his answer: accommodate none of them, at least if the accommodation would inflict harm or shift burdens onto third parties. Exactly what is meant by these assertions, as we shall see, is less than clear. The argument depends on terms like “conscience,” “special,” and “harm,” but the book provides no precise definition of their meanings. The author is vague about what to do when accommodations do not cause harm and when religious practices have no secular analogue.

Organizationally, the book weds four chapters of ambitious and wide-ranging philosophical arguments to a fifth and final chapter primarily addressing two controversial issues of First Amendment law: whether religious practices are entitled to exemptions from formally neutral laws (to which Leiter answers “no”), and whether groups may be excluded from otherwise open public school speech forums because they espouse a religious point of view (to which he answers “yes”).

The first major argument of the book—spread between Chapters One and Four—is that discussions of religious freedom ought to be framed around the concept of “toleration.” By “toleration,” Leiter means protection from coercion (or “eradication”) but something less than neutrality. To be specific, the state may not “jail or annihilate the adherents of the disfavored claims of conscience,” nor may it “directly target or coercively burden their claims of conscience” (absent real harm),19 but it may use public resources and publicly controlled institutions to espouse the state’s own contrary “religious or irreligious” Vision of the Good20 and may exclude dissenters from equal access to public facilities.21 The second philosophical argument—Chapters Two and Three of the book—presents a definition of religion and discusses several prominent justifications for toleration, concluding that none of these theories can justify a special protection for the free exercise of religion, beyond that accorded conduct based on nonreligious beliefs.22 In these chapters, Leiter’s argument consists of two steps. First, he offers a definition of religion as “categorical demands that are insulated from evidence”23—meaning that religion is a phenomenon characterized by insulation from “common sense and the sciences.”24 Second, he examines several prominent justifications for toleration offered by John Rawls, John Stuart Mill, and Frederick Schauer, and in each case concludes that nothing in these justifications warrants tolerating religion specifically.

More surprisingly, in Chapter Five Leiter concludes that this spare doctrine of “principled toleration” also does not justify any special protection against the establishment of religion. As far as “principled toleration” goes, it would be unobjectionable to declare the Roman Catholic Church the established church of the nation, and favor it over all other ideological competitors—so long as dissenting voices are not coercively burdened or silenced. It becomes clear that Leiter’s objection is not to one particular theory of free exercise protections (free exercise exemptions), but to the entire idea of special protection for religious freedom.

At a few extraordinary moments in the book, it appears that the author might even opt for intolerance toward religion—use of the coercive power of the state to discourage or even “eradicate” religious belief,25 on the ground that religious beliefs do real harm to the body politic. Each time, after floating the argument for intolerance, usually in the form of rhetorical questions rather than straightforward claims, he retreats. But each time the retreat is based on the lack of sufficient empirical support for the net harmfulness of religion—not because of the importance of religious freedom to the individual or to liberal democracy. At page 59, for example, he poses the question: “isn’t there reason to worry that religious beliefs, as against other matters of conscience, are far more likely to cause harms and infringe on liberty?,”26 observing that this might “form the basis of an argument for why there are special reasons not to tolerate religion.”27 He follows this suggestion with the tentative disavowal that “I wonder” whether “such a demeaning conclusion about religious belief . . . is warranted,”28 leaving the reader to suspect that his support for toleration hangs on the thread of empirical uncertainty.

And consider this paragraph:

[R]eligious believers overwhelmingly supported George W. Bush, widely considered one of the worst presidents in the history of the United States, whom many think ought to be held morally culpable both for the illegal war of aggression against Iraq as well as the casualties resulting from domestic mismanagement. Of course, if we really thought there were some connection between religious belief and support for the likes of Bush, then even toleration would not be a reasonable moral attitude to adopt toward religion: after all, practices of toleration are, themselves, answerable to the Millian Harm Principle, and there would be no reason ex ante to think that Bush’s human carnage is something one should tolerate.29

If I understand this passage correctly, Leiter is flirting with the idea that it would be justifiable to withhold toleration from religious believers because they have a propensity to support political candidates of whom he disapproves. If that is his notion of “Millian Harm,” sufficient to justify official intolerance toward American religious believers, we are very far from anything recognizable as liberalism or democracy.

Ultimately, Leiter concludes that this “Bush carnage” argument for intolerance is “not warranted,” but not because of any principled commitment to democracy or respect for differing opinion. It is unwarranted because “there is no reason to think” that religious beliefs “are especially likely to issue in ‘harm’ to others.”30 Religion has done good as well as evil, he notes,31 and not all evil is caused by religion.32 Nonetheless, he finds it to be a close question. “Perhaps [religious] beliefs . . . are more harmful, on average, but it seems to me much more empirical evidence would actually be required to support that conclusion.”33 One wishes that the argument for toleration were more robust than that.

Aside from these disquieting passages, the argument of the book rests not on the claim that religious belief is specially harmful, but on the more conventional claim that religion is nothing special.34 Let us turn to those arguments.

I. “toleration”

The title of the book, Why Tolerate Religion?,at first blush sounds anachronistic. The value of religious toleration has not been seriously contested in the Anglo-American world since the seventeenth century. Strictly speaking, the “toleration” issue arose in the context of an established church; the question was whether practitioners of dissenting religions should be permitted to exercise their faiths without penalty. Britain’s celebrated Act of Toleration of 1689, for example, allowed certain sects—dissenting trinitarian Protestants, but not Jews, Roman Catholics, or Unitarians—to conduct worship services without being punished for violation of the Uniformity Acts.35

By the time of the American founding, prevailing opinion had moved beyond toleration. When George Mason proposed in 1776 that the Virginia Declaration of Rights provide for “toleration” of religion, James Madison objected on the ground that “toleration” implies an act of legislative grace. He successfully moved to substitute the term “the full and free exercise of [religion.]”36 In a similar vein, George Washington wrote to the Hebrew Congregation of Newport, Rhode Island that “[i]t is now no more that toleration is spoken of, as if it was by the indulgence of one class of people, that another enjoyed the exercise of their inherent natural rights.”37 It is not an accident that the United States Constitution contains a Free Exercise Clause, not a toleration clause.

But Leiter is not using the word carelessly or anachronistically. He makes clear that by “toleration” he means pretty much what Madison and Washington understood the term to mean—forbearance—with only a slight twist: that the state that today is considering whether to permit the practice of religion is the modern secular state. “[A] genuine ‘principle of toleration,’” Leiter writes, exists only when there is a “dominant group” that “actively disapproves of what another group . . . believes or does.”38 Under his definition of toleration, “one group must deem another differing group’s beliefs or practices ‘wrong, mistaken, or undesirable’ and yet ‘put up’ with them nonetheless.”39 Leiter candidly “reject[s] the view” that “the right posture for the modern state is one of neutrality” toward religion. Rather, the posture of the modern secular state toward religion should be one of “disapproval”40—the only question being whether that disapproval should be tempered with toleration.41

To Leiter, the “contemporary problem, at least in post-Enlightenment, secular nations,” is “why the state should tolerate religion as such at all.”42 Just as the seventeenth-century state, committed to an established church, had to decide whether to tolerate persons of dissenting faiths, the twenty-first-century state, committed to a particular form of secularism, has to decide whether to tolerate religious believers at all. Some might say, following the seventeenth-century philosopher John Locke, that it is futile to attempt to use force to compel belief (or unbelief), because convictions do not yield to external compulsion.43 But Leiter—probably correctly—points out that the modern state in fact has “sophisticated means to effectively coerce belief.”44 The proper question, therefore, is why “we”—meaning the secular state—“morally, ought not to eradicate differing beliefs and practices,” given that we could.45

Leiter claims that “toleration,” understood as putting up with beliefs that the dominant group disapproves of, is “reflected” in the First Amendment of the United States Constitution, and is the “paradigm of the liberal ideal.”46 But this is incorrect: under the United States Constitution, the state does not deem religious belief “wrong, mistaken, or undesirable.”47 On the contrary, our liberal republic takes no stand on the truth or worth of any religious belief as such.48 One of the most widely admired opinions of the Supreme Court states that “[i]f there is any fixed star in our constitutional constellation, it is that no official, high or petty, can prescribe what shall be orthodox in . . . religion, or other matters of opinion.”49 That proscription of official orthodoxy applies to Leiter’s unbelief no less than it does to a conventionally religious establishment. It is no more proper for the state to assume religion is false or unwarranted than to assume that it is true. As James Madison put it, “[t]he Religion . . . of every man must be left to the conviction and conscience of every man,” and it is an “arrogant pretension” to believe that “the Civil Magistrate is a competent Judge of Religious Truth.”50 Indeed, many statesmen at the time of the Founding believed that religious faith was valuable or even essential to republican self-government.51 I am not aware of any statement by a constitutional founder, any decision of the Supreme Court, or any important document in our constitutional tradition that espouses Leiter’s version of toleration: that religion is wrong, mistaken, or undesirable, but we should nonetheless “put up” with it.

Leiter acknowledges there is a competing view to his idea that mere toleration ought to characterize our attitude toward religion. Chapter Four of the book is devoted to explaining why he rejects Professor Martha Nussbaum’s argument that free exercise exemptions are justified by “respect” for the religious beliefs of others, even if we do not share them.52 While the author states elsewhere that his concern is “state toleration” of religion, “as opposed to toleration in interpersonal relations,”53 much of the chapter is propelled by examples of interpersonal relations. The primary argument—spanning seven pages of the twenty-three page chapter—revolves around whether an atheist invited to shabbat dinner should participate in Jewish prayers offered by his host.54 The problem, you see, is that the guest believes that religion is an “(epistemically) culpable false belief,” and therefore is reluctant to show the “respect” to his hosts’ beliefs that participation in the prayers would express.55 That is all well and good. Maybe a guest with those views should politely excuse himself. But what could this possibly have to do with “state toleration”? Leiter argues by way of analogy that because religious beliefs are false or unwarranted, they are “not the kind of belief system that could warrant [affirmative respect],”56 and thus that the state—like the dinner guest—ought rightly to grant religion only toleration. But the state is not in the same position as the guest. The guest, like his host, enjoys full religious freedom to form his beliefs and to act on them. The question he faces is primarily one of etiquette. The state, by contrast, does not hold an official position on whether the Jewish religion, or any other, is false—“culpably” or otherwise. The dichotomy between “tolerating” and “respecting” religion based on individual judgments about religion’s truth or falsity is a red herring when it comes to questions of governance. Constitutional law is not about good manners or respect, but about law, power, coercion, and freedom.

The difference between moral demands on individuals and institutional constraints on the liberal state is fundamental. In the liberal tradition, the government’s role is not to make theological judgments but to protect the right of the people to pursue their own understanding of the truth, within the limits of the common good. That is the difference between “the full and free exercise of religion”57 (Madison’s formulation) and mere “toleration.” Toleration presupposes a “dominant group”58 with a particular opinion about religion (that it is “false,” or at least “unwarranted”),59 who decide not to “eradicate”60 beliefs they regard as “wrong, mistaken, or undesirable.”61

By contrast to Leiter’s “toleration,” religious freedom does not proceed from any official presuppositions about religious truth. It allows everyone, believers and unbelievers alike, the right to form their own convictions about transcendent reality and to live in accordance with them, subject only to the constraint that they must not invade the rights of others or damage fundamental aspects of the overall common good. That is a more attractive vision than Leiter’s, and it is far more consonant with our constitutional principles (even if not always with our practices).

II. the philosophical argument

Leiter’s philosophical argument can be stated in three steps:

1.      Religion is a subset of the broader category, “conscience.”

2.      What sets religion apart as a distinctive subset of conscience is that religious beliefs “are insulated from ordinary standards of evidence and rational justification,” yet lead to “categorical demands on action” (meaning demands that take precedence over competing desires and interests).62

3.      Employing this definition, neither of the two major strands of modern thought, Kantianism (represented by John Rawls) and utilitarianism (represented by John Stuart Mill), supports an argument for special protection for religion. Nor is it supported by Frederick Schauer’s argument from government incompetence.

The argument, however, is not persuasive, for two reasons. First, it depends entirely on the pejorative way in which Leiter defines religion, and second, it falls short in its understanding of Rawls, Mill, and Schauer.

A. Religion as a Subset of Conscience

The central argument in the book is that “if there is something morally important about religious belief and practice that demands legal solicitude, it is connected to the demands of conscience that religion imposes upon believers,” rather than the distinctively religious or “sacred” character of those beliefs.63 The book is an inquiry into whether there is any reason to single out religious conscience for legal protections that are not also extended to nonreligious claims of conscience.

It would therefore seem essential to unpack what is meant by claims of “conscience” as well as what is meant by “religion.” But while the book devotes an entire chapter to a formal definition of “religion,” it provides no definition of “conscience.” I believe, however, that it would be fair to borrow the first portion of Leiter’s definition of “religion” as a definition of “conscience”: a belief system that imposes “categorical demands on action—that is, demands that must be satisfied no matter what an individual’s antecedent desires and no matter what incentives or disincentives the world offers up.”64 Leiter comments that the “categoricity” of commands is a “significant feature” not only of religion, but “of all claims of conscience,”65 and that seems right. In less jargony language, we may translate this “categoricity” as referring to the demands of right and wrong, as opposed to self-interest, whim, habit, or compulsion.66 The important point is that the demands of right and wrong may arise from nonreligious as well as religious systems of belief—although, as Leiter points out, “religion is one of the few systems of belief that gives effect to this categoricity.” According to Leiter, “those who genuinely conduct their lives in accord with the categoricity of the moral demands they recognize are overwhelmingly religious.”67

Leiter defines religion by four criteria, only two of which he says “matter” for the purpose of evaluating the claim for tolerating religion as such.68The first is “categoricity,” as just discussed. The second is that religious beliefs, “in virtue of being based on ‘faith,’ are insulated from ordinary standards of evidence and rational justification, the ones we employ in both common sense and in science.”69 The combination of these two criteria becomes his catch phrase for religion: “categorical demands on action conjoined with insulation from evidence,” or variants on these words.70

Leiter mentions two candidates for further refining the definition of religion: that religion involves “a metaphysics of ultimate reality,”71 and that it offers “existential consolation” for dealing with “the basic existential facts about human life, such as suffering and death.”72 Ultimately, he declines to include these two other elements in his formal definition. The former, he says, is just “a variation on the idea that religious belief is insulated from evidence,”73 and the latter, he says, is not distinctive to religion, but may be found in such nonreligious practices as philosophical reflection, meditation, and therapeutic treatment.74 Thus, “only the first two features [categoricity and insulation from evidence] . . . matter” for these purposes.75 At times, though, Leiter includes “existential consolation” as one of the distinctive features of religion, with the practical effect of “excluding the case of Maoist personality cults, of Marxism, and (probably) of morality.”76

It is difficult to follow Leiter’s method here. At the outset, he asserts that a proper definition must be based on “features that all and only religious beliefs have.”77 But elsewhere, Leiter notes that neither “categoricity”78 nor “insulation from evidence”79 is unique to religion, yet for unexplained reasons these features nonetheless do “matter.” This is all rather confusing and inconsistent. Personally, I think it is futile to draw up a list of features descriptive of religion and only of religion. What makes religion distinctive is its unique combination of features, as well as the place it holds in real human lives and human history.

Leiter assumes that religion is merely a subset of conscience, distinguished primarily by its lack of evidentiary warrant. It would be more precise to see religion and conscience as two overlapping categories, neither fully subsumed within the other. Conscience has to do with convictions about moral right and wrong. Some conscientious convictions have a religious foundation and some do not. Religion is partly about right and wrong, and in that sense overlaps with conscience. But it involves much more than that. Religion typically includes a set of beliefs about the nature of the universe, it prescribes practices that are sometimes more ritualistic than ethical in character (such as taking communion or wearing a yarmulke), and it is embedded in authoritative communities involving texts, stories, institutions, leaders, and tradition. It thus involves much more than conscience, just as conscience comprises more than religion.80 This is important because much litigation involves religious ritual, ecclesiastical form, and tradition that are not strictly matters of “conscience” and have no evident secular analogue.

There are claims of nonreligious conscience that are powerful and coherent enough that they have a moral weight comparable to that of religion. During the Vietnam War, the Supreme Court decided two cases involving conscientious objectors whose beliefs, by their own admission, were not “religious” in the ordinary sense (the sense that Congress used in the conscientious objector statute recognizing exemptions from conscription). The Court’s response was to stretch the definition of religion to include any “sincere and meaningful belief which occupies in the life of its possessor a place parallel to that filled by the God of those admittedly qualifying for the exemption.”81 Similarly, in particular contexts of obvious seriousness, such as protecting medical personnel from being required to perform or assist in abortions, Congress has protected “moral convictions” as well as “religious beliefs.”82

Leiter, however, does not confine the term “conscience” to claims of this serious nature. As his paradigmatic example of nonreligious conscience, Leiter refers to claims based on family tradition and identity,83 which we will discuss in more detail below. He also includes the “lone eccentric, who for reasons known only to him, feels a categorical compulsion,”84 and the Marxist.85 The category is evidently open-ended. If any belief comprising a moral judgment is “conscience,” we would face some wildly counterintuitive claims.

As Leiter points out, “[i]t seems unlikely that any legal system will embrace this capacious approach to liberty of conscience” because it would be tantamount to a “legalization of anarchy!”86 In other words, we do not extend protection to all manifestations of conscience, broadly understood, because we cannot and should not. Accordingly, the United States Constitution provides no protection for liberty of “conscience” as such—although particular manifestations of conscience sometimes receive constitutional protection under other rubrics (freedom of speech, freedom of religion, due process). In fact, although Leiter does not mention it, the drafters of the First Amendment considered using the language of “conscience,” voted it down, and used the term “religion” instead.87 Leiter thinks that was an error. He argues that “conscience” is the morally relevant concept, and it was wrong to single out the religious subset for legal protection.

B. “Insulation from Evidence”

If “categoricity” is the element common to both religious and nonreligious systems of belief making demands on human conduct, it is the “insulation from evidence” that most clearly distinguishes religion in Leiter’s definition, and does almost all the work in his analysis. By smuggling into the definition of religion a feature that makes religious belief seem unreasonable, the book unsurprisingly comes to the conclusion that this very unreasonableness disqualifies religion from a moral claim to special legal solicitude. The conclusion is baked into the premise.

Most obviously, Leiter’s definition stacks the deck by assuming that religious belief “always” is to some degree “false, or at least unwarranted.”88 That is a sectarian premise, predicated on a questionable view about evidence.89 According to Leiter, the “only epistemically relevant considerations” that warrant belief are “those that figure in common sense and the sciences.”90 He goes so far as to say that philosophic attempts to justify religious beliefs are “nothing more than an effort to insulate religious belief from ordinary standards of reasons and evidence in common sense and the sciences, and thus religious belief is a culpable form of unwarranted belief given those ordinary epistemic standards.”91

No religious believer would recognize this description. Religious believers do not think they are “insulating” themselves from all the relevant “evidence.” They think they are considering evidence of a different, nonmaterial sort, in addition to the evidence of science, history, and the senses. It would be more accurate, and less loaded, to amend this second part of Leiter’s definition to say that religion is a system of belief in which significant aspects are not based on science or common sense observations about the material world.

To begin with, much religious thought is not “insulated” at all. Developments in biology, physics, linguistics, archeology, and other disciplines have had profound impact on Biblical hermeneutics and theology in mainstream Protestantism and Roman Catholicism,92 and “practical reason” has played a major role in natural law thinking since at least Thomas Aquinas.93 To be sure, some religious traditions are more insulated from scientific developments than others. The Navajo creation story, for example, is impervious to archeological and linguistic evidence that the tribe migrated to the Southwest from Canada only a few centuries before the arrival of Europeans, and fundamentalist Christian belief in the historicity of Noah’s flood and the literal six-day creation, depending on how these ideas are understood, is much the same. But to say that “insulation from evidence” is a defining characteristic of “all”94 (or even most) religions is simply false. Religion is constantly changing, and constantly interacting with the culture and other ways of understanding the world.

More importantly, the standards established by the scientific revolution, however powerful within their proper domain, are not obviously applicable to such matters as esthetics, morality, values, love, trust, and ultimate meaning. The scientific method does not claim to provide insight into these areas of human understanding. Indeed, some philosophers of science maintain that even science depends on certain leaps of faith, which are not the products of mere observation of material evidence.95

In a footnote, Leiter acknowledges that “of course” there may be matters such as the “meaning of life” that “are insulated from evidence only in the sense that no scientific evidence would seem to bear on them.”96 But he immediately dismisses the importance of this observation on the ground that “[s]uch beliefs are not my concern here, mainly because they are not distinctive to religion.”97 What could he be thinking? His entire argument is built around the idea that religion is “a culpable form of unwarranted belief” precisely because of its “insulation from evidence.”98 If it turns out that religion’s “insulation from evidence” is attributable to the fact that “no scientific evidence bears” on many questions of a religious nature, then religious belief cannot be criticized on these grounds. There is no reason to apply the “ordinary epistemic standards” of science and material observation to questions on which they do not bear. If Leiter is confining his “concern” to beliefs on which “scientific evidence would seem to bear,” he is leaving out most of what is central to religion, including beliefs underlying almost all claims of religious conscience, which are the subject of this book.

Leiter is entitled to confine himself to whatever categories of evidence may strike him as persuasive, but he cannot reasonably label as “culpable” or “unwarranted” the sincere conclusion of many persons, including thinkers of the first rank, that there are nonmaterial aspects of reality supporting religious belief. Leiter can no more disprove the existence of nonmaterial reality than religious believers can prove the existence of God on the basis of material evidence alone. A color-blind person might think the idea of color is bunk, because the evidence of his own eyes fails to reveal it, but that does not entitle him to assume that those who see color are engaged in a culpable form of unwarranted belief. He, not they, might be the one lacking.

As individuals, we might be justified in dismissing the idiosyncratic beliefs of small numbers of persons, especially when these people do not appear rational in other respects. But religious belief has been attested to by millions of seemingly intelligent and rational people over long periods of time, who report that they have experienced, in some way, transcendent reality. There is even, as Leiter admits, a “large literature in Anglophone philosophy devoted to defending the rationality of religious belief.”99 Leiter chooses to disregard this testimonial evidence, along with its philosophical defense, without so much as “address[ing] . . . in any detail”—really, at all—the arguments that are offered.100 Why? The only reason he supplies is that the “dominant sentiment among other philosophers” is that belief in God is “unsupported by reasons and evidence.”101 With all respect, there is no reason to think that members of modern philosophy faculties have any special insights about God. But as we shall see, if you take away Leiter’s conceit that religious believers are culpably insulating themselves from evidence (as opposed to responding to a different kind of evidence), most of Leiter’s conclusions fall of their own weight.

Indeed, even for those who agree with Leiter as a matter of personal conviction that there is no persuasive evidence supporting the truth of religious belief, but agree with Madison and Washington that the truth of religion is not a subject on which the government should take a stand, Leiter’s conclusions do not follow, because they rest on the view that the state should treat religious beliefs and arguments as lacking evidentiary warrant. It is better to proceed on the premise that people may reasonably disagree about the truth or falsehood of religious claims.

C. Rawls

Having offered a definition of religion, the next step in Leiter’s argument involves asking whether unique toleration for religion can be justified by several prominent arguments for toleration. Turning first to the Kantian (or better, “neo-Kantian”) argument, Leiter adverts to John Rawls’s well-known thought experiment in which we choose fundamental principles of justice under which we should be governed as if from behind a veil of ignorance—meaning we do not know what our circumstances (including our moral and religious views) will be, or whether we will be in the majority or minority.102 Rawls concludes that while behind the veil we would choose to protect an equal liberty of conscience. We would not “take chances with [our] liberty,” because the value of being able to form and follow our own moral and religious beliefs outweighs any gain we might achieve from the possibility of being in a majority and imposing our views on others.103 This supports legal protection for freedom of conscience.

In a revealing aside, Leiter questions whether Rawls is correct that people really are better off when “they can freely choose what to believe and how to live.”104 He suggests that “many, perhaps even most” people “make foolish choices about what to believe and how to live,” with the result that they make themselves “miserable.”105 Indeed, these people may “perhaps” not “make real choices at all,” but instead they may be “hostage to social and economic milieux,” which produce only the “illusion of choice.”106

This line of reasoning is ironically reminiscent of the seventeenth-century Puritan preacher John Cotton, an opponent of religious toleration in his day. Cotton argued that the

[f]undamentals are so clear, that a man cannot but be convinced in Conscience of the Truth of them after two or three Admonitions: and that therefore such a Person as still continueth obstinate, is condemned of himself: and if he then be punished, He is not punished for his Conscience, but for sinning against his own Conscience.107

Cotton, like Leiter, thinks that those who disagree with him on the fundamentals are “culpably” wrong, that their foolish ideas will render them miserable for eternity, and that their mistakes are the product of something other than sincere conscience.

Having offered these authoritarian musings, cautiously cushioned in the form of questions rather than assertions, Leiter then disposes of them in this way: “These illiberal thoughts . . . have little purchase these days within the mainstream of English-speaking moral and political theory, though not, as far as I can tell, because they have been refuted systematically.”108 Readers must wonder whether in an environment less constrained than the English-speaking mainstream, Leiter would attempt to pursue these “illiberal thoughts” more seriously, and what his answer would be. In any event, Leiter “put[s] these doubts to one side” and accepts, “[f]or the sake of argument,” Rawls’s conclusion that behind the veil of ignorance we would choose to protect the liberty to form and follow our own beliefs.

He then gets to his real argument. He points out, correctly, that Rawls explicitly includes “moral” along with “religious” obligations in his analysis, and thus that nothing in Rawls’s argument is “specific to religion.” Leiter concludes, therefore, that “the Rawlsian perspective cannot help us evaluate the principled case for toleration of religion qua religion.”109

This is too quick. To be sure, Rawls does not explicitly address whether his thought experiment could be used to evaluate constitutional protections for religion qua religion, but it might. Behind the veil of ignorance, we do not know whether we believe in a supreme authority or not, but if we do, by definition belief in a supreme authority creates obligations superior to all others—in Madison’s words, “dut[ies] . . . precedent both in order of time and degree of obligation, to the claims of Civil Society.”110 Leiter himself recognizes that religious beliefs involve issues of “ultimate reality,”111 meaning “the aspect of reality that is most important for valuable/worthwhile/desirable human lives.”112 From a Rawlsian perspective, from behind the veil of ignorance, there is every reason to protect our capacity to pursue that which “is most important for valuable/worthwhile/desirable human lives.” Indeed, as a matter of historical experience, many hundreds of thousands of real people have regarded their religious beliefs as so important that they sacrificed their lives, fortunes, social standing, opportunities for career advancement, and bodily comfort in order to worship in accordance with their convictions, in the teeth of official hostility and persecution. Their testimony counts for something.

The freedom to carry out our perceived religious obligations in the face of political opposition might be more important than the freedom to carry out our personal conclusions about right and wrong, for two reasons. First, we might think that adherence to the supreme authority of the universe is an ontologically superior obligation to adhering to what we, as fallible persons, might conclude about morality. That is what “the sovereignty of God” would seem to entail. Leiter himself comments that religion is one of the “few systems of belief” that actually “gives effect” to convictions about morality—that “those who genuinely conduct their lives in accord with the categoricity of the moral demands they recognize are overwhelmingly religious.”113 Maybe there is a reason for that. Second, we might think that it is essential for governments to make and enforce moral judgments, even in the face of differences of opinion, but think it not essential for governments to make and enforce religious judgments. It is impossible to conceive of a government that does not enforce norms of right and wrong, but not at all difficult to conceive of a government that leaves religious judgments to individual conscience. At least behind a veil of ignorance, we might think these things, and might think they warrant distinctive constitutional protection for freedom of religion.

What is Leiter’s answer? He says that “it is hard to see how persons in Rawls’s original position, operating behind the ‘veil of ignorance,’ could reason, in particular, about the value of insulation from evidence and the categoricity of demands, let alone existential consolation.”114 This is an ipse dixit, not an argument. It is nothing but an arbitrary exclusion of religious belief (defined in Leiter’s pejorative way) from the original position. The whole point of the original position is that the parties behind the veil of ignorance “do not know, of course, what their religious or moral convictions are.”115 These might include a belief in God. And if that is a possibility, a party in the original position might think it is special and worth protecting, even if Leiter does not.

D. Mill

Leiter’s argument regarding Millian utilitarianism is even more problematic. Mill argued that we can discover truth, or be fully persuaded of the truth, only if we are exposed to a wide range of beliefs and practices—even if some of them are false. As Leiter explains, “truths about how we ought to live”116 support “a wider scope of toleration, one that encompasses practices, not just beliefs.117 Mill’s argument thus supports free exercise of religion, and not just speech or opinion about it.

Leiter, however, says “we can dispense with [these] epistemic arguments for toleration . . . quickly,” because “[t]here is no reason to think, after all, that tolerating the expression of beliefs that are insulated from evidence and reasons—that is, insulated from epistemically relevant considerationswill promote knowledge of the truth.”118 In other words, because religious belief is totally without evidentiary warrant, it cannot possibly contribute to the search for truth. He fits religious belief into an apparent exception Mill draws for mathematical mistakes, where there is “nothing at all to be said on the wrong side of the question.”119

The argument thus depends on Leiter’s tendentious claim that religious believers “insulate” themselves from evidence as opposed to recognizing nonmaterial evidence of a sort that Leiter does not recognize. Put aside Leiter’s personal convictions about the falsity of religion, which reasonable people need not and the liberal state must not accept, and his argument here fails. Leiter pretends to be arguing from Mill, but in fact he is arguing the opposite of Mill. I quote the passage Leiter cites from On Liberty, butin its entirety:

But, some one may say, “Let them be taught the ground of their opinions . . . .” Undoubtedly: and such teaching suffices on a subject like mathematics, where there is nothing at all to be said on the wrong side of the question. The peculiarity of the evidence of mathematical truths is, that all the argument is on one side. There are no objections, and no answers to objections. But on every subject on which difference of opinion is possible, the truth depends on a balance to be struck between two sets of conflicting reasons. Even in natural philosophy, there is always some other explanation possible of the same facts; some geocentric theory instead of heliocentric, some phlogiston instead of oxygen; and it has to be shown why that other theory cannot be the true one: and until this is shown, and until we know how it is shown, we do not understand the grounds of our opinion. But when we turn to subjects infinitely more complicated, to morals, religion, politics, social relations, and the business of life, three-fourths of the arguments for every disputed opinion consist in dispelling the appearances which favour some opinion different from it.120

Leiter takes it to be a faithful extension of Mill’s position that science, like mathematics, presents a way of knowing that ought to be applied to all our beliefs because its empirical force has demonstrated its superiority beyond all argument. But Mill thinks nothing of the sort. For Mill, mathematics is a unique domain of knowledge precisely because the “peculiarity of [its] evidence”—namely, the way that mathematical evidence is not susceptible to objections or answers to objections—renders argument superfluous. Science (“natural philosophy”), by contrast, is just one more place where argument and competition among positions is needed in order to determine and justify our beliefs. And for subjects “infinitely more complicated” than science, such as religion, Mill regards the clash of various epistemic positions as even more essential.121Neither science nor religion can be resolved by dogmatic appeals to authority or the pretense that there is only one side to the question.

Here again Leiter reveals himself as the Anti-Mill. Take his reference to “[t]he large literature in Anglophone philosophy devoted to defending the rationality of religious belief.”122 One might think Leiter would wish to engage with the ideas in this literature, in a Millian spirit, if only to prove why they are wrong. But no. Leiter says it “[s]uffice[s] to observe that its proponents are uniformly religious believers,” and that “much” of this literature has the air of “post-hoc . . . rationalizations.” He then resorts to authority—to the “dominant sentiment among other philosophers,” which, he reports, is on the other side.123 The first avenue of attack is a tautological ad hominem. It is neither surprising nor disqualifying that philosophers who find religious belief rational are likely to be believers, just as philosophers who take the opposite view are likely to be nonbelievers. What does that prove, other than that there is a difference of opinion? The second avenue of attack—Leiter’s appeal to the dominant sentiment among supposed experts—is both elitist and authoritarian, in precisely the sense that Mill condemned. Religious ideas should not be put to a vote, not even of philosophy PhDs.

Careful readers will also note the inconsistency between Leiter’s use of Rawls and his use of Mill. Rawls does not comment on whether his theory would support a special role for religious freedom, beyond that due to secular moral beliefs, and Leiter takes this as tacitly rejecting such a role124—even though we can construct an argument, fully consistent with Rawls’s methodology in A Theory of Justice, that supports such a role. Mill explicitly states that his theory applies with particular force to religion, yet Leiter argues that it does not, employing an argument from expert authority that Mill would never accept. How can this be reconciled?

E. Schauer and “Governmental Incompetence”

Leiter also touches too quickly on one other argument—Fred Schauer’s “argument from governmental incompetence,” made in defense of the freedom of speech.125 Even on the assumption that speech sometimes causes real harm that outweighs any possible benefit, Schauer argues that there is no reason to think that the government will make the right choices about what speech to regulate. Politicians are likely to suppress speech when it advances their own political interests, which is unlikely to coincide with the suppression of speech that causes the most net harm. The same argument can be made about freedom of religion, with even greater force. A cornerstone of the American constitutional tradition of religious freedom is the view—held by all stripes of religious opinion—that the government has no competence to judge religious truth.

Public schools can teach all kinds of nonsense, and people may not like it, but they confine their objections to ordinary channels. When public schools purport to teach religious truth, by contrast—for example, by allowing a prayer at a graduation—it is a constitutional case of the highest order. As a supporter of the Court’s School Prayer Cases,126 I have gone on Christian talk radio to defend the prohibition of collective spoken prayer in school. The natural impulse of the audience tends to be to defend prayer, but when I explain that agents of the government should not be entrusted with the power of teaching our children how and what they should pray, even the most fervent believers usually come to see the wisdom of the decisions. Whatever our views on religion, no one trusts the government to guide our spiritual lives. That is what Schauer is getting at. Even if we would not be principled disestablishmentarians in a theoretical world where government officials are theologically trustworthy, the actual fact of government incompetence is good reason to deny them the power to guide the religious life of the nation.

Leiter’s reaction to Schauer’s argument is brief and baffling. After devoting almost two pages to explaining Schauer’s argument, Leiter offers one (long) sentence in response. He says: “Perhaps this kind of instrumental argument for state toleration is more plausible,” but “it does not tell us why we, morally, ought not to eradicate differing beliefs or practices, it tells us only that we (through the instrumentality of the state) are unlikely to do it right.”127 That is not much of a response.

To begin with, Schauer’s is not just an “instrumental” argument. It goes to the heart of the matter. Government is not omnicompetent. It has a large, but limited, role in human affairs, limited to matters where collective coercive action is necessary and likely salutary. It makes no sense for a people to give its government powers that are outside its competence, and it makes no sense to talk about constitutional design on the assumption that government will always exercise its power intelligently and beneficently. Second, what is wrong with an instrumental argument? We might erect constitutional barriers to governmental action because the activity we are protecting is especially important to the individual or to society, and we might erect constitutional barriers to governmental action because the power we are limiting–the power to “eradicate differing beliefs or practices”–is especially inappropriate to government, or susceptible to abuse. Leiter offers no reason why the latter is less persuasive a reason than the former.

That the state is “unlikely to do it right” is evidently not, to Leiter, a deeply serious objection. He operates on an abstract plane where a magisterial “we”—those who share his own convictions and prejudices—control the levers of power. The entire book is about what this infallible “we” should do about religion. The American tradition of constitutionalism, though, proceeds on the premise that “enlightened statesmen will not always be at the helm”128 and that the “Civil Magistrate is [not] a competent judge of Religious Truth.”129 Leiter’s dismissal of Schauer’s argument misses this important point.

III. the legal arguments

The conclusion of the first four chapters is that there is no “principled argument that picks out distinctively religious conscience as an object of special moral and legal solicitude.”130In Chapter Five, entitled “The Law of Religious Liberty,” the author applies that theoretical conclusion to two practical issues of First Amendment law: whether persons whose religious beliefs conflict with neutral and generally applicable laws are entitled to exceptions or accommodations, and whether state institutions such as schools may deny religious groups equal access to otherwise generally available public resources. As to the first, he argues that singling out religious claims of conscience would be “unfair” because it “arbitrarily selects some subset of claims of conscience for special consideration,”131 although he leaves room for exemptions that would not shift burdens onto others. As to the second issue, Leiter argues that it is “consistent with principled toleration” for the government to discriminate against religious views of which it disapproves and to exclude them from equal access to public property and resources, particularly in the schools132—though he is careful to insist that this discrimination must not extend to “persecution” or the imposition of “coercive burdens.”133

Putting these two positions together, religious beliefs and practices may not be given “special moral and legal solicitude,” but they may be subjected to special civil disadvantages and exclusions. It is “arbitrar[y]” and “unfair” to single out the religious “subset of claims of conscience” when this would protect the religious but not when this would disadvantage them. What theory could support these two conclusions?

A. Free Exercise Exemptions

Professor Leiter’s rejection of free exercise exemptions bears strong superficial similarity to the Supreme Court’s still-controversial 1990 decision, Employment Division v. Smith.134 In an opinion by Justice Antonin Scalia, the Court held that members of the Native American Church have no constitutional right to use the drug peyote in their religious ceremonies, because the Free Exercise Clause provides no protection against neutral laws of general applicability. Leiter, similarly, argues that “there should not be exemptions to general laws with neutral purposes, unless those exemptions do not shift burdens or risks onto others.”135

There are three important differences, however, between Leiter and the Court. First, Leiter’s rationale is entirely different from, even contradictory to, the Court’s. Leiter bases his opposition to exemptions on his belief that it would be “unfair” and “arbitrar[y]” to protect religious beliefs if it is not feasible to extend the same protection to nonreligious claims of conscience.136 The Court, by contrast, primarily based its opposition on the jurisdictional impropriety of allowing judges to weigh religious needs against the importance of governmental purposes.137 The Court did not think it improper for the First Amendment to single out religion, and in fact even stated that legislative exemptions for religious practices may be “desirable.”138

Second, Leiter excludes from his “no exemptions” rule cases where the exemptions would not shift burdens or risks onto others. The Smith Court recognized no such limitation. In fact, because the ceremonial use of peyote does not harm others, Leiter seems to conclude that Smith itself was wrongly decided.139 Once this exclusion is taken into account, Leiter advocates a far broader scope for free exercise exemptions than the general rhetoric of the book suggests—almost as broad, perhaps, as pre-Smith interpretations of the Free Exercise Clause.

Third, Leiter recognizes the danger that a no-exemptions regime might “open the door to state conduct motivated by antireligious animus, but under the pretense of legitimate, neutral objectives.”140 The Smith opinion seems oblivious to that problem. The problem is especially serious because it is exceedingly difficult for courts of law to ferret out official pretense. Leiter argues that

if we had reason to think that it will be impossibly difficult to discriminate between the facade of neutral purpose and actual neutral purpose in legislation that burdens religion—then we might think exemptions for religious claims of conscience the preferable approach, notwithstanding the inequality such an approach entails and notwithstanding the burden on the general welfare.141

That is a generous concession, though it ultimately appears not to sway him from his no-exemption position.

Leiter frames his discussion of the exemptions issue around one illustrative case: whether baptized adherents of Khalsa Sikhism, who have a religious obligation to carry a ceremonial dagger, or kirpan, should be exempted from general school regulations prohibiting students from carrying weapons. Leiter compares this Sikh believer to a hypothetical “rural boy” of the same age whose “family traditions and upbringing” call for him to carry a knife as a symbol of his identity as a man in his community.142 He asks us to think about what should be done in the two cases. Should both boys be exempted? Neither? Only one of them?

Leiter says there “can be no doubt” that his hypothetical rural boy’s felt need to carry a knife is a “conscientious obligation”—indeed, an “equally serious obligation[] of conscience” to that of the Sikh.143 In my opinion, far from there being “no doubt” about this, the idea that the rural boy has a conscientious obligation comparable to the Sikh is highly dubious. Strictly speaking, conscience is an individual’s judgment about right and wrong—such things as not killing innocent persons, telling the truth, and caring for your children. It strikes me as very unlikely that the hypothetical rural boy believes that his family’s tradition of knife carrying is a moral obligation of this nature. It may take away from “who he is” to deny him the right to carry a knife, but it does not make him commit a wrong. There are many practices tied up in ethnic or familial identity that are not moral in nature. This does not make them unimportant, but it does put them in a different category than that of “conscience.” A religious dictate, by contrast, is more than a question of identity; it is a duty.

As Leiter points out, “no Western democracy” would recognize a legal right on the part of the rural boy, though it is easy to imagine that rules against pocket knives might not always be rigorously enforced, especially in rural communities where knives are commonplace and useful. The Sikh, by contrast, has a good case. In the leading decision in the United States, Cheema v. Thompson,144 the Ninth Circuit approved an arrangement under which the Sikh student was exempted from the “total ban” on “weapons,” provided his kirpan had a dull blade of only 3-3 ½ inches, was sewn into its sheath, and was worn under his clothing so as not to be plainly visible.145 (The dissenter’s main point of disagreement was to think that the blade should be still smaller, and riveted to the sheath.146) In other cases, courts have found that kirpans are not “weapons” at all, in light of their design and ceremonial purpose.147 Leiter focuses on a Canadian Supreme Court decision, Multani v. Commission scolaire Marguerite-Bourgeoys,148 in which the Sikh student was allowed to wear his kirpan without these protective limitations.

One might expect Leiter to say that these cases were wrongly decided at a level of principle, because it would be unfair and arbitrary to allow the Sikh student to wear a kirpan when the rural boy has no right to carry a knife. After all, the rural boy’s conscientious claim is “equally serious” to the Sikh’s.149 But that does not seem to be Leiter’s view. “Certainly,” he says, “the state should tolerate the various religious practices of Sikhs under the general rubric of liberty of conscience.”150 Apparently this is so even though no Western democracy protects the right of the rural boy to wear his knife, and Leiter does not argue that they should. Leiter’s reservation about the decisions—and it is an entirely reasonable one, even if I might come out the other way—is that the courts in the kirpan cases gave insufficient weight to the risk of harm to others. If the equality objection (no exceptions for religion unless there would be an exception for secular conscience) were dispositive, Leiter would not need to discuss the risk of harm. Leiter concludes that “both boys should be out of luck,”151 but that is because he thinks an exemption in this context would create a risk of harm to others. If the harm could be minimized or eliminated—as the Ninth Circuit thought it could, through the protective conditions—it appears that only the rural boy would be out of luck.

Note what has happened to Leiter’s argument. When it comes down to the real case of the Sikh boy and his kirpan, the “culpable falsity” of religious belief drops out of the calculus, and the analysis shifts to what he calls the “side-constraint” of not allowing harm to others. Leiter’s position turns out to be “that there should not be exemptions to general laws with neutral purposes, unless those exemptions do not shift burdens or risks onto others.”152 Another way to put this is: “There may be exemptions to general laws with neutral purposes unless those exemptions shift burdens or risks to others.” The real point of difference then becomes: How much burden? How much risk?

It appears that in cases where the Millian Harm Principle is not violated by an exemption, Leiter’s rhetorical case against “special” solicitude for religion turns out not to matter very much, if it matters at all. As Leiter understands, his hypothetical rural boy’s perceived need to carry a knife will not and should not receive legal protection. This is not because the law is hostile or indifferent to nonreligious claims of conscience. It is because the claim is too broad, too undefined, too unfocused to be enforceable as a legal right.

As Leiter recognizes, this practical problem of open-ended subjectivity is not true of religious claims, at least not to the same extent. “After all,” he points out,

a litigant who asserts a claim of religious conscience must reference a religion. Religions typically have texts, doctrine, and commands . . . . Membership in the religion in question usually depends . . . on participation in practices, rituals, and ceremonies. All of this gives the courts a rich evidential base for assessing the genuineness of a claim of conscience.153

These practical differences lead Leiter to the uncomfortable thought that “perhaps we should simply extend legal protection for liberty of conscience only to claims of conscience that are rooted in communal or group traditions and practices that mimic, from an evidential point of view, those of religious groups.”154 After a few pages debating the “unfairness of such inequality,”155 Leiter concludes that “the inequality of treatment of claims of conscience is not necessarily fatal to a scheme of universal exemptions for claims of conscience.”156 Translation: it is permissible, after all, to single out religious claims and those nonreligious claims that “mimic” religious claims, and to give them special solicitude.

We cannot know how different this revised position is from the pure protection of free exercise of religion without knowing how often nonreligious claims “mimic” religion in this sense. This might well be a very small category. Leiter himself observes that “those who genuinely conduct their lives in accord with the categoricity of the moral demands they recognize”—which presumably is the heart of the matter—“are overwhelmingly religious.”157 As already noted, the Supreme Court found that the claims of two Vietnam-era conscientious objectors were close enough that they warranted religious exemptions,158 but there have been no others. In the very situation Leiter uses to illustrate the problem—the Sikh and the rural boy—the latter claim does not sufficiently “mimic” the former to warrant legal protection.

Having talked himself out of the claim that “the inequality of treatment of claims of conscience is . . . fatal to a scheme of universal exemptions for claims of conscience,”159 Leiter drops the subject with no further discussion—until the very end of the book, when he returns to the position that the “selective application” of toleration to the conscience of only religious believers “is not morally defensible.”160 Even then, however, he equivocates. He reiterates his support for a “No Exemptions approach . . . to claims of conscience that are burden-shifting,”161 but as to non-burden-shifting exemptions, it still appears to be his position that religious claims and those that “mimic” religious claims are entitled to exemptions.

How different that is from the current regime depends entirely on what he means by “mimicking” religion and what counts as harm. On the actual practical meaning of those key ideas, Leiter says almost nothing.

B. Harm

The question of free exercise exemptions thus turns out not to hinge on the philosophical arguments of the first four chapters, but instead on the application of the Harm Principle. Some religious exemptions entail harm or the risk of harm to third parties, and some do not. In the former category Leiter gives as examples “exemptions from zoning regulations for religious institutions, exemptions from mandatory vaccination schemes, or exemptions from a ban on knives in the schools.”162 In the latter category are such exemptions as “the right to wear certain religious garb, or to use certain otherwise illegal narcotics in religious rituals.”163 So, the Peyote Case was wrongly decided after all.

The analysis of free exercise claims has always taken harm to third parties into account. Madison wrote that the free exercise of religion should prevail “in every case where it does not trespass on private rights or the public peace.”164 Most of the early state constitutions protected the exercise of religion up to the point that it endangered public peace and good order.165 Prior to the Peyote Case, free exercise litigation turned almost entirely on questions of harm,166 and cases governed by the Religious Freedom Restoration Act167 and the Religious Land Use and Institutionalized Persons Act168 employ that same standard.169 But the idea of “harm,” or of “burden-shifting,” is not self-defining. If these are to be useful legal concepts, courts must be able to make defensible judgments both about degree and nature of harm.

Unfortunately, neither courts nor scholars have given serious analytical attention to what counts as “harm.” Leiter quotes John Rawls as saying that liberty may be limited “to prevent an invasion of freedom that would be still worse,”170 which implies some sort of weighing or balancing of harms, to determine which is “worse.” That is not easy to do with any consistency or predictability. And Leiter refers many times to John Stuart Mill’s Harm Principle, according to which “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.”171 But that statement merely begs the question: What counts as harm?

There are, of course, many easy cases. Leiter mentions that “the state need not tolerate . . . killing the infant children of the alleged heretics.”172 No one will argue with that. But what about parents who make decisions about their children’s upbringing that others—maybe experts, maybe majorities—think are deleterious? A test case might be Wisconsin v. Yoder,173 where the Supreme Court held that Amish families have a free exercise right not to send their children to school after the eighth grade. Was that “harm” in the Millian sense? What about prisoners whose religious practices—for example, a kosher diet—increase the cost to the taxpayers?174 Is that “harm”?What about slitting the throats of chickens and sheep in a religious ceremony?175 When members of three small California Indian tribes sought to block construction of a logging road through their sacred places in a national forest, was the loss of the economic benefits to the logging companies a Millian “harm”?176

Outside the context of free exercise claims, we do not live in a Millian world. A great deal of modern legislation coercively adjusts the burdens and benefits of life, helping some at the expense of others, in ways that Mill presumably would not approve. Many modern free exercise controversies arise in the context of social and economic regulation that coerces transactions and dictates their terms. In our post-Lochner, which is to say post-Mill, world, if the problem is merely economic redistribution, there is generally no constitutional obstacle to these schemes. But what if the regulatory scheme demands a violation of conscience? From the baseline of the regulatory requirement imposed on everyone, an exemption for one individual can be said to “harm” the intended beneficiaries of the law, because they will not receive the benefit. But from the standpoint of the Millian Harm Principle, an exemption to such regulation merely returns the parties to the position they occupied before law coercively intervened.

For example, in the contraceptive mandate cases, the government has decided to shift the cost of obtaining contraceptives (including abortifacient drugs) from the user to her employer, through a mandatory term in the health insurance contract. There is nothing constitutionally objectionable about that for most employers, but what about those for whom providing abortifacients is a violation of conscience?177 Leiter objects to “burden-shifting” religious exemptions, but what if the burden-shifting goes the other way, and the grant of an exemption would return the parties to a clearly constitutional status quo ante?

For another example, a wedding photographer in New Mexico, Elaine Huguenin, declined to provide her services to a lesbian couple, out of the conscientious belief that same-sex marriages are contrary to God’s will.178 The couple easily found another wedding photographer. Were they harmed by Elaine’s refusal to film their nuptials? If Elaine had declined their business because she had another booking, or because she was going on vacation, no one would think they were harmed. It would appear that the only real “harm” was the communicative impact of Elaine’s action—the feeling of offense experienced by the lesbian couple because of Elaine’s reasons. In other contexts, the Court routinely holds that people may not be punished because others are offended by what they say. Yet state officials fined the photographer for her refusal, and the New Mexico Supreme Court recently upheld the fine as constitutionally legitimate.179 Should we treat offense as “harm” in the context of a free exercise claim for exemption?

In an intriguing footnote, Leiter says that “to exempt Catholic priests from performing gay marriages would not be a burden-shifting exemption as long as gay couples can otherwise be married.”180 From the perspective of harm or burden-shifting, that example is not different in any meaningful way from the Elane Photography case, unless the priests’ religious status is the driving factor (meaning that religion is “special” after all). Leiter purports to distinguish the case of a Catholic pharmacist who objects to dispensing morning-after pills on the ground that “depending on the community at issue and the availability of the relevant medicines,” this could be burden-shifting.181 I say “purports” because in the two litigated cases involving such pharmacists, in Illinois and in Washington State, the evidence showed that conscience exemptions did not have meaningful effects on patient access to the drugs.182 Thus, it would seem that, in Leiter’s view, the government should not be able to enforce public accommodation requirements or universal service obligations against service providers with conscientious objections, except in the rare circumstance where the service would not otherwise be available.

One more example: How does the Millian Harm Principle apply to the hiring of ministers by churches?183 Title VII of the Civil Rights Act of 1964 gives everyone a right to obtain employment without discrimination based on sex.184 If a woman goes to seminary and is otherwise qualified for an available position, can an Orthodox synagogue refuse to hire her as a rabbi? Who is shifting burdens onto whom, and relative to what baseline?

Questions of this sort will dominate free exercise litigation for the next decade or two. My sense is that very few free exercise claims seek authorization to invade the private rights of third parties or to inflict harm (in the Millian sense) upon them. Most, instead, resist the blanket enforcement of regulatory schemes that interfere with natural liberty in a way that, in some cases, also burdens conscience. Leiter does not say much about these situations outside of footnote 11 to Chapter Five,185 but that footnote suggests that the logic of his arguments may be more supportive of these claims for exemption than the more generalized rhetoric of the book would suggest.

C. Establishment of a “Vision of the Good”

The book closes with an argument that the establishment of religion is not inconsistent with Leiter’s conception of “principled toleration.” As a heuristic device, Leiter contemplates a “scenario in which the state, instead of disestablishing religion in the public schools, endorses a particular religion (say, Catholicism) and thus declines to let funding for public education be utilized for supporting Hinduism or atheism.”186 This means, among other things, that “public school facilities” (such as classrooms in the afternoon) “would be available to the Catholic Student Society, but not to the Hindus or the atheists or perhaps even to the Republicans!”187 So long as dissenters are permitted to express contrary views using their own resources, including wearing religious symbols or garb to school, and to attend alternative sectarian schools,188 he says this establishmentarian scenario is consistent with “principled toleration.”189

That is probably true. The government could use its prestige, power, and resources to support one vision of religious truth while still leaving dissenters free to dissent. The establishment of religion may be consistent with mere toleration, but it is not consistent with the “full and free exercise of religion”190 that our founders adopted at the federal level in lieu of toleration. About half a dozen states pursued some form of tolerant establishment in the early years of the Republic, when the Religion Clauses did not apply to state governments, but all of them dismantled their establishments by 1833. No one, to my knowledge, mourns their passing.

Toleration might be the most we can hope for in nations of the Middle East, where the population is overwhelmingly of one religious faith and there is a long tradition of union between mosque and state, but for pluralistic liberal democracies, mere “toleration” would be a step backward. From the point of view of religious freedom or of liberal constitutionalism more generally, it is hard to see why anyone would prefer Leiter’s hypothetical Catholic establishment to a regime of religious neutrality. As Madison and others pointed out long ago, the establishment of religion is bad for religion, including the established faith, bad for dissenters, bad for government, and bad for freedom.

Leiter recognizes that it is “possible that a religious or irreligious establishment reduces citizens with differing views to a second-class status.”191 But for some reason this “is a separate question,” which requires a “culturally nuanced inquiry.”192 He says no more against it.

Of course, Leiter has no interest in establishing the Catholic religion. What he defends is the establishment of secularism, where we would use the public schools to inculcate ideologies of a nonreligious nature and prevent voluntary student groups from using the facilities on an equal basis for prayer or Bible study.193 His defense of establishment is a disguised attack on the modern constitutional doctrine that the state must be neutral toward religion and may not deny equal access to otherwise open public facilities to groups on account of their religious point of view.194

Now, the idea of a secular state may sound harmless. We often use the term, loosely, to describe a nonsectarian or nonconfessional state—a state that is not committed to a particular religion or religious worldview.195 But Leiter is using the idea in a more insidious way, to denote a state that is committed to secularism as a substantive position—that is, to what he calls “irreligion, in the form of atheism or otherwise.”196 The establishment of secularism would stand in the same relation to religious beliefs as his hypothetical Catholic establishment stands to Hinduism, atheism, and Republicanism. The whole point of this sixteen-page detour197 into antidisestablishmentarian theory is to legitimate the use of governmental institutions, especially schools, to promote secularism or irreligion and to discriminate against religious speech.

A state that is neutral toward religion is different. Such a state may promote ideas consistent with democratic republicanism, but will not promote religion over irreligion or the other way around. It “may place its imprimatur on values and worldviews that are inconsistent with the claims of conscience of some of its citizens,”198 just as—in the words of the Supreme Court—it may pass laws that “happen[] to coincide or harmonize with the tenets of some or all religions.”199 But it cannot teach religion (though it can teach about religion in a non-catechetical way), and it cannot teach “irreligion” either. And when such a state opens its facilities to private persons for speech of their own choosing, it must neither favor nor disfavor groups on the basis of their religious or other beliefs. We should remember Justice Arthur Goldberg’s admonition in the School Prayer Cases that “a brooding and pervasive devotion to the secular and a passive, or even active, hostility to the religious . . . [is] not only not compelled by the Constitution, but . . . prohibited by it.”200

Leiter’s defense of the establishment of religion brings us back, full circle, to where we began discussion of this book. It seemed odd and anachronistic that Leiter would write of religious “toleration” instead of religious freedom. Toleration was a term associated with the religious establishment. As President Washington wrote to the Hebrew Congregation of Newport, in disestablishmentarian America “[i]t is now no more that toleration is spoken of.”201 It turns out that Leiter wants to return to the earlier regime, but with secularism rather than Anglicanism in charge. I hazard the guess that he will not persuade many readers not already predisposed to that point of view.