Intermediate Questions of
Religious Exemptions--A Research
Agenda with Test Suites

Eugene Volokh*

21 Cardozo Law Review 595 (1999)

    I. Discrimination in Favor of Religion
        A. Preferences for Religion Where Something Major Is at Stake
            1. The Problem
            2. Implementation
            3. The Test Suite
        B. Preferences for Religion That Unduly Burden Others
        C. Discrimination in Favor of Religiously Motivated Speech
            1. The Problem
            2. Distinguishing Speech from Non-Speech
            3. The Test Suite
    II. Applying Strict Scrutiny
        A. Interests in Preventing Harm to Others
            1. Looking to History or Tradition
            2. Distinguishing Direct Harms from Indirect Ones, and Invasions of Rights from Mere Imposition of Negative Externalities
            3. The Test Suite
        B. Paternalistic Interests
        C. Looking to Secular Exceptions from a Law to Decide Whether a Religious Exemption Should Be Granted
    III. Government as Employer, Proprietor, or Educator
        A. Government as Employer
            1. What Is the Test?
            2. If Not Strict Scrutiny, Then What?
            3. But What Do These Formulae Really Mean?
                a. Requests for exemptions from parts of one's job requirements
                b. Requests for exemptions from general courtesy and professionalism rules
                c. Requests for exemptions for conduct that may diminish public confidence in the government employer
        B. Government as Proprietor
            1. What Constitutes a "Substantial Burden"?
                a. Using property in ways that create spiritual harm
                b. Setting up property in ways that make it unusable by believers
                c. Restricting behavior on government property
                d. Restricting behavior on government property in ways that make it unusable by believers
            2. What Is the Proper Level of Scrutiny?
       C. Government as K-12 Educator
            1. What Is the Proper Level of Scrutiny?
            2 Religious Objections to Required Classes
            3. Religious Objections to Conduct Rules


      Does the Constitution require courts to carve out religious exemptions from generally applicable rules? If it doesn't, should legislatures enact statutes that command such exemptions? These might be called primary questions of religious freedom law, and they have been much debated. To some extent, they have recently been resolved by legislatures: religious exemption statutes (commonly known as RFRAs, because they are modeled on the federal Religious Freedom Restoration Act of 1993(1)) now bind the federal government and the governments of several states.(2)

      But between these broad primary questions and the equally heavily debated specific questions related to particular proposed exemptions--exemptions from antidiscrimination laws, weapons laws, drug laws, compelled testimony laws, and the like--lies a range of intermediate questions.(3) How does the RFRAs' textual requirement of across-the-board strict scrutiny of governmental burdens on religion apply to government employment, government property, or government-run schools? What exactly does "least restrictive means of furthering [a] compelling governmental interest"(4) mean? How should courts deal with laws aimed at protecting nontraditional private rights, such as the right to be free from marital status discrimination in housing? Does it violate the Free Speech Clause to protect religiously motivated speech more than nonreligiously motivated speech? 

      These intermediate questions were comparatively underdiscussed even before Employment Division v. Smith,(5) but in any case the answers to them may be different under the RFRAs than under the constitutional exemption regime. RFRAs have more specific, binding text than does the Free Exercise Clause--thus, for instance, they facially require strict scrutiny of all substantial burdens on religious practices, something the pre-Smith Free Exercise Clause jurisprudence did not do. 

      Moreover, as I've argued elsewhere, state RFRAs embody a "common-law model" of religious exemptions, under which, as with the development of the common law itself, courts take the first step in defining the rules but the rules remain subject to legislative revision. This may make it proper for judges to use their moral and practical judgment in this process, something that they've traditionally done when making the common law but that they are often properly reluctant to do when making constitutional rules.(6)

      But just as state RFRAs create opportunities for judicial creativity, they also create opportunities for error and unequal treatment. As long as there are no good general answers to the intermediate questions of religious exemption law, it'll be easy for judges to reach one result for religions or religious practices with which they are familiar or sympathetic, and another for religions or practices that are novel or unappealing. "[I]f arbitrary and discriminatory enforcement is to be prevented, laws must provide explicit standards for those who apply them,"(7) whether they are executive officials or judges. And one task of legal scholarship is to help courts translate vague, general statutory or constitutional language into rules that set forth somewhat more "explicit standards." 

      In this Article, I want to briefly identify--though not resolve--some of these relatively uncharted areas on the religious freedom map, and perhaps provide something of a plan for future research. For each of these areas, I will also suggest some concrete cases that I think pose particularly tough challenges--what computer programmers call a "test suite"--which can be used to check whether any particular proposed doctrine makes sense.(8)

      I'm especially interested in the intermediate questions because I think their answers will likely prove especially useful. The broad primary questions cover such a wide area that unless the answers to these questions are all-or-nothing (e.g., Smith's conclusion that exemptions generally need not be granted), they end up being too vague to say much about how particular cases should be decided (e.g., Sherbertv. Verner's(9) strict scrutiny test). The specific questions about exemptions for particular practices cover such a small area that they naturally invite special pleading on behalf of the author's predilections in this particular case, with little attention to the precedential consequences of the answer in cases that are structurally similar. But answering the intermediate questions can often yield rules that are both relatively determinate and thought through with an eye to likely future cases. 

      Test suites are particularly good tools for evaluating the proposed answers to intermediate questions.(10) First, test suites help us "think things, not words."(11) Lawyers often talk in abstract generalities that don't resolve concrete cases--"compelling government interest," "balancing government interests against employees' religious freedom," and the like; proposals phrased in such general terms can't be evaluated or even understood until we try to apply them to concrete problems. In the process of application, we might realize that the formula in practice means something different, or perhaps is so indeterminate that it really means nothing at all. 

      This should lead us to be cautious, because it shows that even if the proposal can be applied in ways that we think are correct, in someone else's hands it can just as easily be used to reach the wrong results as the right ones. More optimistically, the process of applying the proposal to concrete cases can help us refine and clarify the verbal formulae to make them more determinate and more useful. 

      Second, test suites help us avoid endorsing unsound proposals. When we feel strongly about a particular case, we may be tempted to accept whatever rule seems to do justice in that case, and then regret it in later cases where the logical consequences of the newly established precedent become clear. Thinking ahead about such cases can lead us to realize that the original proposal is wrong, or at least flawed and should be amended. And it's particularly helpful if the test suite contains a variety of fact patterns that appeal to different political preferences--for instance, includes exemption claims that match the favorite practices of the far left, left, center, right, far right, and libertarians. 

      Of course, the test suites I suggest are at most first drafts; commentators who propose answers to these intermediate questions should modify the test suites as they see fit. But I hope my suggestions can prove to be useful steps towards making any proposed religious exemption principles be as clear and sound as possible. 

I. Discrimination in Favor of Religion 

      Religious exemption regimes treat people whose objections stem from their religions (whom I'll call "religious objectors") better than they treat people whose equally deeply held objections stem from their secular consciences (whom I'll call "conscientious objectors"). If you're an atheist, agnostic, or even a religious person whose beliefs in this particular case flow from your nonreligious conscientious views, you categorically lose. If I hold equally strong beliefs because of my religion, I may win. 

      One of the primary questions of religious exemption law is whether this preference for religion is constitutionally and morally sound, and this question becomes even tougher post-Smith. When the Free Exercise Clause was read as compelling religious exemptions, one could have argued that the Constitution itself created this preference for religion. But when religious exemptions are required by statutory, not constitutional, command, it becomes more plausible to argue that: (1) the (federal or state) constitutional norm of equal treatment without regard to religiosity trumps the statutory preference for religious objectors over conscientious objectors; or (2) the legislature that's considering a RFRA is morally required to treat the two categories of objectors equally, whether or not the Constitution also embodies such a command. 

      Some may argue that the answer to this primary question is categorically "there's no constitutional or moral problem with preference for religious objectors over conscientious objectors." Others argue that the answer is equally categorically "such preferences are wrong." These two positions have been discussed elsewhere,(12) so I won't go into them here. 

      But what about the possibility that some religion-only exemptions are constitutional and others aren't? This middle ground has often been suggested, but unfortunately it has not yet been adequately defined and defended. 

A.Preferences for Religion Where Something Major Is at Stake

1. The Problem 

      Say that a proposed religious exemption involves relief from a truly major burden, such as an exemption from the draft. Even if the government may give religious objectors preferential access to peyote or preferential treatment for their Sabbath observance, it doesn't necessarily follow that the government may give them preferential protection from risk of death, or from being put in a position where they must kill. In fact, even some supporters of religion-only exemptions take this view.(13)

      There are two possible reasons for concluding that some such preferences may be unconstitutional or at least morally unsound. One is simply concern about discrimination: Putting some people at greater risk than others of dying or of having to kill simply because their value systems aren't religiously based may be particularly troubling. 

      The other problem is the risk that a religion-only exemption would pressure people to become religious or at least to profess religion, something the United States Supreme Court has said is one of the evils against which the Establishment Clause guards.(14) It might not be very likely that a nonreligious person would join a religion (or claim to do so) just to use peyote; but it's quite likely that at least some people who want to avoid the draft will claim a religious belief and even start going to church if that helps them take advantage of a religious objection regime.(15)

      This approach--some preference for religion is permissible, but too much is not--is appealing, but it raises the obvious and still unanswered question: How large must the requested immunity be to become too large? 

      It's not enough to just ask whether the exemption involves some discrimination or provides some inducement to become religious or profess religion; virtually all religion-only exemptions share these traits. One might ask whether "the magnitude of the burden being removed from religion outweighs any inducement to religion,"(16) but that just raises the next question of how to weigh these two rather different items. 

      The "how large is too large" question might be purely academic if exemption debates almost never involved really substantial immunities--if, for instance, the draft exemption question were sui generis.(17) But in fact many such debates do involve serious matters. 

      There is, for instance, an eminently credible case under the RFRAs for a right to assisted suicide, where the patient or the doctor feels a sincere religious obligation to participate in the procedure.(18) At the same time, I suspect that many of those who want to commit assisted suicide for nonreligious reasons, or who want to assist another's suicide for nonreligious reasons--both of which are matters of great moral gravity--do so out of a deeply held conscientious belief: a morally urgent desire to live and die with dignity, to spare one's relatives continuing anguish, or to help another ease his suffering. 

      Is it right to allow religious people the solace, dignity, and opportunity for charity that assisted suicide provides, but to deny it to secular people? And isn't there a "strong likelihood"(19) that some patients or doctors who want the exemption will be pressured into feigning a religious objection, and perhaps even joining a religious group that believes in assisted suicide so as to make their objections more credible? 

      Likewise, some judges have shown a sympathy to religion-based claims of immunity from forced testimony against one's family members.(20) Having to testify against one's loved ones is a tremendous burden, to which many people have the deepest of conscientious objections. Exempting only those whose objections are religious is a serious discrimination, and granting such objections may well pressure people to feign religious belief and engage in religious conduct: For instance, if being a synagogue-going Orthodox Jew makes it easier to persuade a court that one ought not have to testify against one's family, then many secular Jews may well claim Orthodoxy and even take up some orthodox practices in order to take advantage of such an exemption. 

      In other cases, the claim that there's "too much" discrimination is weaker, but still plausible, especially if we focus on the pressure to claim religious beliefs. Some people do want to get Saturdays off for secular reasons that may rise to the level of conscientious obligation--for instance, to spend more time rearing their children. Others want to discriminate against unmarried or homosexual couples in housing because they oppose premarital sex or homosexuality on deeply felt moral grounds, even setting aside religion.(21) Still others want to consume peyote for important secular reasons, for instance because they believe it important to expand their consciousness by altering it.(22)

      One might believe that courts should reject the equality arguments in these cases, but it's still important to develop some rules or guidelines that will help courts reach these results. Among other things, the existence of such guidelines might diminish the likelihood that courts will decide these questions based on subtle or unconscious prejudice in favor of or against certain religious groups or practices. 

2. Implementation 

      What are the possible consequences of concluding that religion-only exemptions are sometimes unconstitutional or at least wrong? 

      (1) One possibility is that under the Establishment Clause or one of its state equivalents, granting a religious exemption is in certain cases unconstitutional. Thus, a court might say: "True, the state RFRA may authorize religious exemptions from assisted suicide laws, but such an authorization is unconstitutional because it gives too big a preference to religious objectors. Your religious request for an exemption therefore loses." Alternatively, a court might reach the same result by concluding that there's a compelling government interest in not unduly discriminating against the nonreligious, or unduly pressuring people into professing religion: "It would be unconstitutional for the state RFRA to be read as authorizing religious but not conscientious exemptions from assisted suicide laws, so the government has a compelling interest in preventing such an unconstitutional preference, and therefore the religious request for an exemption from assisted suicide laws loses." 

      (2) Courts might sometimes conclude that a religion-only exemption would be constitutional, but still too discriminatory to be granted: "Even though it would be constitutional to have a religious, but not conscientious, exemption in this case, discrimination on a matter as important as assisted suicide is still unfair, and the government has a compelling interest in preventing such unfairness. The religious request for an exemption from assisted suicide laws thus loses." 

      (3) Alternatively, a court might conclude that equal treatment requires not rejection of religious exemption claims but acceptance of conscientious exemption claims. This was the reasoning of Justice Harlan's concurrence in Welsh v. United States,(23) in which he argued that the statutory draft exemption only applied to religious objectors and was therefore unconstitutional, but the unconstitutionality should be remedied by equalizing up rather than equalizing down. 

      Likewise, such an approach might fit well with the RFRAs' "least restrictive means" requirement: Though the government clearly has a compelling interest in avoiding constitutional violations, and might even have a compelling interest in preventing constitutional but unfair preferences for religion, exempting both conscientious objectors and religious objectors would serve these interests but at the same time provide the religious exemption that the RFRAs' texts presumptively demand. The court would thus say: "While there's a compelling interest in preventing religious discrimination in assisted suicide, denying the exemption to the religious objector is not the least restrictive means of serving the interest. Rather, the least restrictive means is to grant the exemption both to religious objectors and to conscientious objectors." 

      On the other hand, this less restrictive alternative--extending the exemption to conscientious objectors--might sometimes be unsatisfactory because it would unduly jeopardize the government's other compelling interests: For instance, one could argue that even if the judicial system could function effectively while allowing the rare religious objection to compelled testimony by family members, it would be substantially hurt if it had to allow all conscientious objections. In such a case, the combination of the two compelling interests--preventing "too much" discrimination in favor of religion and protecting the judicial system against the loss of a great deal of evidence under a general conscientious objection regime--might be enough to defeat the religious objector's claim. 

      (4) Finally, if courts conclude that the federal and state Establishment Clauses tolerate all preferential exemptions for religion, even as to the "big items," this would actually make legislators' jobs tougher, because it would require them to ask whether such preference for religion is nonetheless sometimes morally unsound. 

      If legislators think that the preference is indeed unsound, though not constitutionally forbidden, then they wouldn't just be able to enact a general religion-only RFRA and expect the courts to sort it all out on an as-applied basis; rather, they'd have to themselves try to prevent the unfair preference for religion. One possible way to do this would be to enact RFRAs that cover all conscientious objectors, not just religious ones: Even if some preferences for religious objectors are morally proper, the difficulty of telling which are and which aren't, coupled with the risk that the courts will uphold both the sound and the unsound preferences for religion, may counsel in favor of an across-the-board conscientious exemption regime. Another way would be to craft RFRAs that grant religious exemptions in some cases--those where the discrimination is small enough to be unproblematic--and conscientious exemptions in others.(24)

      I'm not sure which of these alternatives is the right one, but I'm pretty sure that deciding which alternative is right is an important and still largely unresolved problem. And unless the legislature decides to just enact a RFRA that covers both religious and conscientious exemptions, then even after this problem is solved there will remain the question of just where the line between permissible and impermissible preference for religious objections should be drawn. 

3. The Test Suite 

      As the above discussion suggests, I think that any "preferences for religion are permissible unless something really major is at stake" proposal should confront something like the following test cases: 

      (a) Conscientious objectors claim that a draft exemption available only to religious objectors--imagine that Congress statutorily overrules Welsh v. United States(25)--is unconstitutional. This is not quite a RFRA inquiry, since draft exemptions probably wouldn't be available under a RFRA alone,(26) but it does illuminate the Establishment Clause question. 

      (b) Those who believe as a matter of conscience that they ought to commit assisted suicide, or ought to assist another's suicide, claim that RFRAs should treat them the same way as religious objectors. 

      (c) Same as (b), as to those who believe as a matter of conscience that they ought not testify against family members. 

      (d) Same as (b), as to those who want unemployment compensation after having been dismissed for refusing to work on Saturdays because of a conscientious obligation to spend weekends with their children. 

      (e) Same as (b), as to those who want to use peyote to expand their consciousness. 

B.Preferences for Religion That Unduly Burden Others

      Some commentators have suggested that the Establishment Clause may impose another limitation on religious exemptions--"In accommodating religious activities, the legislature should not impose disproportionate costs on other citizens or activities."(27) Language in two Supreme Court cases, Estate of Thornton v. Caldor(28) and Texas Monthly, Inc. v. Bullock,(29)supports this view.(30)

      I'm not quite sure that such a limitation is sound. I agree that the burden on others is relevant to whether the government has a compelling interest in denying exemptions, and I agree that discrimination against conscientious objectors raises Establishment Clause problems. But I'm not sure that the burden imposed by the exemption on others, who aren't conscientious objectors to the law and thus have no claim to equal treatment with the religious objectors, should be relevant to the Establishment Clause analysis. Given that laws are generally allowed to impose all sorts of burdens on people, it's not clear to me that religious exemption laws that impose comparable burdens are made unconstitutional by the existence of these burdens. Nonetheless, some people--including some Justices--take this view, and it thus can't be lightly dismissed. 

      If this view is to be implemented, there's need for a doctrine that distinguishes which burdens on others are proper and which are "disproportionate." Many accommodations (of religion or conscience) impose some burden on others.(31) Unemployment compensation, for instance, generally comes indirectly out of the pockets of employers; Adele Sherbert's gain was her employer's loss.(32) Refusing to testify when called by a criminal defendant or a party in a civil case imposes a burden on someone who would otherwise have had a legally secured right to the objector's testimony. Exempting religious schools from certain zoning regulations may give such schools an edge in competing against secular private schools.(33) Religiously motivated discrimination in housing against unmarried couples is seen by many as imposing a burden on the supposed victims of such discrimination, though others deny that this should be a legally cognizable burden. Likewise, as even some defenders of religion-specific exemptions admit, both "exemptions from military services . . . and many forms of tax exemption . . . impose burdens on non-beneficiaries, forcing them to shoulder added responsibilities for defense or revenue."(34)

      The Establishment Clause controversies in these cases can be avoided if a court concludes that religious exemptions aren't required because the government has a compelling interest in preventing such a burden on others.(35) But this need not always be the case: Much depends on what courts conclude qualifies as a "compelling government interest"(36) and on what exactly are the boundaries of the Establishment Clause "no exemption if there's a disproportionate burden on others" doctrine. If there is such a doctrine, and if it does indeed limit or modify the implementation of a RFRA (all the implementation alternatives mentioned in Part I.B would also apply to this doctrine), it would be useful to have some proposals outlining just what the doctrine should be. And the test suite for such a proposal would include the cases mentioned in the text: unemployment compensation; draft exemption; tax exemption; the duty to testify; and housing discrimination. 

C.Discrimination in Favor of Religiously Motivated Speech

1. The Problem 

      Even if the Establishment Clause and its state equivalents have nothing to say about discrimination in favor of religion, the Free Speech Clause and its state equivalents may bar discrimination in favor of religiously motivated speech. 

      Say that Alice feels a religious motivation or even a religious compulsion(37) to picket in front of an abortion provider's home, despite a city ordinance banning residential picketing, and Bob feels a secular motivation to picket in front of the home of the leader of Operation Rescue. Under Frisby v. Schultz,(38) Bob has no right to picket; content-neutral speech restrictions are tested only under intermediate scrutiny, and residential picketing bans pass muster. But under a RFRA, Alice would get an exemption from the residential picketing ban unless the ban passes strict scrutiny, not just intermediate scrutiny. It's far from clear that such bans do pass strict scrutiny, so Alice may well win, while Bob will surely lose. 

      The same applies in many other situations where speech restrictions are tested under standards lower than strict scrutiny: religious claims to solicit funds at county fairs,(39) religiously motivated billboards in no-billboard zones,(40) religiously motivated flyers posted on telephone poles in a city that generally bars such postings,(41) religious claims to publish books that infringe copyright,(42) religious broadcasters' claims to be free of a reborn Fairness Doctrine,(43) and so on. RFRA seems to suggest that religiously motivated speakers would have greater speech rights than secularly motivated speakers,(44) a problem that didn't exist under the pre-Smith Free Exercise Clause jurisprudence, where the Court scrutinized claims of religious exemption from speech restrictions under the same standard as that used in Free Speech Clause cases.(45)

      Does such discrimination in favor of religiously motivated speech violate the Free Speech Clause? The discrimination is not on its face content-based, since it doesn't turn on whether the person is saying religious things, and it's not justified with reference to the communicative impact of the speech:(46) The law is aimed at preventing substantial burdens on religiously motivated behavior, an interest that has nothing to do with the communicative impact of the speech. 

      In fact, the Court has upheld certain discriminations among speakers that were justified by the speakers' identities rather than the content of their speech. Regan v. Taxation with Representation(47) upheld a law that let tax-exempt veterans' organizations, but not other tax-exempt organizations, lobby with tax-exempt funds, though the Court rested this conclusion in large measure on the fact that the law involved a speech subsidy, not a speech restriction. Madsen v. Women's Health Center, Inc.(48) upheld a speaker-specific injunction, treating it as content-neutral because the government's purpose was the prevention of trespass and obstruction by the speakers. Turner Broadcasting System v. FCC(49) upheld a requirement that cable operators devote some of their space to over-the-air broadcasters, rather than to channels only carried on cable, reasoning that the government's purpose was the content-neutral one of preserving free over-the-air broadcasting. 

      On the other hand, discrimination among speakers, even when it's not justified by the content of the speakers' likely speech, raises obvious concerns. Many political debates involve disproportionately religiously motivated speakers on one side and disproportionately secularly motivated speakers on the other. Giving a special preference to the religiously motivated speakers may substantially skew the content of the debate, perhaps as much as a content-based restriction would. And it might also be just wrong--given the importance and constitutional significance of speech--for the government to allocate speech opportunities based on religiosity, even if preference for religion is generally constitutionally permissible as to behavior that is not itself constitutionally significant. 

      The Court has at times recognized this, insisting that certain speaker-based restrictions must be subjected to strict scrutiny even when they aren't justified based on content. Thus, Minneapolis Star & Tribune Co. v. Minnesota Commissioner of Revenue(50) struck down a tax imposed only on large newspapers, saying that the power to tax only a small set of speakers presents too much potential for abuse. Recently, Greater New Orleans Broadcasting Ass'n v. United States(51) struck down a ban on casino advertising that exempted certain casinos, partly on the grounds that "[e]ven under the degree of scrutiny that we have applied in commercial speech cases, decisions that select among speakers conveying virtually identical messages are in serious tension with the principles undergirding the First Amendment."(52) Similarly, First National Bank of Boston v. Bellotti(53) said that "[i]n the realm of protected speech, the legislature is constitutionally disqualified from dictating the subjects about which persons may speak and the speakers who may address a public issue."(54)

      The constitutionality under the Free Speech Clause of preferential treatment for religiously motivated speech thus seems unresolved. I am personally persuaded by Alan Brownstein's argument that discrimination based on a speaker's motivation is so close to discrimination based on content, and so comparatively far from the discrimination upheld in TWR, Turner, and Madsen, that it must be considered unconstitutional.(55) Nonetheless, the argument is hardly open and shut, and there's certainly need for further work on this question. 

      Finally, even if a preference for religiously motivated speech doesn't violate the Free Speech Clause, two questions remain: (1) Does the government still have a compelling interest in promoting equality in speech opportunities, an interest that would defeat a RFRA claim to exemption from a speech restriction? (2) If the answer to the first question is "no," then is the preference still morally unsound, which would mean that legislatures should craft RFRAs to prevent such a preference? As I suggested with regard to the religious preference generally, the absence of any judicial remedy for unfair preference for religion only makes it more important for the legislature to prevent such unfairness in the first place. 

2. Distinguishing Speech from Nonspeech 

      Suppose principles of speech equality do require religiously motivated speech to be treated the same as secularly motivated speech. Religiously motivated speech, then, must be excluded from the RFRAs' across-the-board protection for religiously motivated behavior, whether under the Free Speech Clause, under a compelling interest analysis, or through a specially crafted statutory provision.(56)

      How broad will this exclusion be? Much religiously motivated conduct is speech, which has provided considerable protection for religion in the past.(57) Requests for exemptions of religious meetings from zoning requirements are generally aimed at facilitating religious speech--prayer, sermons, and the like.(58) This is also true for claims of exemption from child labor laws or minimum wage laws when the labor involves distribution of literature.(59)

      Preferences for other religiously motivated conduct might also implicate Free Speech Clause issues, though less clearly so. For instance, religious schools' claims of exemption from certain antidiscrimination laws affect the schools' ability to choose whom to speak to and whom to let into their debates.(60) An exemption that lets bankrupt debtors donate money for religious purposes often subsidizes religious speech.(61) Claims of exemption from compulsory education laws likewise touch on parents' right to control what is said to their children: After all, part of the Amish claim in Yoder was that "expos[ure] of their children to worldly values" would interfere with their children's "integration . . . into the Amish religious community."(62) Claims of exemption from a duty to testify(63) implicate one's right not to speak.(64)

      I can imagine at least three possible lines that can be drawn here. First, one might say that the Free Speech Clause only bars exemptions for religiously motivated speech from laws that target speech activity. Special religious exemptions from residential picketing laws, antisolicitation rules, compelled testimony laws, and possibly compulsory education laws would therefore be unconstitutional preferences for religiously motivated speech(65) over other speech. But exemptions of religious universities from antidiscrimination laws, exemptions of debtors from fraudulent transfer laws, exemptions of religious assemblies from zoning laws, and exemptions of religious literature sellers from child labor laws would be permissible, because those laws cover a broad range of conduct without singling out speech. 

      The premise of religious exemption law, though, is precisely that laws of general applicability should not enjoy any "talismanic"(66) immunity from religious freedom principles simply because they don't target religion: if the laws in effect interfere with the practice of religion, they are burdens on religion. Given this, it seems odd to make the linchpin of the no-discrimination-in-speech-activities analysis be whether the law at issue targets speech. The RFRAs' strict scrutiny, whether applied to laws that target speech or laws that don't, would sometimes treat religiously motivated speech better than other kinds of speech. If there's something wrong with this discrimination, it seems equally wrong regardless of the breadth of the law at issue. 

      A second approach would be to ask whether the activity involved is a "speech activity"; if it is, then the propriety of applying any restriction to this activity must be judged under the applicable Free Speech Clause principles, not under RFRA strict scrutiny. The question then might be (to borrow a test from the free speech area, though such borrowing is by no means obviously correct) whether the conduct intends to convey and is likely to convey a particular message(67)--either to outsiders or to fellow worshippers--or would otherwise conventionally be a form of speech.(68)

      Third, one could try to determine which religious exemptions are in fact likely to materially privilege religiously motivated speech in public contests with nonreligiously motivated speech. Thus, one might say that allowing religiously motivated residential picketing but not secularly motivated picketing may favor one side over the other in certain debates, but favoring religiously motivated meetings in residential areas over secularly motivated meetings is less likely to do so, because the religiously motivated meetings generally tend to be directed much less to public debate and much more to speech within a religious group. On the other hand, I suspect that this sort of test probably requires too much case-by-case speculation about the nature of certain political debates to work effectively; it's also not completely clear to me that its core premise is correct--a preference for religiously motivated speech that helps foster the growth of religious ideological communities and not secular ones may be as troublesome as a preference that more directly affects public debate. 

3. The Test Suite 

      Any proposal for avoiding Free Speech Clause problems by excluding speech from the RFRAs' strict scrutiny must, it seems to me, confront this issue. The following test suite, drawn from the examples above, might help: 

      (a) A county fair regulation that bans solicitation for all causes.(69)

      (b) A city ordinance that bans all residential picketing.(70)

      (c) A compulsory education law that requires all children to be educated until age sixteen.(71)

      (d) A Bankruptcy Code provision that in effect bars debtors from giving away money to any person or cause.(72)

      (e) A child labor law that applies to a wide range of labor, including children selling pamphlets.(73)

      (f) A zoning law that bans fairly large, frequent assemblies in private homes.(74)

II. Applying Strict Scrutiny 

      Under state RFRAs, courts must determine whether a law serves a "compelling governmental interest," and whether applying the law and denying the exemption request is the "least restrictive means" of furthering that interest. How exactly are they to do this? 

      One possible answer is "the same way they did it under the Sherbert regime that RFRAs were meant to restore." Unfortunately, this isn't much of an answer, since the Sherbert-era strict scrutiny test was quite vague--much more so than strict scrutiny in free speech or race classification cases--and even many of those who supported it in principle roundly criticized its implementation.(75)

      But beyond this, the analysis under RFRAs should differ from the constitutional analysis. Under RFRAs, courts aren't overruling legislative judgments; rather, they are implementing the legislature's own judgment that religious exemptions ought often be granted. Moreover, under RFRAs, the courts' decisions aren't final--if a legislature believes an exemption unduly hurts others, directly or indirectly, it can enact a law repealing or modifying that particular exemption. This is why I argue that state RFRAs create a "common-law exemption regime," under which courts act much as they did in framing the common law: by taking the lead role in creating legal rules, but recognizing that the legislature may always supersede these rules. 

      It would therefore be proper for courts to be considerably less hesitant in creating exemptions under state RFRAs than they were under the pre-Smith Free Exercise Clause case law. Various decisions that under a constitutional regime would rightly be condemned as improper imposition of the courts' own moral or factual judgments become perfectly proper when the courts' lawmaking is both pursuant to the legislature's command and revisable by the legislature. After all, the creation of the common law was all about courts exercising their moral and factual judgment in deciding what the law ought to be. 

      But courts will properly worry, when granting exemptions, about what precedent they'll be setting for future exemptions. A proposed test that will distinguish exemption requests that should be granted from those that should be rejected might ease such slippery slope concerns by identifying defensible stopping points along the slope. And such proposals will probably work best if they are neither too narrow (and thus too obviously tailored to one particular problem) nor too broad (and thus probably so vague and general that they aren't much more helpful than "least restrictive means of serving a compelling government interest"). 

A.Interests in Preventing Harm to Others

      It would be nice if someone could define what makes an interest "compelling." But this is a tall order. The Court has never precisely defined this term in any of the contexts in which strict scrutiny has been applied, and the world might be too complex for any nontautological definition to be created. Still, one partial proposal that's often made is that preventing "impos[ition] on the basic rights of another person"(76) or "the prevention of negative externalities"(77) is per se a compelling governmental interest: my religious beliefs never give me the right to trespass on your property, touch you without your consent, defame you by making secularly verifiable but false factual claims about you, infringe your copyright, and so on. "I have no right to walk across my neighbor's land without permission, no matter how vital it may be to my religion or how minor an imposition it may be on my neighbor. The free exercise clause does not allow believers to shift the cost of their religious practices on others."(78) And this makes sense; religious liberty is appealing when it's a claim of "immun[ity] . . . from civil jurisdiction [when religious exercise] does not trespass on private rights,"(79) or of "free[dom] to practice our religions so long as we do not injure others,"(80) but is not so appealing when it's a claim of a right to act in ways that do infringe others' rights. 

      The trouble, as many have realized, is that this requires a determination of what constitutes "impos[ition]" on "basic rights" or the creation of "negative externalities," particularly: (1) which rights are basic and which negative externalities count; and (2) which actions in fact impose on these rights. Many religious exemption conflicts arise precisely because of disagreements about the boundaries of people's rights to be free of certain impositions by fellow citizens. If one thinks that people have a right not to be discriminated against in housing based on marital status--much like they have rights not to be physically attacked or have their property trespassed on--then landlords' claims to exemptions from marital status discrimination laws should fail: "[I]t is unwarranted even to grant a single exemption for a religious claim" when even an "individual instance of the particular religious activity imposes on the basic rights of another person."(81)

      Likewise, if one thinks each litigant has a basic right to subpoena witnesses (subject to certain recognized limits) to defend his life, liberty, or property from unjust deprivation, then a witness's refusal to testify would be a violation of the litigant's rights. On the other hand, if one thinks that people don't really have such rights, or that the rights aren't "basic," then it becomes much harder (though not impossible) to claim that the government has a compelling interest in denying the exemption. 

      Similarly, the debates about Sikhs' claims of exemption from weapons laws rest on a factual disagreement over whether such conduct will in fact indirectly impact others' "basic rights." Everyone agrees that we have a right to be free from unjustified physical attack; the premise of bans on carrying guns and knives is that such conduct may (not must, but may) lead to violations of this right, which is why some argue that private possession and carrying of arms in fact creates a negative externality.(82) Though the prohibited conduct doesn't itself violate the basic right, many people believe that the risk of harm that flows from the conduct is great enough to justify punishing weapons carrying as an "impos[ition]" on the "basic right[s]" of others. Others, though, argue that weapons carrying by law-abiding adults in fact does not increase the risk of crime, and might even decrease it.(83)

      I have argued elsewhere that the problem with a constitutional exemption model is precisely that courts generally ought not second-guess legislative moral or factual judgments about when conduct starts to impose on others' basic rights.(84) On the other hand, under RFRAs, courts have no choice but to make these judgments, and in the context of those statutes such second-guessing is perfectly proper, because it's done on the legislature's own orders and subject to legislative revision. 

      Still, the fact that courts may and must make these judgments doesn't tell courts how they are to make these judgments. Are there any general principles courts can follow in deciding these issues? 

1. Looking to History or Tradition 

      Some suggest that courts should make these decisions based on history: Tom Berg, for instance, suggests that this rights-of-others principle should be at least presumptively limited to "historic, libertarian-oriented interest[s] of [o]ther individual[s],"(85) thus excluding, for instance, rights under antidiscrimination law. There is much that's appealing in this, and much that's off-putting. On one hand, given the need to draw the definition of the basic rights of others from some source other than the current laws themselves--since RFRAs are supposed to provide exemptions from those very laws--longstanding American traditions seem a good starting point. That a right has been recognized by the common law for centuries helps make the case that the government interest in protecting the right is relatively compelling; that a right is only one or two generations old tentatively suggests the contrary. 

      On the other hand, the common law is far from pure, and may have often erred in its decisions as to what constitute a person's "basic rights." Even libertarians will probably take this view when it comes to common-law overregulations: Few, for instance, think that the common law's recognition of the torts of alienation of affections or criminal conversation was its finest hour.(86) And many nonlibertarians take this view when it comes to underregulation, believing for instance that the pure employment-at-will doctrine, under which an employer's rights included the right to fire someone for any reason or no reason at all, and under which the employee had no recognized countervailing rights, was a mistake. 

      I largely support employment-at-will, but many others do not, and I'm not sure that state RFRAs should be interpreted as endorsing my views rather than theirs. If modern society has concluded that certain behavior does violate the rights of others, why should courts defer more to the judgment of the past than of the present, even when religious claims are involved? There may be a satisfactory answer to this question, but it seems to me to require more examination. 

2. Distinguishing Direct Harms from Indirect Ones, and Invasions of Rights from Mere Imposition of Negative Externalities 

      Similarly, quite a few people suggest that courts should distinguish "direct" harms from "indirect" ones.(87) Others propose a similar distinction between invasions of actual rights, recognized by the law as rights, and imposition of negative externalities. 

      But, first, indirect harms can be just as serious as direct ones, and, second, the "directness" of a harm often depends on how we define the interest being harmed. Speech that libels you directly harms your reputation; infringements of your copyright directly interfere with your property right in your work. But neither of these harms actually itself tangibly hurts you. Your life is directly affected only by the consequences of those harms, such as lost business opportunities caused by people thinking badly of you, or lost profits caused by people not buying lawful copies of your book because they have gotten infringing ones. So the real harm to your pocketbook is indirect; we can call it direct only because the law has declared your reputation and your copyright to be protected rights, and these legal rights (and not your tangible assets) are directly affected. 

      On the other hand, the very fact that almost all behavior has some effect on others suggests that a legal regime committed to protection for individual liberty must be willing to ignore some such effects as too small or otherwise irrelevant. And when we shift--as I argue we must when considering the state RFRA regime--from the constitutional question of "When may the legislature regulate conduct to prevent harm to others?" to the common-law question of "When should courts allow certain conduct in the absence of a legislature's conclusion that it harms others?" courts must draw some lines between those externalities that justify restriction and those that don't. Certainly this line-drawing has been seen as a legitimate, even necessary, judicial function in the evolution of the common-law rules of tort, property, contract, evidence, and the like. The same is true for the continuing evolution of religious exemptions. 

      Moreover, indirect claims of harm generally rest on primarily factual assumptions. If the claim is that discrimination based on marital status is just plain wrong, because it violates a moral right not to be discriminated against, then any instance of such discrimination--including religiously motivated discrimination--violates this right. But if the claim is that such discrimination materially interferes with people's ability to find housing, and that the underlying violation of rights is the denial of a supposed right to have plenty of access to places to live, then only those instances of discrimination that actually have this effect constitute rights violations. 

      Thus, in some neighborhoods, where housing is scarce and many landlords dislike unmarried couples, religious exemptions may violate this right to have access to a wide range of housing. In other neighborhoods, though, where there are many vacancies and most landlords don't care about their tenants' sex lives, the occasional religious exemption won't violate such a right. There, it might make more sense to grant the exemption, so long as one thinks that the harm is indirect (denial of a right of access to a wide range housing) rather than direct (denial of a right to freedom from discrimination in each transaction), and that courts will do a good job answering these factual questions.(88)

      So some sort of a direct/indirect distinction may ultimately be sound, and common-law decisionmaking does implicitly draw this sort of distinction at times--consider, for instance, tort law's proximate cause doctrine. But it would be helpful to work out the details of such a distinction, rather than just resting on courts' ability to tell at a glance which interference with another's rights is "direct" enough to bar an exemption claim. 

3. The Test Suite 

      My recommended test suite for this question consists of the following cases: 

      (a) A religious objection to trespass laws, brought by someone who feels a religious obligation to travel onto private property to visit a sacred shrine (for instance, land seen as holy by American Indians, or the site of a visitation of the Virgin Mary).(89)

      (b) A religious objection to copyright infringement laws, brought by someone who feels a religious obligation to disseminate, either in the original form or in a modified one, certain religious teachings that are protected by copyright law.(90)

      (c) A religious objection to laws barring discrimination based on race, sexual orientation, or marital status in housing, brought by someone who feels a religious obligation not to rent to interracial couples, homosexual couples, or unmarried couples.(91)

      (d) A religious objection to bankruptcy law's restrictions on gifts by bankrupt debtors, brought by a bankrupt debtor who wants to donate to a religious institution.(92)

      (e) A religious objection to bans on carrying large knives in public, brought by a Sikh who feels a religious obligation to carry a large kirpan.(93)

      (f) A religious objection to bans on sale or possession of marijuana, brought by a Rastafarian who feels a religious obligation to use marijuana.(94)

B.Paternalistic Interests

      Another categorical proposal is that the government shouldn't have a compelling interest in "protecting the members of the religious community" from harms that flow "from the consequences of their religious choices."(95) Thus, the government would have to exempt from minimum wage laws employees who feel a religious duty to work for free; exempt from social security taxes Amish farmers who believe that support for the aged is the exclusive responsibility of the religious community; and exempt from antidiscrimination laws religious universities which feel a religious obligation to prohibit interracial dating among their students (who are presumably largely members of the same religious group as the university administrators).(96) Similar arguments can be made in favor of exemptions from drug laws and from laws restricting the handling of dangerous snakes. 

      I think this is a promising argument, again especially because exemptions granted under state RFRAs are revisable by legislatures. A court that carves out an exemption from a paternalistic law isn't telling the legislature, "You may not protect people in this situation from their own folly, or protect others from the indirect consequences of such folly." Rather, it's saying, "Per your instructions in the state RFRA, we have concluded that religiously motivated people in this situation should be allowed to do these seemingly foolish things; if you still want to protect them from themselves or protect others from the indirect consequences of this foolishness, of course feel free to pass a law saying so specifically, and we'll dutifully enforce it." 

      And this is a sensible thing to say in many situations, especially because religious commitments which place a high value on extratemporal consequences may make it reasonable for the religious observer to do things that would seem foolish if the person were acting under a solely secular value calculus. I take it that few would think it irrational for someone to take a vow of poverty or chastity if that person sincerely believes that such a vow would have spiritual benefits for him (just as even non-opera-buffs wouldn't think it irrational for someone to spend a lot of money on opera tickets if that person sincerely finds that opera gives him tremendous joy). Likewise, it may not be so irrational for someone to sacrifice some protection against economic dangers or even physical ones if his value calculus places a high value on the extratemporal consequences to be gained from such actions. 

      Nonetheless, any defense of this antipaternalism principle must deal with the three basic defenses of paternalistic justifications, defenses that courts may find persuasive even under a strict scrutiny religious exemption regime. 

      The first defense is the irreversible choice argument. It's true, the argument goes, that the government generally shouldn't protect me against my own choices, especially religious choices. If I choose wrong, I'll have the opportunity to learn my lesson for the future--or I might conclude that the spiritual benefits of the choice exceed the temporal cost, in which case I'll see that the choice isn't really wrong for me, and I can then continue making it. However, some choices, especially ones that risk (or guarantee) death or physical addiction, may preclude future choice; in these cases, the government may properly try to save me from myself today so that I can have the opportunity to live tomorrow and possibly thank the government for its intervention. 

      I'm not sure this argument is sound, but it's certainly endorsed outside the religious context by many actors in our legal system--hence many of the paternalistic laws we see--and any antipaternalism argument must confront it. The classic test cases for this are: (1) religiously motivated suicide by healthy people, such as a widow burning herself on her husband's funeral pyre or the mass suicides of the Jonestown and Heaven's Gate cults; (2) religiously motivated suicide and assisted suicide of the terminally ill; and (3) religiously motivated conduct that seriously risks life, for instance handling poisonous snakes, drinking strychnine,(97) or riding a motorcycle without a helmet.(98) It seems to me that in case (3), courts ought to grant exemptions at least so long as the conduct is on average no riskier than dangerous behavior (perhaps such as skydiving) that the law does not forbid: If the law tolerates risking one's life to get an adrenaline rush, it seems appealing to carve out an exemption (again, subject to legislative revision) for similarly risking one's life for spiritual reasons. Nonetheless, this is a close question, and cases (1) and (2) are harder still. 

      The second defense of paternalism is the pressured choice argument: sometimes the law bars certain voluntary behavior because of the risk that it isn't completely voluntary. One classic example of this is dueling; even those who endorse a right to suicide might want to prohibit dueling, because of the danger that the duelist's decision to risk his life might be unduly pressured by social norms of honor or courage. Similarly, a common argument against assisted suicide is that the seemingly voluntary decision to commit suicide might be the result of pressure from relatives who are greedy or who just want to be done with the ordeal of a loved one dying.(99)

      It would be appealing to argue that the "least restrictive means" of dealing with this is to distinguish those who are being pressured from those who aren't, but this often won't work. The pressure here falls far short of legally punishable coercion and may be largely unspoken; how could courts reliably decide whether a would-be suicide decided to commit suicide "on his own" or "under pressure," especially given that the subject of the proceedings would presumably claim that he really wants to do this (even if he was in fact pressured into it)? Moreover, the inquiry into such "undue pressure" may pose constitutional problems of its own: How is a court to judge, for instance, whether there's undue pressure in saying that "if you don't do this [join us in our cult suicide, spare your family expense and continued pain, handle snakes], you'll [be damned to hell/offend God/commit a sin]"?(100) And yet the concern about pressure is hardly frivolous; we can anticipate that in some substantial number of cases, this pressure will take place. One may therefore argue that the compelling interest in preventing people from being pressured into ending or risking their lives can only be served by a per se prohibition on the life-ending or life-threatening activity. 

      The argument that some claims of a religious right to engage in risky conduct will flow from religious community pressure is reinforced by the fact that many religious communities do routinely "pressure" people into doing what "they really don't want to do." Often this is precisely the sort of pressure for which many praise religion: It's pressure to do right, because of a fear that if one doesn't do right, God will punish you or your community will condemn you. But it can also be a pressure to drink strychnine, to use peyote, to commit ritual suicide, or to work for a religious employer for what would in other contexts be seen by many as an "exploitative" wage. (This might cut against the argument, made earlier, that snake-handling should be permitted if it's less risky than skydiving; it's at least possible that the pressure from a religious group to engage in snake-handling would usually be greater than the pressure from a peer group to engage in skydiving.) 

      These points may help persuade people that paternalism arguments are often legitimate; or they may help persuade them that the whole "undue pressure" justification should be rejected, because even though we can imagine situations where such pressure exists, it's so hard for the law to disentangle bad pressure from good pressure that it's both more libertarian and more egalitarian for the law to accede to private choices. The argument for neither side is open-and-shut, and a thorough examination of when the risk of pressured choice justifies paternalism in religious exemption cases would be extremely helpful. 

      The third defense of certain seemingly paternalistic regulations is the effects on others argument: the law must prevent certain behavior in order to prevent harm to others, not just to the claimant. I've discussed above the general problems raised by harm-to-others arguments, but here I want to mention two extra complexities that arise when we consider harm to others in the context of a rationale that is traditionally seen as paternalistic. 

      The most obvious difficulty is that the paternalistic focus of the main justification for a restriction may obscure the possible effects-on-others rationales that can sometimes support the restriction even when the paternalistic justification is rejected. Consider, for instance, the argument that certain religious objectors should get an exemption from minimum wage laws because "if members of the Alamo religious movement are inspired to work for the glory of God for long hours at no pay, their neighbors are not injured and the government has no legitimate power to intervene."(101) This claim, by exclusively focusing on the most common rationale for minimum wage law (preventing unfair exploitation of workers), ignores the effect of an exemption on competitors and their employees: the concern that subminimum wages "affect many more people than those workers directly at issue in this case and would be likely to exert a general downward pressure on wages in competing businesses."(102)

      One possible response to this--that "the evidence in the [Alamo] case did not show any price cutting"(103)--seems to me to not adequately consider likely business realities. Business is all about competition, and an economically valuable break given to one competitor is likely to hurt other competitors. Even if a business that's exempted from minimum wage laws doesn't cut prices as such, it may increase services while keeping prices the same; and even if at a particular time it might do none of this, the ability to cut prices or increase services will serve it well when competition gets tougher. And this erroneous assertion of "no effects on third parties" in the minimum wage exemption context is emblematic of the more common inclination to ignore subtle third-party effects when trying to rebut what seems to be a primarily paternalistic argument. 

      The better response, I think, would be to admit that there is an effect on third parties, but that the effect is of a kind or a degree such that an exemption should be granted despite it. There really is no right to be free from "unfair competition" by employers who are exempted from minimum wage law, the argument would go, so there's no inherent harm in allowing an occasional exemption. And though there might be a more serious harm if some employers routinely had to face such competition from many other businesses, this is unlikely to happen, since only very few businesses will get this exemption. 

      The second complication arises when the argument isn't just that the government shouldn't protect me from my religious choices, but also that it shouldn't protect my co-religionists from my religious choices. This is the root of the claims that, for instance, "Amish farmers should not be compelled to participate in the government-sponsored social security system when they believe that support for the aged is the exclusive responsibility of the religious community," and that Bob Jones University's discriminatory policies had "direct effects [that were] purely internal to the religious group."(104)

      But at least some of the Bob Jones students, despite their general adherence to the University's religious tenets, might believe that interracial dating is proper, might want to engage in it, and might thus be "harmed" by the Bob Jones policy in the sense that race discrimination is generally seen as a harm.(105) Likewise, at least some employees of Amish farmers, even those who generally adhere to Amish beliefs, might have no religious scruples against participating in social security, and might in fact be thinking about leaving the community in the future and thus losing whatever community-provided benefits they might get. We're certainly familiar with many Catholics who disagree with some Catholic beliefs, or Jews who disagree even with their own rabbis; the same is possible with all sects. Should the law protect such dissident members of the sect against harms--discrimination, "exploitative" employment, and the like--imposed on them by their coreligionists? 

      In any event, all these questions are, I think, best answered by considering a substantial range of test cases, in which one's moral intuitions might at first point in a variety of directions. For instance: 

      (a) A demand for religious exemption from a law banning assisting another in a suicide, brought by a doctor who wants to help a dying patient die, or brought by the patient who wants a doctor's help. 

      (b) A demand for religious exemption from a law authorizing commitment of would-be suicides, brought by a physically healthy person who feels a religious obligation to commit suicide, and a demand for religious exemption from a law banning assisted suicide, brought by someone who helped physically healthy fellow cult members commit suicide. 

      (c) A demand for religious exemption from a law banning the drinking of strychnine (an example of an extremely dangerous activity).(106)

      (d) A demand for religious exemption from a law banning the handling of poisonous snakes (an example of a less dangerous activity).(107)

      (e) A demand for religious exemption from a law requiring motorcyclists to wear a helmet (an example of a less dangerous activity, but one that--unlike in examples (c) and (d)--many people actually do want to engage in).(108)

      (f) A demand for religious exemption from the social security tax, brought by employers who say that it would be sinful for them to pay such a tax and that their employees are very likely to agree and to instead use the religious community's own old-age provisions.(109)

C.Looking to Secular Exceptions from a Law to Decide Whether a Religious Exemption Should Be Granted

      Courts are sometimes urged to decide whether a religious exemption should be granted by considering whether the relevant law has many exceptions for secular conduct. If it does, the argument goes: (1) the law is not really "generally applicable," and thus exemptions are required even under the post-Smith Free Exercise Clause;(110) (2) the government interest can't really be that compelling for RFRA purposes, since if it were compelling then the law would have been written to serve the interest without exception;(111) or (3) granting a religious exemption would be a less restrictive but pretty much equally effective way of serving the interest, since the existence of the secular exceptions shows that modest exclusions don't prevent the law from accomplishing the government's goals.(112)

      I have argued elsewhere that the Free Exercise Clause argument is mistaken;(113) courts should not insist, as a constitutional matter, on religious exemptions from laws that don't discriminate against religion, whether or not the laws contain secular exceptions. But when courts decide these cases under RFRAs, the situation is considerably different. 

      If courts applying RFRAs are acting as common-law-makers--creating rules using their own moral and practical judgment, subject to revision by legislatures--then considering existing secular exceptions makes sense. The common-law process has always proceeded by analogy; when a court considers a proposed common-law defense or exception, we would expect it to consider existing defenses or exceptions. 

      For instance, if a court is deciding whether to recognize a psychotherapist-patient privilege, which is an exception to the general rule that witnesses have a duty to testify, it's quite proper for the court to consider the existence of other privileges. It might be hard for the court to abstractly weigh the policy arguments for and against the privilege, but it's much easier to compare the arguments to those for and against the existing privileges.(114) If the court concludes that the case for a new privilege is at least as strong as the case for the existing ones, then it's reasonable for it to create such a privilege, subject as always to revision by the legislature. 

      Nonetheless, the claim can't be simply that a religious exemption must be granted so long as the law includes a "pattern of exemptions [for secular interests]."(115) This is so for four reasons. 

      First, and most generally, virtually all laws have many exceptions. Few legal principles can be perfectly articulated in a single phrase. Trespass law has exceptions for adverse possession, necessity, law enforcement, and so on. Laws banning race discrimination in employment have many exceptions; breach of contract law has even more. The Copyright Act contains an operative section followed by fifteen sections of exceptions. Statutory rape laws often except acts committed by someone who is close enough in age to the minor, or acts committed by the minor's spouse. Even bans on intentional homicide have exceptions, such as for execution of a lawful sentence, killing in war, police killing of a dangerous fleeing felon, and killing in self-defense or in defense of another.(116) Unless one wants to claim that courts should grant religious exemptions from all these laws and many more, one must accept that not all secular exceptions should lead to religious exemptions. 

      Second, it seems to me that courts generally shouldn't carve out religious exemptions from laws that secure private rights, even when the laws contain secular exceptions. Consider the trespass, breach of contract, copyright infringement, and murder examples in the previous paragraph. The moral claims underlying these private rights may be too complex to be defined in a single phrase, or the law might conclude that the private rights must sometimes give way to certain countervailing interests. But the fact that our private rights are complex or not absolute doesn't mean that they should be further limited to satisfy the religious beliefs of others. 

      Third, courts must look carefully at each secular exception's scope to determine whether it is indeed analogous to the proposed religious exemption. Some exceptions might be narrow, others broad. Some might be limited to cases where the government interest or claim of private right defended by the law is particularly weak--for instance, when the Copyright Act exempts some noncommercial copying that's unlikely to greatly affect the value of a copyright. Others might rest on a legislative judgment that a countervailing interest is especially strong; consider exceptions for medicinal marijuana use. Others might be self-limiting because there are other checks on the excepted conduct. Still others might have been enacted because the legislature assumed that a total ban would be unlikely to be effectively enforced in some cases, and that trying to enforce such a total ban would have undesirable practical side effects. 

      Some proposed religious exemptions might be fairly similar to the existing exceptions and others might be quite different in scope or in justification. It would help to have a better theory of how to perform this comparison, but at least it's clear that the comparison has to look closely at the particular characteristics of the secular exceptions, rather than just at their existence or their number. 

      Fourth, courts will often have to compare the moral significance of an interest that justifies an existing exception with the moral significance of the desire to accommodate religion. Sometimes this comparison may cut in favor of the exemption claim. For instance, if a law restricting certain modes of slaughtering animals "exempts any person slaughtering and selling 'not more than 20 head of cattle nor more than 35 head of hogs per week'"(117)--presumably because the legislature wanted to avoid financial burdens on small farmers--it seems plausible that a similar exemption should be given to people who have religious reasons to slaughter animals in the otherwise-prohibited way. Accommodating felt religious needs is probably at least as worthy a goal as helping small farmers. Of course, this is a moral judgment, but one that under the common-law model judges are free to make, since the legislature can reverse them if its moral judgment is different. 

      But I take it that the existence of a self-defense exception to homicide law doesn't mean that courts should also allow a religious justification exception to homicide law, or even a narrower "religious justification plus consent of the victim" defense for human sacrifice.(118) True, to some religious people protecting one's immortal soul from the sin of defying God's will (the interest that would be served by granting the exemption) is more important than protecting one's life from killing or one's body from serious injury or rape (the interest served by the self-defense exception). But it doesn't follow that the law must take the same view. State RFRAs command that religious motivation be treated with considerable respect by the law, perhaps as much respect as some government interests; but it needn't be treated with as much respect as all government interests.(119)

      In any event, I again don't quite know what the right rule, rules, or even guidelines should be here, but I do think that some of the examples given above form a useful test suite for evaluating proposed rules and guidelines: 

      (a) A person who feels a religious obligation to visit an Indian sacred site or a visitation of the Virgin Mary--both on private land--claims an exemption from trespass law.(120)

      (b) A person who feels a religious obligation to disseminate another's copyrighted religious work claims an exemption from copyright law.(121)

      (c) A person who feels a religious obligation (i) not to testify against one's kin or (ii) not to testify in civil court against any co-religionist claims an exemption from the duty to testify.(122)

      (d) A person who feels a religious obligation to use (i) peyote, (ii) marijuana (especially in a state where marijuana laws have an exception for medical use), or (iii) cocaine claims an exemption from the drug laws.(123)

      (e) A person who feels a religious obligation to use alcohol at a religious gathering in a church claims an exemption from a law that bans alcohol consumption outside the home but has exceptions for many secular activities.(124)

      (f) A person who feels a religious obligation to discriminate based on (i) race in employment or (ii) marital status in housing claims an exemption from antidiscrimination law.(125)

III. Government as Employer, Proprietor, or Educator 

A.Government as Employer

1. What Is the Test? 

      Under free speech law, the government acting as employer has far more authority to restrict people's speech than does the government acting as sovereign. First, content-based restrictions on government employee speech need not pass strict scrutiny, but only the far weaker Connick/Pickering(126) test. Second, as to speech that is actually part of the employee's job description, where the employee is speaking on behalf of the government, the government might have virtually unlimited power to control what the employee says (though that's not completely clear).(127) In the Sherbert era, at least some courts took a similar view as to religious exemption cases, reasoning that the government acting as employer had extra authority to compel its employees to do certain things, even if these commands incidentally violated the employees' religious beliefs.(128)

      RFRAs, though, are quite explicit: Whenever the government incidentally burdens religious exercise--and if loss of unemployment benefits qualifies, presumably so would the loss of a job--the government action must pass strict scrutiny, with no exceptions for the government acting in special capacities.(129) Thus, for instance, if a government employee feels religiously obligated to berate his coworkers for doing things that he considers sinful, the government must tolerate this even if such an accommodation causes offense and morale problems; it seems hard to see how avoiding such problems can be said to be a "compelling" government interest.(130) One of the reasons the Connick/Pickering test implements a sub-strict scrutiny standard for the government acting as employer is precisely that under traditional government-as-sovereign strict scrutiny review, preventing offense to coworkers or clients wouldn't be a compelling interest.(131)

      Similarly, if a police officer refuses on sincere religious grounds to guard an abortion clinic,(132) the government might also have to accommodate him unless the court concludes that (as in the Free Speech Clause context) there's no substantial burden when the government insists that you do the job you're paid to do. While there's a compelling interest in preventing crimes against abortion clinics, the less restrictive means of serving the interest would be to simply assign another police officer to this duty. 

      Could these results be right? Here are a few possible answers: 

      (1) Whether you like these results or not, RFRAs require them. 

      (2) These results are not mandated, because the government does have a compelling interest in saving money or preserving morale. This, though, seems possible only if one tremendously waters down the compelling interest prong, which would end up watering down the test in the government-as-sovereign context, too, and this strikes me as the wrong outcome: The two religious exemption claims described above should lose precisely because the government should be able to require its employees, who are paid to perform certain tasks, to act in certain ways even if it couldn't impose the same requirement on private citizens. 

      (3) Telling an employee "We will employ you to do jobs X, Y, and Z, but if you refuse to do Z (such as guarding abortion clinics), we won't continue employing you" is not a substantial burden. In the Free Speech Clause context, the Court has generally held that there's no free speech burden when the government offers money to pay for certain government-endorsed speech but not other speech. In such a situation, "the Government is not denying a benefit to anyone"--or at least is not unconstitutionally denying a benefit--but is instead simply insisting that public funds be spent for the purposes for which they were authorized."(133) Likewise, the argument would go, there's no burden on religious practice when the government offers to employ someone to do a certain job that violates some people's religious beliefs: the government is simply insisting that people who get public funds do the things for which the funds were allocated. 

      This argument might take care of the police officer hypothetical, though it probably won't take care of the rude coworker hypothetical, where the free speech jurisprudence acknowledges that analogous speech restrictions do impose a burden on free speech rights (albeit one that's judged only by the Connick/Pickering standard). 

      (4) The RFRAs need not be interpreted literally. Legislatures that were enacting RFRAs, the argument would go, weren't thinking of the special situation of the government acting as employer, but were primarily focused on the government acting as sovereign.(134) Especially given the declarations in some RFRAs that generally endorse the pre-Smith case law, and the language of "restoration" of this case law, the strict scrutiny test might just have been meant as shorthand for "the pre-Smith regime, which was mostly strict scrutiny but sometimes something else"; thus, courts should continue refining this regime without being bound to the literal terms of strict scrutiny. 

      (5) Even if the statute is read literally as establishing a right to be free from substantial burdens on one's religion unless the burdens pass strict scrutiny, the government acting as employer may demand that employees surrender some part of this right as a condition of getting a government job. Thus, for instance, the government may say: "We will hire you only on the condition that you act civilly to one another, even if this means refraining from doing what your religion requires." 

      True, such a condition must itself be scrutinized by the courts, by analogy to the unconstitutional conditions doctrine, because the government doesn't have a completely free hand in demanding that people surrender their rights in order to get a government benefit. But the unconstitutional conditions doctrine does leave the government with more power to attach some conditions to benefits than the government has in unilaterally imposing commands when acting as sovereign. Thus, the argument would go, creating a special level of scrutiny for burdens on religious practice imposed by the government acting as employer is faithful to the RFRA text: The text does create a right, but a right that the government can, within limits, insist that the employee sign away if he wants to get a government paycheck. 

      I've tried to make a sympathetic case in the above paragraphs for interpreting RFRAs as providing a lower standard of review in government-as-employer cases, but the case is far from open-and-shut. Imposing any standard short of strict scrutiny runs enough against the letter of the law that such a result requires a serious defense--as does the position that the RFRAs must mean what they say, and thus seriously constrain government employers despite the oddity of the results that this might create. 

2. If Not Strict Scrutiny, Then What? 

      Let's say the case for a lower standard of review is made. What should this standard be? 

      One obvious answer is to borrow it from the free speech context, but two of the Connick/Pickering test's three components aren't easily adaptable to religious exemptions. Under Connick and Pickering, the government: (1) may restrict any speech on matters of purely private concern; and (2) may restrict speech on matters of public concern; if (a) the damage caused by the speech to the efficiency of the government agency's operation outweighs (b) the value of the speech to the employee and the public. Element (1) is inapplicable to religious exemption claims, since religious beliefs are not usefully divided into those of "public concern" and "private concern." The same is probably true as to element (2)(b): Secular courts can't estimate the "value" of religious behavior to a religious observer--for instance, by trying to determine whether or not the practice is "central" to the observer's religion--at least beyond determining whether the observer sincerely believes the practice is religiously compelled or motivated (the inquiry required by the substantial burden prong).(135)

      It's conceivable that courts might take a different view in government-as-employer cases. After all, determining how "valuable" speech is to a speaker or to the public--rather than letting the speaker and the public make this decision themselves--and then subjectively weighing this value against the government interest is itself in some tension with elements of standard free speech analysis,(136) but the necessities of the government-as-employer context persuaded the Court that such inquiries were called for. Perhaps one can similarly argue that an inquiry into the value or the centrality of a practice to the observer, though improper as a general matter, is more sensible when the government is acting as employer. Still, such a position would need a pretty powerful defense. 

      But if courts try to borrow the Connick/Pickering inquiry without elements (1) and (2)(b), then all they will be borrowing is an inquiry into the degree of the damage caused by the religious practice to the efficiency of the government agency's operation--and it won't even be clear what this damage is to be compared against. The borrow-from-free-speech-law solution, then, needs a good deal of work. 

      Alternatively, courts could borrow from Title VII's religious accommodation case law. Under Title VII, all employers, governmental or otherwise, must accommodate employees' religious practices so long as the accommodation requires only "a de minimis cost" to the employer and does not impose any significant burden on coworkers.(137) And since most exemption claims against government employers will probably be brought under Title VII as well as under a RFRA, applying the same test under both statutes would simplify such cases. 

      Moreover, the Title VII test is in some ways analogous to the "least restrictive means" part of the textual RFRA test, though with the "compelling interest" prong dramatically relaxed: Under the accommodation requirement, the government may not burden an employee's religious practice unless there's no "less restrictive means"--i.e., no reasonable accommodation--that would cheaply and equally well serve the government's interest in efficient operation. And the Title VII test may also be in some ways similar to the portion of the Connick/Pickering test that can survive the shift from free speech to religious exemptions. The accommodation requirement may be seen as a weighing of the damage caused by the religious practice to the efficiency of the government agency's operation (what I call element (2)(a)) against a unitary standard of "de minimis cost." Just as a government employer may punish employee speech if the burden of the speech on the employer's operation is too great, so a government employer may punish an employee's religiously motivated conduct if the burden of accommodating the conduct is more than de minimis. 

      Nonetheless, the case for borrowing this standard also needs a further defense. Title VII's duty to make only very cheap accommodations seems only a pale shadow of the RFRAs' use of strict scrutiny, on its face one of the most rights-protective tests in constitutional law. Moreover, the Title VII standard was selected with an eye towards the burden that it would impose on private employers, and may thus be relatively weak because of the need to prevent employees from imposing the costs of their religion on another private party.(138) RFRAs, which are aimed at constraining government interference with religion, may plausibly be seen as demanding greater accommodation from the government. Again, the right answer is far from clear, but such an answer is necessary. 

3. But What Do These Formulae Really Mean? 

      Perhaps, though, it's a mistake to focus too much on a search for the right formula, be it strict scrutiny, Connick/Pickering (with the proper adaptations), or "reasonable accommodation." All these supposed "tests" may be too vague to provide any really predictable results; both Sherbert-era strict scrutiny in religious exemption cases generally and Connick/Pickering in government employee free speech cases have been condemned on these grounds.(139) Maybe it's better to look more to the kinds of ways in which religious practices may interfere with the government employer's interests, and identify more precise tests for each of these situations. 

a.Requests for exemptions from parts of one's job requirements

      As I've mentioned above, some religious objectors, such as the police officer who objects to guarding abortion clinics and the employee who objects to working on certain days,(140) are asking more than the free speech claimants in the Connick/Pickering cases ever did. The free speech claimants are willing to do their jobs, but also want to say certain things, either while doing the jobs or outside office hours. The police officer and the Sabbatarian actually want to not do parts of their jobs; the free speech analogy might be an employee whose job requires him to say certain things to the public--a post office clerk, a radio announcer, and so on--but who refuses on the grounds that he finds those statements objectionable. The Connick/Pickering test probably doesn't even generally apply to such refusals to say what you're paid to say.(141)

      One might by analogy argue that state RFRAs should similarly not apply to refusals to perform part of one's assigned job tasks. The government's insistence that I do what I'm paid to do should not be seen as a "substantial burden," or, even if it is a substantial burden, it should be a permissible condition for the government to insist that to get a certain job I must waive any RFRA rights to refuse to do the tasks that would be assigned to me. 

      I might still have some rights to an accommodation under Title VII, which has been interpreted to cover even requests for exemptions from one's assigned job duties. This, though, is because Title VII is an equality rule rather than a liberty rule; the "religious accommodation" provision actually forbids employers "to refuse to hire or to discharge [or discriminate against] any individual . . . because of such individual's . . . religion," "includ[ing] all aspects of religious observance and practice"(142)--firing a police officer because he refuses to do part of his job for religious reasons is thus seen as discharging an individual because of his religious observance. 

      Antidiscrimination rules generally apply without regard to any substantial burden threshold; even discrimination in provision of purely discretionary government benefits violates an antidiscrimination rule. RFRAs, though, are liberty rules, and contain a substantial burden prong--it makes sense that their scope would thus be narrower than the scope of Title VII's equality rule, even if the strength of their protection within that scope (strict scrutiny rather than reasonable accommodation) is greater. 

      But if this analysis is rejected, and RFRAs are interpreted to require exemptions from at least some job requirements, a helpful test suite for such a proposal might include: 

      (a) A police officer who refuses, for religious grounds, to defend abortion clinics. 

      (b) A postal worker who refuses, for religious grounds, to process draft registration forms. 

      (c) A nurse at a government-run hospital who refuses, on religious grounds, to assist in abortions. 

      (d) An employee at a government facility whose duties include raising and lowering a flag, but who refuses, on religious grounds, to do so. 

      (e) A postal worker who refuses, on religious grounds, to deliver mail that he considers blasphemous. 

      (f) A bus driver who refuses, on religious grounds, "to hand out coupons for free hamburgers as part of a promotion to boost ridership."(143)

b.Requests for exemptions from general courtesy and professionalism rules

      Another area that might deserve special analysis is employee requests to act in ways that would normally be seen as unprofessional, rude, or annoying, and that the government claims interfere with morale or customer relations. This can include employees who feel religiously obligated to wear long hair or a beard, contrary to rules premised on the notion that a short-haired or clean-shaven look is more "professional";(144) who feel religiously barred from shaking hands with people of the opposite sex, which some clients or coworkers might find rude;(145) who feel a religious obligation to wear anti-abortion pins or religious symbols that coworkers might find rude or offensive;(146) who feel religiously motivated to say "God bless you" or "Praise the Lord" to patrons, which some patrons might find intrusive;(147) who feel religiously moved to criticize their coworkers' lifestyles;(148) and the like.(149)

      This set of cases raises several common issues. First, they share a relatively similar government interest that is probably hard to call "compelling" but that may deserve consideration when the government is acting as employer. 

      Second, in at least some situations the perceived unprofessionalism or rudeness of certain behavior might seem considerably less once people recognize that the behavior is religiously motivated. If I refuse to shake hands with you, you might plausibly assume that I'm trying to slight you; but if I politely explain that I'm acting this way because of a religious belief that doesn't rest on any condemnation of you, you'll probably feel considerably less bothered. On the other hand, this need not be the case: Consider one woman's account of "quiver[ing] with anger as she remembers the Orthodox rabbi who refused to shake her hand . . . . 'I held out my hand and he completely ignored it. They told me afterwards that he wasn't allowed to shake it because I was a woman . . . . I was unclean!'"(150) As I understand it, many Orthodox Jews would generally respond that the bar on handshaking doesn't rest on the notion that women are unclean, but as is often the case in questions of manners, observers' interpretations of an action may not match the actor's intentions. 

      Third, some of the objections that clients or coworkers may have to the behavior could rest on dislike of a particular religion, and one can argue that government employers should not ratify such religious hostility by requiring the employee to yield to it.(151) Thus, one might argue that the government ought not be able to interfere with an employee's wearing a Star of David, a pentagram,(152) or a religious swastika,(153) even if the symbol is causing serious tension with coworkers or is driving away clients in a situation where the government is competing for public patronage. 

      On the other hand, it seems unsound to insist that the government ignore all hostility to an employee's religiously motivated conduct. Some behavior--for instance, wearing, for religious reasons, a button that says "Accept Jesus or you'll burn in Hell" or that says "Ban Religion" and contains a stop sign printed over a cross(154)--seems likely enough to insult customers and coworkers, and to make it harder for the office to do its job, that the government acting as employer ought to be able to stop its employees from acting this way. The same may be true as to people wearing swastikas, no matter what the motivations for them. But if wearing the buttons or the swastikas is to be treated differently from, say, wearing a pentagram, then the law will have to draw the line between "genuinely rude" behavior that is "reasonable" for coworkers or patrons to object to, and the "not really rude" behavior that's "unreasonable" or "intolerant" for coworkers or patrons to object to. 

      I'm not sure where this line should be drawn, or even whether it can be drawn in a judicially administrable way. Perhaps the difficulties with drawing it should lead courts to focus solely on the magnitude of actual or plausibly predicted public reaction, without regard for whether or not the reaction is improperly "intolerant." But inquiring into these tough questions seems ultimately more productive than simply debating whether the right test is strict scrutiny, Pickering, or reasonable accommodation. 

      A possible test suite here would include: 

      (a) A male government employee who feels a religious obligation to wear (i) long hair or (ii) a beard.(155)

      (b) A government employee who feels a religious obligation (for instance, because she has taken a vow to this effect) to wear, exposed to coworkers, a sign with a picture of a dead fetus and the words (i) "Abortion Is Murder" or (ii) "Mothers Who Get Abortions Are Murderers."(156)

      (c) A government employee who is in a job that normally involves handshakes with customers, but who feels religiously obligated not to shake the hands of customers of the opposite sex.(157)

      (d) A government employee who feels a religious obligation or motivation to say "God Bless You" or "Praise the Lord" to patrons whom he serves.(158)

      (e) A government employee who feels a religious obligation to wear (i) a cross, (ii) a Star of David, (iii) a pentagram,(159) (iv) a swastika as a Hindu religious symbol,(160) or (v) a swastika as a religious symbol of an Aryan Nation religious group that views non-Whites and Jews as inferior.(161)

      (f) A government employee who feels a religious obligation to admonish coworkers who behave immorally, either (i) pointing out, from a religious motivation but without using explicitly religious language, the moral flaws in their actions, or (ii) explicitly exhorting them to abide by God's law.(162)

c.Requests for exemptions for conduct that may diminish public confidence in the government employer

      It may also be profitable to consider, as a group, cases where the government employer fears that an employee's religiously motivated conduct may diminish public confidence in the employer. Let me provide the test suite up front; assume in all cases that the employer has a general policy banning the relevant conduct, whether religiously motivated or not, and the claim is that the religious objectors should be exempted from this policy. 

      (a) A prosecutor's office in a state where (i) homosexual conduct or (ii) polygamy is illegal fires someone for engaging in a religiously motivated homosexual or polygamous marriage. The justification given is that "the employment in this case is . . . a kind of employment in which appearances and public perceptions and public confidence count a lot," and though "[p]ublic perception is . . . not knowable precisely," the objector's "acts [are] likely to cause the public to be confused and to question the [office's] credibility."(163)

      (b) A mayor fires a high-ranking employee at an agency that enforces, among other things, a law banning sexual orientation discrimination, for making a religiously motivated speech condemning homosexuality as an "abomination." The justification given is that the city has "a compelling governmental interest [in] the preservation of the integrity of its antidiscrimination policies" and the firing is "the only effective way to remedy the damage [the employee's] statements . . . wreaked on the credibility of those policies."(164)

      (c) A mayor fires a high-ranking employee at an agency that enforces, among other things, a law banning religious discrimination, for stating in a sermon that those who don't accept Christ will be condemned to hell. The justification given is that the city has a compelling government interest in the preservation of the integrity of its antidiscrimination policies, and the firing is the only effective way to remedy the damage the employee's statements wreaked on the credibility of those policies.(165)

      (d) Returning to a case I've mentioned before, a police department refuses to honor a police officer's request to be transferred from duties that require him to guard abortion clinics. The justification given is that "[t]he importance of public confidence in the neutrality of its protectors is so great that a police department or fire department or equivalent public-safety agency that decides not to allow recusal by its employees should be able to . . . escape any duty of accommodation."(166)

      (e) A government-run drug rehabilitation agency fires employees who it learns overtly use peyote and marijuana for religious purposes.(167) The justifications given are that (i) the addicts whom the employees are supposed to counsel may question the employees' "just say no to drugs" zero-tolerance advice when they know that the employees don't practice what they preach, and (ii) the public will doubt the employees' zeal in urging the zero-tolerance policy. 

      (f) A police department fires a clerical employee for being a known member of a racist (either white supremacist or racist black nationalist) religious group. The justification given is that, though the employee had "performed his duties in exemplary fashion," his continued employment "would severely impede the progress made in nurturing trust between [various racial groups] and the Sheriff's office."(168)

      These cases share several important characteristics. First, in each of them it is possible for the government employee to zealously do his job given the accommodation. Someone who opposes homosexuality certainly can assiduously enforce gay rights laws if allowed to remain at the civil rights agency. A prosecutor who is homosexual certainly can, if ordered to do so, vigorously prosecute people who violate sodomy laws. Granting a police officer a religious exemption from duty that he finds immoral may actually increase the chances that he will zealously do his job. 

      Second, rational citizens may nonetheless be skeptical that the objectors will indeed vigorously perform all their duties. There's nothing bigoted or unreasonable in doubting that someone who opposes homosexuality will assiduously enforce gay rights laws, that someone who is homosexual will forcefully prosecute people who violate laws against homosexual conduct, or that a drug-using counselor will be effective at conveying a zero-tolerance message to his clients. 

      The risk of half-hearted enforcement may be smaller if the employee can be exempted from enforcing laws of which he disapproves--if the prosecutor doesn't assign sodomy cases to the gay employee (not as a matter of religious accommodation but just as a matter of management discretion), or if the police department allows the objecting police officer to not cover abortion clinics. But even with these accommodations, the public may remain skeptical. If there's an emergency at an abortion clinic or at the home of someone who turns out to be an abortion provider and the police officer is summoned to attend to it, will he do his best? Will a prosecutor who believes it's proper to commit one felony also be more sympathetic to violators of other laws that he considers unjust? 

      Third, in some of the cases, some of the public's concerns may in theory be palliated by making clear that the conduct is religiously motivated, that the employee promises to do his duty outside the zone of the exemption, and that the employee's managers will oversee the employee carefully to make sure that happens. If the police chief says, "Yes, out of respect for people's religions, the law requires us to accommodate this police officer's religious objections, but he promises to set his religion aside--as all of us promise to set our prejudices aside--when doing the rest of his job," some members of the public may accept this. Likewise, if the drug rehabilitation counselor who is challenged by a recovering addict says, "Yes, I use peyote, but only for religious purposes and under circumscribed rules, and my religion tells me that I should never use other drugs," the addict might be persuaded. 

      But, fourth, in all these cases the public's concerns may remain even after the justifications and the narrow scope of the religious exemption are made clear. The public might not trust, for instance, that the police employees will in fact set their biases aside; and while perhaps the public should have known all along that government employees have various biases, there's a difference between knowing this in the abstract and being confronted with specific evidence about specific employees. The government employer might therefore quite rationally be concerned with the effect of the public's perceptions on public confidence in the department. 

      This strikes me as an especially thorny problem, and while it seems to me that RFRAs require exemptions to be granted in some of these cases but not in others, it's not clear what exactly the distinction should be. But focusing on a test suite containing cases that are viscerally appealing to various political viewpoints should be particularly helpful here. 

B.Government as Proprietor

      Just as the government generally has more power over employees' speech, so it has more power over the speech of people who are using the government's property. Outside traditional public fora such as streets and parks, and outside places that the government has intentionally devoted to public speech (so-called designated public fora), the government may generally restrict speech, even based on its content, so long as the restriction is viewpoint-neutral and reasonable.(169)

      As with employment, though, the RFRAs' text doesn't seem to give the government this sort of extra discretion where religious exercise is concerned; the text calls for strict scrutiny without regard to the government's role as proprietor. A ban on soliciting money in airports, for instance, is within the government's power under the Free Speech Clause, because it is viewpoint-based and, in the Court's judgment, reasonable.(170) On the other hand, it's far from clear that the ban would pass strict scrutiny, so if the solicitation is religiously motivated (or compelled) and the RFRAs' strict scrutiny applies, the government would presumably have to allow it. Can this be right? 

1. What Constitutes a "Substantial Burden"? 

      As in the other sections, I want to highlight the important questions, rather than try to answer them. And the first question under the RFRA inquiry is: When does a regulation imposed by the government as proprietor constitute a substantial burden? 

a.Using property in ways that create spiritual harm

      There are at least four ways in which the government can affect religious practice in its capacity as proprietor. The first is when the government uses its property in a way that some believe spiritually harms them: The classic examples are Lyng v. Northwest Indian Cemetery Protective Ass'n,(171) where the government built a logging road through an Indian sacred ground, and Bowen v. Roy,(172) where the government recorded a child's social security number--believed by the child's parents to be "the mark of the beast"--in the child's files. 

      In these cases, the religious observers were doubtless sincerely concerned that the government's actions imposed religious harms on them, but the government was not using legal coercion to compel or even pressure people to violate their religion's commands. The Supreme Court concluded that the government's actions were therefore not substantial burdens on religious practice; for RFRA purposes, the threshold question here would be whether the RFRAs codify these holdings (in my view the likely conclusion), or whether state courts remain free to read "substantial burden" broadly enough to cover this behavior.(173)

b.Setting up property in ways that make it unusable by believers

      The second way the government as proprietor can affect religious beliefs is by creating programs or places that members of certain religious groups can't effectively use. The government might put up an electronic gate at a public housing complex, which keeps an Orthodox Jewish tenant from being able to easily enter and exit on the Sabbath.(174) A government-run cafeteria might not provide vegetarian, kosher, or halal dishes, which means that people of certain religions who use the building can't take advantage of the cafeteria's services.(175) A park might contain sculptures or signs that some people see as indecent or blasphemous, and those people may conclude that it would be spiritually harmful for them to be in a place that contains such material.(176) And a government-provided education program--such as the public schools themselves--might be conducted in a way that makes it impossible for people of certain religions to take advantage of it.(177)

      Here, there are two burden questions: Is there a "substantial burden" here at all, given that the government isn't making it legally impossible for anyone to do anything, but only making it religiously impossible for them to take advantage of the property? And even if the answer to the first question is potentially yes, are certain government benefits--for instance, the ability to eat in a certain cafeteria--so slight that it is not a substantial burden to provide them in ways that certain religions can't use? 

c. Restricting behavior on government property

      When the government as proprietor imposes neutral rules that bar or require certain behavior on government property, then the substantial burden test seems easier to meet. In such a case, the government is indeed requiring people to do what their religion forbids, or to not do what their religion compels, at least on that particular piece of property. Thus, for instance, government bans on solicitation(178) or demonstrations on government property do coercively restrict the behavior of those whose religion moves them to solicit or demonstrate. Likewise, government attempts to keep anyone (other than government employees) from going onto certain property coercively restrict the behavior of those who feel religiously moved to travel to that property, for instance because it's a sacred American Indian site or the location of a visitation of the Virgin Mary. 

      There still remains, though, the question whether this coercion should be seen as a substantial burden for RFRA purposes. (I assume here that the claimant sincerely believes that he ought to follow the urgings of his religion on this government property, and doesn't feel that doing it elsewhere would be an adequate substitute.) One can at least argue that there is generally no substantial burden when the government opens its property for only a particular purpose, and then insists that people who use that property use it for that purpose and not for purposes of their own. 

      This might explain why the Free Speech Clause requirements for restrictions in nonpublic fora are only reasonableness--a bit more demanding than a rational basis test, but not much(179)--plus viewpoint-neutral. Such a rule may reflect the judgment that allowing only certain behavior on government property is a generally permissible condition on a benefit rather than a burden, and therefore the only concern is that this benefit be distributed without viewpoint discrimination. Of course, one might argue in response that any attempt by the government to fence off some land as its own property is an interference with the liberty of those who want to travel there, and is thus a substantial burden on their rights.(180)

d.Restricting behavior on government property in ways that make it unusable by believers

      Finally, the government as proprietor can also affect religious beliefs through neutral rules that both prohibit certain behavior and thereby make the property unusable by certain religious believers even when they want to use it for the very purpose for which it is intended. Thus, for instance, a ban on wearing kirpans in a hospital may stop male Sikhs from using the hospital as patients,(181) and a rule requiring visitors to certain government buildings to wear photo badges may keep those whose religions forbid having their photos taken from using the building altogether.(182)

      Here, unlike in the second category (the government simply setting up the forum in a way that makes it useless to religious observers), the government is indeed coercively restricting people's behavior; and unlike in the third category (the government simply restricting behavior on government property), the burden on religious observers may be greater than that imposed on speakers by nonpublic forum speech restrictions. A ban on demonstrating in a government office building or soliciting money in an airport doesn't prevent the speakers from using the office building or the airport as normal citizens or travelers--but a knife ban or a photo badge requirement does completely deprive the observers of access to the government property involved. Especially when the benefit involved is quite significant, the burden may thus be quite substantial. 

2. What Is the Proper Level of Scrutiny? 

      If a restriction imposed by the government acting as proprietor does constitute a substantial burden, the next question is what the right level of scrutiny should be. One possibility, of course, is strict scrutiny (as per the RFRA text), though one can argue that this is far too strong, and that the government should be able to do its job without having to grant access to its property to people who want to engage in religious practices that interfere with that job. 

      Another is a reasonableness test for government property that isn't traditionally open to the public, and Ward v. Rock Against Racism(183) intermediate scrutiny for government property that is traditionally open, by analogy to the Free Speech Clause doctrine. This treats religious freedom the same way that free speech is treated, and helps avoid the concerns about discrimination between religiously motivated and secularly motivated speech activities discussed in Part I.C, which may be especially useful because many government-as-proprietor cases involve religious activities that are also speech. On the other hand, this test might be too weak (albeit a bit stronger than the rational basis test(184)), especially when applied to conduct restrictions that have the effect of denying someone the opportunity to use government property for the normal purposes for which other citizens use it. And there might still be other alternatives, if one concludes that the RFRA text permits them. 

      Again, I don't know the right answer, but I hope that some of the examples discussed above can form a good test suite for evaluating any possible proposals: 

      (a) An Orthodox Jewish tenant in a public housing complex objects to the government putting up an electronic gate, which keeps him from being able to easily enter and exit on the Sabbath.(185)

      (b) A person with a religiously mandated diet objects that (i) a government building cafeteria or (ii) a government-run hospital fails to provide vegetarian, kosher, or halal meals.(186)

      (c) Someone who is religiously forbidden to be exposed to certain material (because it's blasphemous or indecent) objects to the government placing such material in (i) a government building that he must use or (ii) a public park that he wants to use.(187)

      (d) A person who feels a religious duty to solicit money objects to a ban on such solicitation in a government-run airport.(188)

      (e) Someone who feels a religious duty to hold a prayer vigil in the hallway outside a legislator's office objects to a ban on demonstrations in such places. 

      (f) A Sikh who feels a religious duty to possess a long, sharp dagger objects to a ban on deadly weapons in a government-run hospital.(189)

      (g) A person who feels a religious duty not to let photos of himself be taken objects to a photo badge requirement in a government building (i) where threats of violence are fairly rare, or (ii) where threats of violence are not so rare.(190)

C.Government as K-12 Educator(191)

1. What Is the Proper Level of Scrutiny? 

      Kindergarten through twelfth grade education, just like government employment, is full of requirements that students listen, write, and act in ways the school requires. School authorities often regulate clothing (for gym classes even if not for other classes), hair length, and personal possessions, and they constantly decide what students must hear and say in the classroom. 

      Speech restrictions imposed by the government acting as K-12 educator are commonly said to be governed by the Tinker v. Des Moines(192) test, under which the government may restrict speech when there's concrete reason to believe that it will "materially and substantially interfere with the requirements of appropriate discipline in the operation of the school," or when it "inva[des] the rights of others."(193) This, though, covers out-of-class speech and passive in-class speech, such as pins or armbands. As to speech that's part of the classwork--answers the student gives to teachers' questions, papers the student writes as assignments--the government may have more power still: Students can be and are often marked down for expressing ideas that the instructor believes are illogical or unsound, and thus are pressured into expressing adherence to ideas that they might otherwise think are wrong. West Virginia Board of Educationv. Barnette(194) suggests that there might be some limit to what the school may insist that students say, but as a general matter schools compel classwork speech and punish nonconforming classwork speech all the time.(195)

      As in the government-as-employer and proprietor contexts, state RFRAs contain no exception for the government as K-12 educator. Moreover, the argument that the government is just imposing a condition on a government benefit--"if you want a government-provided K-12 education, you'll just have to accept this restriction"--doesn't work, because education (at least until a certain age) is legally compulsory. One can still theorize that legislatures didn't mean to require strict scrutiny in the government-as-K-12-educator context because they only meant to "restore" the pre-Smith case law, but this argument too is weaker than in the government-as-employer context: The few pre-Smith decisions in this area tended to apply strict scrutiny rather than Tinker or any other test.(196)

      So the question of what the proper test should be remains open; but, as in the other areas, the harder and more interesting questions may arise when we try to go beyond the vague contours of "compelling government interest" or "materially and substantially interfere with the requirements of appropriate discipline." And again it might be profitable to look separately at two different categories of exemption requests, which I sketch very briefly below. 

2. Religious Objections to Required Classes 

      Consider the following test suite: Students (or, more likely, the students' parents) claim that their religions prohibit them from: 

      (a) engaging in compulsory military training;(197)

      (b) engaging in co-ed dance class;(198)

      (c) engaging in co-ed physical education (because this will involve wearing clothing that they think indecent);(199)

      (d) taking AIDS education;(200)

      (e) being in classes where audiovisual equipment is used;(201)

      (f) taking sex education;(202)

      (g) learning about evolution;(203) or 

      (h) reading textbooks that contain passages to which they believe it would be religiously wrong for them to be exposed.(204)

      In cases (d) through (h), assume that the very exposure to the offensive items or teachings is seen as religiously forbidden; such mandated exposure would therefore constitute a substantial burden, because it would require people to engage in certain actions (listening, watching, or reading certain things) that their religion says are impermissible.(205) In each case, the students or the parents ask that the students be excused from the class. 

      Excusing the students would not directly hurt the education of the other students in the same way that offensive student speech or offensive religious conduct might. The government's concerns are generally that exemptions might deprive students of knowledge that we think would be valuable for them; that the students' ignorance may indirectly hurt society (a common argument in AIDS education and sex education cases);(206) that setting up alternatives short of total exclusion of the students from the class may be administratively burdensome or might foster "religious divisiveness";(207) and that excluding the students from certain parts of a class may make it hard to fairly grade them. When are these concerns compelling enough to justify a school's denying the exemption request? 

3. Religious Objections to Conduct Rules 

      Some students or parents may object to rules of conduct that, for instance: 

      (a) require the wearing of school uniforms;(208)

      (b) restrict long hair;(209)

      (c) bar certain classes of jewelry or insignia on the grounds that they might be used as gang symbols;(210)

      (d) bar certain insignia (such as pentagrams, religious swastikas, or signs that convey criticism of other religious groups) on the grounds that they are offensive to classmates and thus disruptive of schoolwork;(211) or 

      (e) bar the carrying of knives.(212)

      Here, the school's concern is generally that the conduct will either disrupt studies by distracting classmates, causing fights, or making classmates feel threatened, or will cause a longer-term erosion of standards of discipline and obedience. In some measure, the fate of each class of regulation may turn on facts related to that regulation, or even to the experience and environment of the particular school that's adopting the regulation. On the other hand, there are likely to be some common questions that can profitably be considered together: How much evidence is required to support a prediction of disruption? To what extent can the school rely on predictions about a general decline in obedience and discipline that would flow from granting any exemption--predictions which will necessarily be hard to prove? Must the school first try to educate classmates about the religious practice in order to minimize the risk that misperception of the practice will lead to disruption, and then deny the exemption only if this less restrictive alternative fails? 


      Those who support the RFRA religious exemption regime should be sobered by the fact that another religious exemption regime had already been tried and, after three decades of experience, rejected. And this rejection wasn't just a 5-4 vote by a single result-oriented Court majority. For years before Smith, the Court had applied strict scrutiny in ways that were indeterminate, inconsistent, and disingenuous; even exemption supporters have criticized the Court's jurisprudence in this area on these grounds.(213) Moreover, of the five Justices in the Smith majority, only two (Rehnquist and Stevens) had been historic opponents of the exemption regime, and one (White) had historically been a fairly strong supporter, even casting the most pro-exemption vote of any of the Justices in Bowen v. Roy,(214) only four years before Smith. Perhaps the problems with applying the Court's doctrine caused Justice White and some of the others to become disaffected with it. 

      This suggests to me that RFRAs are more likely to succeed if there's a body of thought-through, relatively determinate, well-supported doctrinal proposals that courts can draw on. While judges may sometimes tolerate ad hoc, particularistic decisions when their downstream precedential effects are unclear, I suspect judges generally don't really like them. Given the choice, most judges would feel much better about clearer, more rule-like solutions that are more predictable in their effects and that provide more guidance to lower courts and to government decisionmakers. And one of the functions of legal scholarship is to try to fill this need. 

(*Professor, UCLA School of Law ( Many thanks to Alan Brownstein, Rick Duncan, Rick Garnett, Sam Levine, Vladimir Volokh, and Jim Weinstein for their help. At the author's request, selected footnotes and text quotations do not conform precisely to The Bluebook. 

1 42 U.S.C. §§ 2000bb to 2000bb-4 (1994). 

2 See Ariz. Rev. Stat. § 41-1493.01 (1999); Conn. Gen. Stat. Ann. § 52-571b (West 1998); 1998 Fla. Sess. Law Serv. ch. 98-412 (West); 775 Ill. Comp. Stat. Ann. 35/10 (West 1999); R.I. Gen. Laws § 42-80.1-3 (1998); 1999 S.C. Acts 38, to be codified at S.C. Code Ann. § 1-32-40 (Law. Co-op. 1999); Tex. Civ. Prac. & Rem. Code § 110.001 (1999); see also Ala. Const. amend. No. 622 (approved by referendum Nov., 1998). But see A.B. 1617, 1997-98 Reg. Sess. (Cal. 1997) (vetoed Sept. 27, 1998); H.R. 865, 144th Gen. Assembly, 1997-98 Reg. Sess. (Ga. 1997); H.B. 2916, 77th Leg., 1998 Reg. Sess. (Kan. 1998); H.B. 167A, 1998 Spec. Sess. A (La. 1998); S.B. 515, H.B. 1041, 1998 Reg. Sess. (Md. 1998); S.B. 678, H.B. 4376, 89th Leg., 1997 Reg. Sess. (Mich. 1997); H.B. 1470, 155th Sess. of the General Court (N.H. 1997) (defeated Feb. 18, 1998); S.B. 321, A.B. 903, 208th Leg. (N.J. 1998); S.B. 5673, A.B. 8499, 220th Annual Leg. Sess. (N.Y. 1997); H.B. 3469, 70th Leg. Assembly, 1999 Reg. Sess. (Or. 1999); H.B. 2068, 181st Gen. Assembly, 1997-98 Reg. Sess. (Pa. 1997); S.B. 3054, H.B. 3051, 100th Gen. Assembly, 1998 Reg. Sess. (Tenn. 1998); H.B. 550, 1997-98 Leg. Sess. (Vt. 1997-98); H.B. 1, 1998 Leg. Sess. (Va. 1998). The language of the state statutes generally closely tracks the language of the federal RFRA, 42 U.S.C. §§ 2000bb to 2000bb-4 (1994). See, e.g., 1998 Fla. Sess. Law Serv. ch. 98-412 (West) ("The government shall not substantially burden a person's exercise of religion [unless] application of the burden to the person . . . [i]s in furtherance of a compelling governmental interest; and . . . [i]s the least restrictive means of furthering that compelling governmental interest."). The federal RFRA was written to bind state governments as well as the federal government, but after City of Boerne v. Flores, 521 U.S. 507 (1997), it remains in effect only as to the federal government. See Christians v. Crystal Evangelical Free Church (In re Young), 141 F.3d 854 (8th Cir.), cert. denied, 119 S. Ct. 43 (1998). 

3 This division of religious freedom questions into primary questions, intermediate questions, and specific questions owes much to my colleague Dan Lowenstein's similar analysis of political theory in Political Bribery and the Intermediate Theory of Politics, 32 UCLA L. Rev. 784 (1985). The division is simply meant to be helpful in thinking about the subject, and not dispositive of any doctrinal questions; this spares me the hard, unnecessary, and perhaps impossible task of specifically defining the lines between primary, intermediate, and specific questions. 

4 Religious Freedom Restoration Act of 1993, 42 U.S.C. § 2000bb-1(b)(2). 

5 494 U.S. 872 (1990). 

6 See Eugene Volokh, A Common-Law Model for Religious Exemptions, 46 UCLA L. Rev. 1465, 1474-76, 1487-90 (1999). 

7 Grayned v. City of Rockford, 408 U.S. 104, 108 (1972). 

8 Before a software company distributes a program to the users, it's expected to test it as much as possible--to give it various kinds of inputs and make sure it reaches the right results. Of course, one can never test the program in every possible situation, but better to test it in a few rather than not at all. 

      This is why programmers design test suites. A test suite for a calculator program might, for instance, have the following test cases (among many others): 

      (1) Check to make sure 2+2 is 4. 

      (2) Check to make sure 3-1 is 2. 

      (3) Check to make sure 1-3 is -2 (because the program might work differently with positive numbers than negative). 

      (4) Check to make sure 1/0 yields an error message. 

      If the program doesn't yield the right results with these cases, then the author needs to change its logic. Test suites are critical to delivering quality code, and programmers often spend a huge amount of effort on them. The main software product that I wrote contains 100,000 lines of source code and 30,000 lines of test suites. Cf. KDM v. Reedsport Sch. Dist., 1999 U.S. App. LEXIS 29800 (9th Cir. Nov. 15, 1999) (Kleinfeld, J., dissenting): 

      One traditional test of the correctness of a legal proposition is to apply it to a case where we can be confident of the right answer, and see whether the proposition yields the wrong answer. We do the same thing to test a spreadsheet, by entering data where we know what the answer should be to see if the spreadsheet properly generates it. 


9 374 U.S. 398 (1963). 

10 I stress that I'm not claiming that test suites for legal doctrines are a remarkably novel idea: Most legal reasoning implicitly uses test suites whenever it asks what the consequences of a particular proposal might be. My modest contribution here is simply to suggest that test suites be explicitly prepared and published before proposed solutions are developed, and that they be created by someone other than the designer of such solutions. 

11 "We must think things not words, or at least we must constantly translate our words into the facts for which they stand, if we are to keep to the real and the true." Oliver Wendell Holmes, Jr., Law in Science and Science in Law, 12 Harv. L. Rev. 443, 460 (1899). 

12 For samples of the "no preference for religion allowed" school, see City of Boerne v. Flores, 521 U.S. 507, 536 (1997) (Stevens, J., concurring); Christopher L. Eisgruber & Lawrence G. Sager, Why the Religious Freedom Restoration Act Is Unconstitutional, 69 N.Y.U. L. Rev. 437, 453-54 (1994); William P. Marshall, The Case Against the Constitutionally Compelled Free Exercise Exemption, 40 Case W. Res. L. Rev. 357, 388-94 (1989-90); and Ellis West, The Case Against a Right to Religion-Based Exemptions, 4 Notre Dame J.L. Ethics & Pub. Pol'y 591, 600-04 (1990). 

      For samples of the "preference for religion in exemptions from government compulsion is no problem" school, see Corporation of the Presiding Bishop v. Amos, 483 U.S 327; and Texas Monthly, Inc. v. Bullock, 489 U.S. 1, 39 (1989) (plurality opinion) (Scalia, J., dissenting). 

13 See, e.g., Thomas C. Berg, What Hath Congress Wrought? An Interpretive Guide to the Religious Freedom Restoration Act, 39 Vill. L. Rev. 1, 48 (1994); cf. Wygant v. Jackson Bd. of Educ., 476 U.S. 267 (1986) (plurality opinion) (holding that even where race classifications in government employment are permitted, they generally may not involve discrimination in firing decisions, even if they may involve discrimination in hiring). 

14 See, e.g., Lee v. Weisman, 505 U.S. 577, 593 (1992); see also Michael W. McConnell, Free Exercise Revisionism and the Smith Decision, 57 U. Chi. L. Rev. 1109, 1146-47 (1990): 

      [Under] the Free Exercise Clause . . . the government is not required to create exemptions that would make religious believers better off relative to others than they would be in the absence of the government program to which they object. The purpose of free exercise exemptions is to ensure that incentives to practice a religion are not adversely affected by government action. By the same token, government action should not have the effect of creating incentives to practice religion. 


15 See Berg, supra note 13, at 48: 

      [T]he crucial factor is whether the exemption . . . encourag[es] persons to switch to (or feign) religious identification in order to gain it. . . . [M]ost religious conduct is not in this category. It is the sort of behavior in which one would not engage without the distinctive religious motivation. But exemptions from military service obviously present such a benefit, as do many forms of tax exemption. 


16 Id. at 50. 

17 Note that under the Court's Sherbert/Yoder-era religious exemption doctrine, the draft exemption is not mandated under strict scrutiny. See Gillette v. United States, 401 U.S. 437, 461-62 (1971). 

18 See Volokh, supra note 6, at 1528-29; see also Brief of 36 Religious Organizations, Leaders, and Scholars as Amici Curiae in Support of Respondents, Vacco v. Quill, 521 U.S. 793 (1997) (No. 95-1858), and Washington v. Glucksberg, 521 U.S. 702 (No. 96-110) (concluding that a right to assisted suicide is covered by the state constitutional right to privacy, and stating that a doctor had a right to "determine his own ethical, religious, and moral beliefs in declining or agreeing to assist"); McIver v. Krischer, No. CL-96-1504-AF, 1997 WL 225878, at *11 (Fla. Cir. Ct. Jan. 31, 1997) (adopting this argument), rev'd, 697 So. 2d 97 (Fla. 1997); Winthrop Drake Thies, Shall the Dying Be Denied Their Religious Freedom?, Newark Star-Ledger, Feb. 6, 1997, at 26: 

      [W]here the dying patient, in accordance with his religious beliefs, concludes that a loving and compassionate God has by the terminal process called him; and where the patient's physician believes that God asks him to play the part of the good Samaritan in releasing the dying person from suffering through death, we may not in effect say, "Your religion is different from mine and thus not worthy of respect. You can't die the way you religiously feel called upon to die. And you, Doctor, can't respond to your patient the way your religion calls on you to respond, although your suffering patient cries out for your help." The First Amendment guarantees to each of us--including each dying person and his physician--the free exercise of our religions. 


19 See Berg, supra note 13, at 50 ("The issue is whether there is a strong likelihood of inducement of religious practices over other practices."). 

20 Compare In re Grand Jury Empaneling of the Special Grand Jury, 171 F.3d 826 (3d Cir. 1999) (holding that RFRA doesn't allow a religiously motivated refusal to testify against a family member, at least in this case), and In re Doe, 842 F.2d 244, 245-48 (10th Cir. 1988) (same, under Free Exercise Clause), with In re Greenberg, 11 Fed. R. Evid. Serv. 579 (D. Conn. 1982) (holding the opposite, under Free Exercise Clause), and In re Grand Jury Empaneling, 171 F.3d at 837 (McKee, J., dissenting) (same, under RFRA). Cf.Grossberg's Parents Ask to Keep Talks Confidential, Newark Star-Ledger, Nov. 26, 1997, at 43: 

      The parents of Amy Grossberg, the college student accused of killing her newborn in Delaware . . . argued in court papers that talks with their daughter should be kept secret and that it is a violation of their right to the free exercise of religion [for prosecutors] to force them to divulge information. Rabbi Joel Roth, a legal expert at the Jewish Theological Seminary in New York City, confirmed yesterday he wrote an affidavit for the Grossbergs, stating that "under Jewish law, a mother and/or a father are not allowed to give testimony against their child in any legal proceeding." 

      Id.; see also Benjamin Weiser, Theological Discussion on Testifying Emerges in Terrorism Case, N.Y. Times, Aug. 8, 1999, at 33 (discussing a witness who refused to testify on the grounds that in his view the Koran taught Muslims to do nothing that "will cause harm to innocents," and that testifying in this case would be seen by his community as "a major sin" and "a betrayal of our beliefs as Muslims"). 

21 I've heard some people argue that the only possible opposition to homosexuality is religious, but this seems to me factually false; the officially and bitterly atheist Soviet Union, for instance, criminalized homosexuality. See, e.g., Criminal Code of the RSFR art. 121 (1965). Obviously this doesn't show the morality of bans on homosexuality, but it does show that even some strongly antireligious people have taken the view that homosexuality is immoral. See John M. Finnis, Law, Morality, and Sexual Orientation, 69 Notre Dame L. Rev. 1049 (1994) (citing Immanuel Kant, The Metaphysics of Morals 277-79, 424-26 (1797) (Mary Gregor trans., Cambridge Univ. Press 1991) (arguing, on essentially secular grounds, that homosexual conduct is wrong)). 

22 Though peyote is reportedly not pleasant to consume, there certainly have been non-American Indian religious groups that have concluded that consuming it produces spiritual benefits: see, for example, Peyote Way Church of God, Inc. v. Smith, 742 F.2d 193 (5th Cir. 1984). It's not hard to imagine nonreligious people who might take the same view. Consider, in a related context, Dr. Timothy Leary, who believed that psychedelic drugs help foster personal growth. See Leary v. United States, 383 F.2d 851, 857 (5th Cir. 1967). 

23 398 U.S. 333, 356-61 (1970) (Harlan, J., concurring in the judgment). 

24 See Ira C. Lupu, The Case Against Legislative Codification of Religious Liberty, 21 Cardozo L. Rev. 565 (1999). 

25 398 U.S. at 333 (reading statutory religious objector exception to include atheists); cf. id. at 356-61 (Harlan, J., concurring in the judgment) (arguing that the majority's reading is unsound, but concluding that the Establishment Clause requires equal treatment of atheist conscientious objectors and religious ones).

26 See Gillette v. United States, 401 U.S. 437, 461-62 (1971). 

27 Berg, supra note 13, at 48; see also Thomas C. Berg, The Constitutional Future of Religious Freedom Legislation, 20 U. Ark. Little Rock L.J. 715, 737 (1998) ("If RFRA in a given situation imposes serious or direct costs on other persons, disproportionate to the burdens removed from religious freedom, that would be a reason to strike down that particular application under the Establishment Clause."). 

28 472 U.S. 703 (1985). 

29 489 U.S. 1 (1989) (plurality opinion).

30 See Estate of Thornton, 472 U.S. at 709-10 (striking down a religious accommodation on the grounds that it provides "no exception when honoring the dictates of [religious objectors] would cause the employer substantial economic burdens or when the employer's compliance would require the imposition of significant burdens on other employees required to work in place of the [objectors]"); see also Texas Monthly, 489 U.S. at 15: 

      When government directs a subsidy exclusively to religious organizations that is not required by the Free Exercise Clause and that either burdens nonbeneficiaries markedly or cannot reasonably be seen as removing a significant state-imposed deterrent to the free exercise of religion, . . . it provide[s] unjustifiable awards of assistance to religious organizations. 

      Id. (emphasis added). 

31 I agree with Tom Berg that some cases, such as those involving soldiers' religious headgear, may not involve "direct or significant burdens on others," but I don't agree with the claim that a "large majority" of religious exemption claims involves no such burdens. Thomas C. Berg, The New Attacks on Religious Freedom Legislation, and Why They Are Wrong, 21 Cardozo L. Rev. 415 (1999). 

32 See Volokh, supra note 6, at 1514 n.54. 

33 Cf. Renzi v. Connelly Sch. of the Holy Child, 61 F. Supp. 2d 440 (D. Md. 1999) (using the Establishment Clause to strike down an explicit statutory exemption of this sort, in part on the grounds that such an exemption may harm private schools that compete with the religious schools). 

34 See Berg, supra note 13, at 48. This conclusion is not unanimous, see, for example, Michael Farris, Facing Facts: Only a Constitutional Amendment Can Guarantee Religious Freedom for All, 21 Cardozo L. Rev. 689 (1999), but the existence of the dispute only highlights the need for a test that will help courts resolve such disputes. 

35 See infra Part II.A. 

36 See Religious Freedom Restoration Act of 1993, 42 U.S.C. § 2000bb-1(b)(1) (1994). 

37 One of the as yet unresolved primary questions about state RFRAs is whether they guarantee exemptions only for religiously compelled conduct or also for religiously motivated conduct. See Mack v. O'Leary, 80 F.3d 1175, 1178 (7th Cir. 1996) (Posner, J.); Douglas Laycock & Oliver S. Thomas, Interpreting the Religious Freedom Restoration Act, 73 Tex. L. Rev. 209, 231-34 (1994); Douglas Laycock, The Remnants of Free Exercise, 1990 Sup. Ct. Rev. 1, 24-27; Ira C. Lupu, Why the Congress Was Wrong and the Court Was Right--Reflections on City of Boerne v. Flores, 39 Wm. & Mary L. Rev. 793, 806 (1998); Steven C. Seeger, Note, Restoring Rights to Rites: The Religious Motivation Test and the Religious Freedom Restoration Act, 95 Mich. L. Rev. 1472 (1997). 

38 487 U.S. 474 (1988). 

39 See Heffron v. ISKCON, 452 U.S. 640 (1981). 

40 See Jody Veenker, God Speaks to Commuters; Religious Messages on Billboards, Christianity Today, July 12, 1999, at 10. 

41 See Members of the City Council v. Taxpayers for Vincent, 466 U.S. 789 (1984) (upholding a ban on such flyers under less than strict scrutiny). 

42 See infra note 90. 

43 Cf. Red Lion Broadcasting v. FCC, 395 U.S. 367 (1969) (upholding the Fairness Doctrine under less than strict scrutiny in a case involving a religious broadcaster, though talking about the First Amendment generally without specifically discussing the Free Exercise Clause in particular); Scott v. Rosenberg, 702 F.2d 1263, 1272 (9th Cir. 1983) ("Courts have approved the application of FCC rules to religious groups on the same basis as applied to secular groups."); King's Garden, Inc. v. FCC, 498 F.2d 51, 59 (D.C. Cir. 1974) ("Unlike a religious newspaper, a sectarian radio or television station must, as King's Garden readily concedes, adhere to the 'fairness' and 'personal attack' doctrines and produce some programs of general community interest."). It's easy to imagine, for instance, a religious broadcaster who sincerely thinks it sinful to allow a blasphemous or evil response to be aired on his station. 

44 See Geoffrey R. Stone, Constitutionally Compelled Exemptions and the Free Exercise Clause, 27 Wm. & Mary L. Rev. 985, 994-96 (1986) (discussing "the special embarrassment that exists when free speech and free exercise claims coalesce"). 

45 See Heffron v. ISKCON, 452 U.S. 640 (1981); see also cases cited supra note 43. The Senate Judiciary Committee's report on the federal RFRA states that "where religious exercise involves speech, as in the case of distributing religious literature, reasonable time, place, and manner restrictions are permissible consistent with first amendment jurisprudence," S. Rep. No. 111, at 9 (1993), reprinted in 1993 U.S.C.C.A.N. 1892, 1898, but I don't think that this statement can undo the explicit textual command to apply strict scrutiny, and in any event, it's at most tenuously relevant to state RFRAs. 

46 See, e.g., Boos v. Barry, 485 U.S. 312, 321 (1988). 

47 461 U.S. 540, 550-51 (1983). 

48 512 U.S. 753, 763 (1994). 

49 512 U.S. 622, 645 (1994). 

50 460 U.S. 575, 591 (1983). 

51 119 S. Ct. 1923 (1999). 

52 Id. at 1935. 

53 435 U.S. 765 (1978). 

54 Id. at 784-85. 

55 See Alan E. Brownstein, State RFRA Statutes and Freedom of Speech, 32 U.C. Davis L. Rev. 605 (1999); see also Haff v. Cooke, 923 F. Supp. 1104, 1114-15 (E.D. Wis. 1996) (holding that RFRA may not be read to create a religious exemption from a prison policy banning possession of racist material, because such an exemption would violate the Free Speech Clause). Compare Daniel A. Farber, The First Amendment 115 (1998) (arguing, in a different context, that "a distinction based on motivation [of speech] differs from one based on content"), with Alan E. Brownstein, Alternative Maps for Navigating the First Amendment Maze, 16 Const. Comm. 101, 140-41 (1999) (disagreeing with Farber on this point). 

56 Cf. Brownstein, State RFRA Statutes and Freedom of Speech, supra note 55, at 643 n.86. 

57 See, e.g., Wooley v. Maynard, 430 U.S. 705 (1977); West Virginia Bd. of Educ. v. Barnette, 319 U.S. 624 (1943). 

58 This issue won't arise in some zoning exemption cases, such as cases involving churches that want to run soup kitchens in contravention of zoning ordinances. 

59 See Prince v. Massachusetts, 321 U.S. 158 (1944); cf. Farris, supra note 34, at 698 (arguing that Prince was mistaken, and that religious exemptions for such activity should be granted). A related question is whether the government must or may exempt sales of religious literature from generally applicable sales taxes. Cf. Carl H. Esbeck, The Establishment Clause as a Structural Restraint on Government Power, 84 Iowa L. Rev. 1, 80 (1998) (arguing that such an exemption is constitutionally compelled by the Establishment Clause itself, and disagreeing with the Court's recent precedent to the contrary). 

60 See generally Eugene Volokh, Freedom of Speech in Cyberspace from the Listener's Perspective, 1996 U. Chi. Legal Forum 377, 390-97 (discussing First Amendment significance of private entities' ability to choose whom to allow to participate in a discussion; the education context may differ in significant respects, but the decisions to include or exclude students still affect the content of the speech in schools). 

61 See In re Young, 141 F.3d 854 (8th Cir.), cert. denied, 119 S. Ct. 43 (1998). Congress has recently provided a special statutory exemption for charitable contributions generally, see Religious Liberty and Charitable Donation Protection Act of 1998, Pub. L. No. 105-183, 112 Stat. 517 (1998), but it's still interesting to ask whether in the absence of such a law a RFRA exemption for religious contributions alone would have been constitutional. 

62 406 U.S. 205, 211-12 (1972). 

63 See supra note 20. 

64 A purely Free Speech Clause claim for a right not to testify would fail, presumably because strict scrutiny is inapplicable to this sort of speech compulsion; but of course the purely Free Speech Clause claims in all the cases I describe in the text would fail, since in all those cases the Free Speech Clause does not require strict scrutiny. The problem is that RFRAs do require strict scrutiny, which raises the prospect of one standard being applied to secularly motivated speech (or silence) and another to religiously motivated speech. 

      If the compelled speech doctrine does require strict scrutiny here, but the subpoena power is still upheld under strict scrutiny, then presumably the subpoena power would likewise be upheld under strict scrutiny as against a RFRA challenge. 

65 Or, in the case of compulsory education laws, parents' religiously motivated control over the speech to which their children are exposed. 

66 See Employment Div. v. Smith, 494 U.S. 872, 901 (1990) (O'Connor, J., concurring) (arguing that "there is nothing talismanic about neutral laws of general applicability"). 

67 See Texas v. Johnson, 491 U.S. 397, 404 (1989); Spence v. Washington, 418 U.S. 405, 410-11 (1974). 

68 See Hurley v. Irish-American Gay, Lesbian & Bisexual Group, 515 U.S. 557, 569-70 (1995). 

69 See Heffron v. ISKCON, 452 U.S. 640 (1981). 

70 See Frisby v. Schultz, 487 U.S. 474 (1988). 

71 See Wisconsin v. Yoder, 406 U.S. 205 (1972). 

72 See Christians v. Crystal Evangelical Free Church (In re Young), 141 F.3d 854 (8th Cir.), cert. denied, 119 S. Ct. 43 (1998). 

73 See Prince v. Massachusetts, 321 U.S. 158, 160-61 (1944). 

74 See, e.g., Grosz v. City of Miami Beach, 721 F.2d 729, 738 (11th Cir. 1983), vacated and remanded, 82 F.3d 1005 (11th Cir. 1996); Laurie Reynolds, Zoning the Church: The Police Power Versus the First Amendment, 64 B.U. L. Rev. 767 (1985). 

75 See infra note 91. For this reason and for various others, I have argued that strict scrutiny is the wrong standard for RFRAs to use, see Volokh, supra note 6, at 1494-505, but that's the standard that the legislatures have selected. 

76 Berg, supra note 13, at 15. 

77 Michael W. McConnell & Richard A. Posner, An Economic Approach to Issues of Religious Freedom, 56 U. Chi. L. Rev. 1, 46 (1989). 

78 Id.

79 Letter from James Madison to Edward Livingston (July 10, 1822), in 9 The Writings of James Madison 100 (Gaillard Hunt ed., 1901). 

80 McConnell, supra note 14, at 1128. 

81 Berg, supra note 13, at 15. 

82 See, e.g., Jens Ludwig, Concealed-Gun-Carrying Laws and Violent Crime: Evidence from State Panel Data, 18 Int'l Rev. L. & Econ. 239 (1998) ("The net effects of shall-issue laws [which allow most law-abiding adults to get a license to carry a concealed gun] are as difficult to predict as those of widespread gun ownership, though shall-issue laws have an even greater potential for positive and negative externalities."). 

83 See generally John Lott, More Guns, Less Crime (1998) (arguing that gun carrying by law-abiding adults actually decreases violent crime). It's not clear whether this finding would carry over to knife carrying by Sikhs, but one component of the gun findings--that lawful licenseholders almost never commit crimes with their guns, see id. at 11--seems intuitively relevant to knives, too. 

84 Volokh, supra note 6, at 1515-29. 

85 Berg, supra note 13, at 15. 

86 These torts have generally been abrogated by statute. See, e.g., Cal. Civ. Code § 43.5 (1999). 

87 See, e.g., Berg, supra note 13, at 15. 

88 See Swanner v. Anchorage Equal Rights Comm'n, 874 P.2d 274, 284 (Alaska 1994) (discussing the two kinds of potential harms that are said to flow from housing discrimination). 

89 See Lyng v. Northwest Indian Cemetery Protective Ass'n, 485 U.S. 439 (1988) (involving an Indian sacred site on government property); see also Man Who Claimed to Talk with Virgin Mary Clams Up, Orlando Sentinel, Oct. 3, 1993, at A22 (stating that the owner of the property on which supposed visitations of the Virgin Mary had occurred, and which had attracted "throngs of faithful," "has posted 'No Trespassing' signs"). 

90 See, e.g., Worldwide Church of God v. Philadelphia Church of God, No. CV-97-5306-JSL (S.D. Cal. 1999), appeal pending, No. 99-55850 (9th Cir.); Urantia Found. v. Maaherra, 895 F. Supp. 1335 (D. Ariz. 1995); Urantia Found. v. Maaherra, 895 F. Supp. 1329 (D. Ariz. 1995); Bridge Publications, Inc. v. Vien, 827 F. Supp. 629 (S.D. Cal. 1993). 

91 For claims of a religious right to discriminate based on marital status, see, for example, Swanner, 513 U.S. at 981-82 (Thomas, J., dissenting from denial of certiorari) (suggesting that the claim should have been accepted); Thomas v. Anchorage Equal Rights Comm'n, 165 F.3d 692 (9th Cir. 1999) (accepting the claim, under a Free Exercise Clause "hybrid rights" rationale); Swanner v. Anchorage Equal Rights Comm'n, 874 P.2d 274, 284 (Alaska 1994) (rejecting such a claim); Smith v. Fair Employment & Hous. Comm'n, 51 Cal. Rptr. 2d 700 (Cal. 1996) (three justices concluded that the law didn't substantially burden religious exercise, three justices concluded that it did burden religious exercise and that it failed strict scrutiny, and one justice did not reach the question); Jasniowski v. Rushing, 685 N.E.2d 622 (Ill. 1997) (reversing with no opinion, a lower court decision, 678 N.E.2d 743 (Ill. App. Ct. 1997), which rejected such a claim); Attorney Gen. v. Desilets, 636 N.E.2d 233, 242-43 (Mass. 1994) (concluding that the result depends on factual findings); McCready v. Hofius, 586 N.W.2d 723 (Mich. 1998) (rejecting such a claim), vacated and remanded, 593 N.W.2d 545 (Mich. 1999) (appearing to reverse course, with little explanation); State v. French, 460 N.W.2d 2, 9-11 (Minn. 1990) (accepting such a claim). 

      For claims of a religious right to discriminate based on race (though in employment rather than housing), see EEOC v. Fremont Christian Sch., 781 F.2d 1362, 1368 (9th Cir. 1986); McLeod v. Providence Christian Sch., 408 N.W.2d 146, 151 (Mich. Ct. App. 1987). The classic example of a claimed religious right to discriminate against interracial couples is Bob Jones University v. United States, 461 U.S. 574 (1983). 

92 See In re Young, 82 F.3d 1407 (8th Cir. 1996), vacated, Christians v. Crystal Evangelical Free Church (In re Young), 521 U.S. 1114 (1997). 

93 See Cheema v. Thompson, 67 F.3d 883, 886 (9th Cir. 1995) (discussing a claimed right to wear a blunt dagger sown in its sheath); Nijjar v. Canada 3000 Airlines Ltd., Canadian Human Rights Trib. (July 9, 1999) (discussing a claimed right to wear a fairly long, sharp dagger). 

94 See United States v. Bauer, 84 F.3d 1549, 1559 (9th Cir. 1996); State v. Balzer, 954 P.2d 931, 941 (Wash. Ct. App. 1998); cf. Farris, supra note 34, at 697 (generally arguing for a very pro-exemption standard, but seemingly agreeing that exemptions from marijuana laws ought not to be required). 

95 McConnell, supra note 14, at 1145; see also McConnell & Posner, supra note 77, at 47-48; Akhil Reed Amar, The Bill of Rights: Creation and Reconstruction 43, 255-56 (1998) (suggesting that "some religious practices that affect only the religious community itself (with no externalities imposed on religious nonbelievers)" should be constitutionally immunized from regulation); id. at 386 n.100 (tentatively suggesting that even voluntary human sacrifice might be shielded by the Free Exercise Clause); Akhil Reed Amar, Intratextualism, 112 Harv. L. Rev. 747, 820 & n.285 (1999) (suggesting that the Fourteenth Amendment's incorporation of the Free Exercise Clause should be read as protecting practices "[w]here only co-religionists are involved" and not ones that "impose[] a harm on nonbelievers," with "harm" used "in the sense elaborated by John Stuart Mill and his followers"). 

96 See Farris, supra note 34 at 696, 698 (making the first two of these arguments); McConnell, supra note 14, at 1145 (making all three). 

97 See, e.g., State ex rel. Swann v. Pack, 527 S.W.2d 99 (Tenn. 1975) (involving a claimed religious freedom right to handle poisonous snakes and drink strychnine). 

98 See Buhl v. Hannigan, 16 Cal. App. 4th 1612, 1618 n.3 (1993) (describing such an exemption request, but rejecting it under Employment Division v. Smith); Dhillon v. Ministry of Transp., British Columbia Human Rights Trib., (May 11, 1999) < (accepting such an exemption request); Sikh Women Attack New Delhi's Helmet-Must Rule, Agence France Press, Nov. 4, 1997 (describing Sikh women's argument that they may not wear helmets because "wearing a 'topi' (cap) is considered to be against the basic [tenets] of our religion"). 

99 See Washington v. Glucksberg, 521 U.S. 702, 736 (1997); see also Nelson Lund, Two Precipices, One Chasm: The Economics of Physician-Assisted Suicide and Euthanasia, 24 Hastings Const. L.Q. 903, 914-16 (1997); Martha Minow, Which Question? Which Lie? Reflections on the Physician-Assisted Suicide Cases, 1997 Sup. Ct. Rev. 1, 21-23. 

100 Courts have at times engaged in such inquiries in cases where a donor or a donor's heirs ask to set aside a transfer to a religious institution or leader on the grounds that the transfer was a result of "undue influence." I'm troubled by undue influence inquiries even in religious donation cases, because I suspect that judgments about which forms of religious influence are "undue" will often (though not always) lead to improper consideration of whether the religion seems unreasonable, excessively authoritarian, or too threatening of extratemporal consequences. See, e.g., Carpenter v. Horace Mann Life Ins. Co., 730 S.W.2d 502 (Ark. App. 1987). But in any event, even if it's legitimate for courts to engage in this highly subjective and imprecise inquiry when deciding about money, I'm not sure that it's proper for them to do this when life is at stake. See Susan R. Martin & Henry J. Bourguignon, Physician-Assisted Suicide: The Lethal Flaws of the Ninth and Second Circuit Decisions, 85 Cal. L. Rev. 371, 396 (1997) (arguing, in my view persuasively, that courts won't be able to reliably determine whether undue influence was present in an assisted suicide decision). Oregon does try to implement such a solution under its pioneering assisted suicide legislation, see Or. Rev. Stat. ch. 127.890, sec. 4.02(2) (1987) ("A person who coerces or exerts undue influence on a patient to request medication for the purpose of ending the patient's life . . . shall be guilty of a Class A felony."), but I know of no evidence on how this provision has been working. 

101 McConnell, supra note 14, at 1145. Michael Farris likewise argues in favor of religious exemptions from minimum wage laws, and likewise ignores the unfair competition concern. See Farris, supra note 34, at 698. His argument is that the current minimum wage regime already excepts many job categories, and therefore wouldn't be undermined by yet one more exemption; but exempting a religious business from a law that covers competing businesses would tilt the playing field against the secular businesses in a way that wouldn't happen with exceptions of a whole job category, which generally would affect all competing businesses. 

102 Tony & Susan Alamo Found. v. Secretary of Labor, 471 U.S. 290, 302 (1985). 

103 McConnell & Posner, supra note 77, at 49. 

104 McConnell, supra note 14, at 1146; see also Paul v. Watchtower Bible and Tract Society, 819 F.2d 875 (9th Cir. 1987) (concluding that Free Exercise Clause required an exemption from a tort action for shunning because the plaintiff "[was] a former Church member" and "[c]hurches are afforded great latitude when they impose discipline on members or former members"). 

105 The objections to the policy in Bob Jones University v. United States, 461 U.S. 574 (1983) were brought by the IRS, not by the students, and most Bob Jones students probably agreed with the discriminatory rule. But if all of them agreed with the rule there would be little need for a rule; a nonbinding exhortation would have sufficed. The existence of the rule suggests that it would have at some time been enforced. 

106 See State ex rel. Swann v. Pack, 527 S.W.2d 99 (Tenn. 1975). 

107 See id.

108 See supra note 98. 

109 See United States v. Lee, 455 U.S. 252, 260 (1982). 

110 See, e.g., Fraternal Order of Police Newark Lodge No. 12 v. City of Newark, 170 F.3d 359 (3rd Cir. 1999) (accepting this argument); Rader v. Johnston, 924 F. Supp. 1540 (D. Neb. 1996) (same); Richard F. Duncan, Who Wants to Stop the Church: Homosexual Rights Legislation, Public Policy, and Religious Freedom, 69 Notre Dame L. Rev. 393, 427-28 (1994); Garrett Epps, What We Talk About When We Talk About Free Exercise, 30 Ariz. St. L.J. 563, 597 (1998); Laycock, supra note 37, at 50-51. 

111 See Michael Stokes Paulsen, A RFRA Runs Through It: Religious Freedom and the U.S. Code, 56 Mont. L. Rev. 249, 269-70 (1995) ("An interest not pursued with paramount rigor and consistency is not a paramount interest."). 

112 See Laycock, supra note 37, at 50 ("In general, the allowance of any exemption is substantial evidence that religious exemptions would not threaten the statutory scheme."). 

113 See Volokh, supra note 6, at 1539-42. 

114 See, e.g., Jaffee v. Redmond, 511 U.S. 1, 9-11 (1996) (comparing proposed psychotherapist-patient privilege with the spousal and attorney-client privileges); id. at 1934 (Scalia, J., dissenting) (comparing proposed privilege with a parent-child privilege, which has generally been rejected by the courts). 

115 Laycock, supra note 37, at 51. 

116 See Volokh, supra note 6, at 1540 (giving detailed citations). 

117 See Laycock, supra note 37, at 50-51. 

118 But see Amar, supra note 95, at 386 n.100 (tentatively suggesting that religious exemptions for voluntary human sacrifice might be proper and even constitutionally mandated). 

119 See Thomas C. Berg, Slouching Towards Secularism, 44 Emory L.J. 433, 467 (1995) (arguing that there should be a "'most favored nation' status for religious freedom: if anyone else gets an exemption, religious practice must also get an exemption" (quoting in part Laycock, supra note 37, at 49)). 

120 See supra note 89. 

121 See supra note 90. 

122 See supra note 20. 

123 See Volokh, supra note 6, at 1517 n.167 (collecting examples of such claims as to peyote and marijuana). 

124 I owe this hypothetical to Richard Duncan. 

125 See supra note 91. 

126 See Connick v. Myers, 461 U.S. 138 (1983); Pickering v. Board of Educ., 391 U.S. 563, 568 (1968). 

127 See, e.g., Rust v. Sullivan, 500 U.S. 173, 203 (1991) (concluding that the government could require contracting doctors who are paid to advise patients about non-abortion family planning to limit their advice to non-abortion procedures); Miles v. Denver Public Sch., 944 F.2d 773, 777 (10th Cir. 1991) (concluding that Connick/Pickering test didn't apply to schoolteacher's in-class speech to students). The leading judicial debate on this subject is in Yniguez v. Arizonans for Official English, 69 F.3d 920 (9th Cir. 1995) (en banc), rev'd on procedural grounds, 517 U.S. 1102 (1996), between Judge Kozinski's dissent and Judge Reinhardt's majority opinion and separate concurrence. 

128 See, e.g., Baz v. Walters, 782 F.2d 701, 708-09 (7th Cir. 1986); Philadelphia Lodge No. 5 v. City of Philadelphia, 599 F. Supp. 254, 258 (E.D. Pa. 1984) (seeming to apply an even more deferential standard than Pickering); Doherty v. Wilson, 356 F. Supp. 35, 41 (M.D. Ga. 1973) (adopting a Pickering-like approach); Barlow v. Blackburn, 798 P.2d 1360, 1366 (Ariz. Ct. App. 1990) (seeming at times to apply Pickering and at other times to apply strict scrutiny). 

129 Some decisions applying state constitutional religious freedom protections likewise assume that strict scrutiny applies even to the government acting as employer. See, e.g., Seabrook v. City of New York, No. CIV.9134, 1999 WL 694265, at *6 (S.D.N.Y. Sept. 7, 1999) (upholding under strict scrutiny a policy barring corrections officers from wearing skirts on duty, though without engaging in any serious inquiry into less restrictive standards). 

130 To see why this is so, consider a hypothetical law banning offensive picketing outside government buildings on the grounds that the picketing offends government employees and hurts morale in the government office. Such a law would, I assume, be unconstitutional, which means that avoiding offense and morale problems must not be compelling interests. 

      Some argue that there might be a compelling interest in preventing certain kinds of religious harangues--ones that focus on the recipient's religion or irreligion--because such harangues can become "religious harassment" and thus violate employment discrimination law. Cf. Meltebeke v. Bureau of Labor & Indus., 903 P.2d 351 (Or. 1995) (upholding a limited Free Exercise Clause defense to an accusation of harassment by repeated proselytizing); Eugene Volokh, Freedom of Speech and Workplace Harassment, 39 UCLA L. Rev. 1791 (1992) (arguing that even insulting religious speech should often be seen as protected by the First Amendment). Nonetheless, even this supposedly compelling interest would arise only when the harangues are "severe or pervasive" enough to create a "hostile or abusive work environment." Eugene Volokh, What Speech Does "Hostile Work Environment" Harassment Law Restrict?, 85 Geo. L.J. 627, 627 (1997). An individual or occasional berating of a coworker would not rise to this level, so the compelling interest wouldn't justify a ban on such occasional rude speech; but I take it that the government acting as employer should be allowed to restrict even such isolated rude speech to prevent the friction it causes, which suggests that the standards for the government acting as employer must be different from those for the government acting as sovereign. 

131 See Waters v. Churchill, 511 U.S. 661, 672-75 (1994) (plurality opinion). 

132 As to objections to police duties related to protecting activities that the police officer thinks are seriously sinful, see Ryan v. Department of Justice, 950 F.2d 458, 462 (7th Cir. 1991) (rejecting, under Title VII, a religious exemption claim by an FBI employee who refused to investigate destruction of government property by anti-war groups, and who refused to accept a duty swap, but not applying strict scrutiny); Rodriguez v. City of Chicago, 975 F. Supp. 1055, 1060 (N.D. Ill. 1997) (rejecting such an objection to guarding abortion clinics, without reaching a strict scrutiny analysis), aff'd, 156 F.3d 771 (7th Cir. 1998) (concluding that the city had adequately accommodated the officer by offering him a transfer to a district that didn't have an abortion clinic); Parrott v. District of Columbia, CIV.A.91-004, 1991 WL 126020 (D.D.C. June 25, 1991) (rejecting a police sergeant's request for exemption from duties which involved the arrest of peaceful anti-abortion trespassers), aff'd by unpublished opinion, 959 F.2d 1102, 1992 WL 75053 (D.C. Cir. Mar. 20, 1992); cf.Rodriguez, 156 F.3d at 778 (Posner, J., concurring), concluding that the city should not have any obligation to accommodate such an objection because of "[t]he importance of public confidence in the neutrality of its protectors." 

      As to objections to actually directly or indirectly helping in the performance of abortions, see Tramm v. Porter Mem'l Hosp., No. H 87-355, 1989 U.S. Dist. LEXIS 16391, at *33 (N.D. Ind. Dec. 22, 1989) (concluding that a hospital had to reasonably accommodate a nurse who objected to cleaning instruments that had been and would be used in abortions, applying Title VII and the Free Exercise Clause); Kenny v. Ambulatory Centre of Miami, 400 So. 2d 1262 (Fla. App. 1981) (concluding that a hospital had to reasonably accommodate a nurse who had objected to assisting with abortions, applying state law but borrowing by analogy from the Title VII undue hardship cases). 

      As to objections to administrative duties involving processing clerical materials that help in causes that an objector thinks are immoral, see American Postal Workers Union v. Postmaster Gen., 781 F.2d 772, 777 (9th Cir. 1986) (concluding that government employer had a duty to reasonably accommodate, by arranging transfers to other jobs, postal workers who had a religious objection to processing draft registration forms); McGinnis v. United States Postal Serv., 512 F. Supp. 517, 523 (N.D. Cal. 1980) (finding the government had a duty to reasonably accommodate, by offering a transfer to another window that wasn't used for registration materials); Haring v. Blumenthal, 471 F. Supp. 1172 (D.D.C. 1979) (concluding that the IRS had an obligation to exempt an employee from having to work on tax-exempt status applications from abortion clinics and other organizations that the employee thought it sinful to deal with). See also Gavin v. Peoples Natural Gas Co., 613 F.2d 482, 484 (3d Cir. 1980) (discussing, but not resolving, a Jehovah's Witness employee's objection to part of his job tasks, which involved having to raise and lower a flag); Palmer v. Board of Educ., 603 F.2d 1271, 1274 (7th Cir. 1979) (concluding, under Free Exercise Clause rather than Title VII, that a school district had no obligation to allow teacher "to refuse to participate in the Pledge of Allegiance, the singing of patriotic songs, and the celebration of certain national holidays," when teacher's job was in part "to teach patriotic matters to children"); Best v. California Apprenticeship Council, 207 Cal. Rptr. 863, 868 (Ct. App. 1984) (concluding that an apprentice training organization--which was treated by state law as an employer--had an obligation to accommodate an apprentice's religious objection to working in a nuclear power plant); David Haldane, Panel Backs Fired Vegetarian Bus Driver, L.A. Times, Aug. 24, 1996, at A18 (discussing a case in which the EEOC concluded that a transportation agency must accommodate a vegetarian bus driver's religious objections to handing out hamburger coupons as part of the agency's promotion aimed at boosting ridership); Felhaber et al., Bits and Pieces, Minn. Employment L. Letter, Sept. 1997 (reporting that the case against the transportation agency was settled for $50,000). 

133 Rust v. Sullivan, 500 U.S. 173, 196 (1991). 

134 See S. Rep. No. 111, at 8-9 (1993), reprinted in 1993 U.S.C.C.A.N. 1892, 1898 ("The committee expects that the courts will look to free exercise cases decided prior to Smith for guidance in determining whether the exercise of religion has been substantially burdened and the least restrictive means have been employed in furthering a compelling government interest."). One difficulty with the argument that RFRA was only meant to restore strict scrutiny for the government acting as sovereign is that Sherbert itself, which was very much considered in the RFRA debate, involved the government acting as welfare distributor, not as sovereign. See, e.g., 42 U.S.C. §§ 2000bb(b) (1994) ("The purposes of this chapter [include] . . . (1) to restore the compelling interest test as set forth in Sherbert v. Verner, 374 U.S. 398 (1963)."). 

      Sidney Rosenzweig suggests that the federal RFRA was specifically not intended to create any extra protection for government employees beyond what is provided by the Title VII undue hardship test (though he suggests that even given this standard, RFRA is relevant because it affects other laws that influence the undue hardship analysis). See Sidney A. Rosenzweig, Restoring Religious Freedom to the Workplace: Title VII, RFRA and Religious Accommodation, 144 U. Pa. L. Rev. 2513, 2526 (1996). As evidence, he points to the Senate Report's statement that "[n]othing in this act shall be construed as affecting religious accommodation under Title VII of the Civil Rights Act of 1964." S. Rep. No. 111, at 13 (1993), reprinted in 1993 U.S.C.C.A.N. 1892, 1903. 

      This statement, though, on its face says only that claims brought under Title VII won't be affected; it says nothing about extra claims brought under RFRA. To come to the conclusion that RFRA's facial strict scrutiny requirement doesn't apply to the government acting as employer, one has to read the report as saying that "nothing in this act shall be construed as creating new religious exemption claims for government employees beyond those brought under Title VII of the Civil Rights Act of 1964"--in my view, quite a stretch, especially given that the statutory text seems to dictate the opposite. (Nothing in the context surrounding the statement in the Senate Report sheds extra light on the subject.) 

      The statement in the report seems to be just an assurance that RFRA won't diminish protections offered under other laws (or perhaps won't broaden restrictions on private employers, which are covered by Title VII but not RFRA); the drafters of RFRA in fact seemed to have liked such assurances. See 42 U.S.C. §§ 2000bb-3(c) (1994) ("Religious belief unaffected. Nothing in this Act shall be construed to authorize any government to burden any religious belief."); 42 U.S.C. § 2000bb-4 (1994) ("Establishment Clause unaffected. Nothing in this Act shall be construed to affect, interpret, or in any way address [the Establishment Clause]. Granting government funding, benefits, or exemptions, to the extent permissible under the Establishment Clause, shall not constitute a violation of this Act. As used in this section, the term "granting," used with respect to government funding, benefits, or exemptions, does not include the denial of government funding, benefits, or exemptions."). 

      Finally, note that the statement in the Senate Report on the federal RFRA would in any event be of virtually no relevance to the interpretation of state RFRAs. 

135 See Employment Div. v. Smith, 494 U.S. 872, 887 (1990) ("It is not within the judicial ken to question the centrality of particular beliefs or practices to a faith, or the validity of particular litigants' interpretations of those creeds." (quoting Hernandez v. Commissioner, 490 U.S. 680, 699 (1989))). But see Bryant v. Gomez, 46 F.3d 948, 949 (9th Cir. 1995) (applying a centrality requirement, without discussing Smith's and Hernandez's rejection of such a requirement). 

136 See generally, e.g., Cynthia L. Estlund, Speech on Matters of Public Concern: The Perils of an Emerging First Amendment Category, 59 Geo. Wash. L. Rev. 1 (1990). Cf. Smith, 494 U.S. at 886-87 ("It is no more appropriate for judges to determine the 'centrality' of religious beliefs before applying a 'compelling interest' test in the free exercise field, than it would be for them to determine the 'importance' of ideas before applying the 'compelling interest' test in the free speech field."). 

137 TWA v. Hardison, 432 U.S. 63, 84 (1977). 

138 See Estate of Thornton v. Caldor, 472 U.S. 703, 709-10 (1985) (striking down a religious accommodation on the grounds that it provides no exception for cases when "honoring the dictates of [religious objectors] would cause the employer substantial economic burdens or when the employer's compliance would require the imposition of significant burdens on other employees required to work in place of the [objectors]"). 

139 See, e.g., Rosalie Berger Levinson, Silencing Government Employee Whistleblowers in the Name of "Efficiency," 23 Ohio N.U. L. Rev. 17, 42 n.121 (1996) ("[E]ven before Connick, the protection afforded free speech rights of government employees was uncertain under the Pickering balance."); William P. Marshall, supra note 12, at 411 (1989-90) ("The cases [under strict scrutiny in religious freedom cases] have been inconsistent . . . ."); Michael W. McConnell, Religious Freedom at a Crossroads, 59 U. Chi. L. Rev. 115, 170 (1992) ("The [strict scrutiny] test . . . provided little guidance to legislatures or lower courts . . . ."); Lawrence Rosenthal, Permissible Content Discrimination Under the First Amendment: The Strange Case of the Public Employee, 25 Hastings Const. L.Q. 529 (1998); West, supra note 12, at 608 (describing the strict scrutiny test in religious freedom cases as leading "to unpredictable results" and being "vague enough that courts can manipulate it to justify almost any decision they make"). 

140 Cf. Tiano v. Dillard Dep't Stores, Inc., 139 F.3d 679, 683 (9th Cir. 1998) (involving a request for accommodation of a two-week-long pilgrimage to a visitation of the Virgin Mary, but concluding that this particular claimant didn't show that she had a religious obligation to go on this pilgrimage at this particular time); id. at 683 (Fletcher, J., dissenting) (concluding that the claimant did have a religious obligation, and that the employer had a duty to accommodate it); EEOC v. Universal Mfg. Corp., 914 F.2d 71, 74 (5th Cir. 1990) (concluding that an employer may have a duty to accommodate an employee's religious obligation to not work during a seven-day holiday period). 

141 See supra note 127. 

142 42 U.S.C. §§ 2000e-2(a)(1), 2000e(j) (1994). This proviso only applies "unless an employer demonstrates that he is unable to reasonably accommodate to an employee's or prospective employee's religious observance or practice without undue hardship on the conduct of the employer's business." Id. § 2000e(j). 

143 Haldane, supra note 132, at A18. All these cases, except the one in item (e), are based on cases cited supra note 132. 

144 See, e.g., EEOC v. United Parcel Serv., 94 F.3d 314, 320-21 (7th Cir. 1996) (holding that accommodation of beard was required); Rourke v. New York State Dep't of Correctional Servs., 915 F. Supp. 525, 538 (N.D.N.Y. 1995) (holding that accommodation of long hair was required); EEOC v. Sambo's of Georgia, Inc., 530 F. Supp. 86, 93 (N.D. Ga. 1981) (holding that accommodation of beard was not required, because it would impose an undue hardship by offending customers, many of whom "prefer restaurants whose managers and employees are clean-shaven"). 

145 Cf. Jane Applegate, Religious Conflicts Affecting Work Place, Fresno Bee, Feb. 1, 1998, at C2 (describing Orthodox Jewish women's belief that they "couldn't shake [male business associates'] hands because Orthodox women are not permitted to touch men other than their husbands"); Catherine Welsh, Reflections on Platonic Friendship, America, Nov. 1, 1997 (describing Moslem women's similar beliefs); see also Miller v. Drennon, 966 F.2d 1443 (4th Cir. 1992) (considering a claim that a county violated an emergency worker's rights "by scheduling him to work 24-hour shifts in single bedroom substations with a female partner, contrary to his religious beliefs," and concluding that offering to allow shift swaps was a reasonable accommodation, but rendering no opinion as to the right result under strict scrutiny). 

146 See Wilson v. U.S. West Communications, 58 F.3d 1337, 1341 (8th Cir. 1995) (considering a claim of religious obligation to wear a two-inch button showing a color photograph of an 18- to 20-week-old fetus, with the phrases "Stop Abortion" and "They're Forgetting Someone," and rejecting the claim on the grounds that this particular employee's religious beliefs in fact did not require her to wear the button exposed); infra note 153 (discussing incidents involving potentially offensive religious symbols). 

147 See Banks v. Service Am. Corp., 952 F. Supp. 703, 711 (D. Kan. 1996) (concluding that such a claim could win because accommodating the claim isn't an undue hardship, despite the 20 to 25 complaints that the employer had gotten from patrons); Johnson v. Halls Merchandising, Inc., No. 87-1042, 1989 WL 23201, at *2 (W.D. Mo. Jan. 17, 1989) (concluding that the employer need not accommodate an employee's religious belief "which required her to preface nearly every sentence she spoke with the phrase 'In the name of Jesus Christ of Nazareth,'" because the employer had a "legitimate and reasonable interes[t] [in] operat[ing] a retail business so as not to offend the religious beliefs or non beliefs of its customers"). Johnson might have been resolved on the narrower grounds that prefixing every sentence with an eight-word phrase would simply make it too cumbersome to communicate, but the court's reasoning rested not on this but on the concern about offense. 

148 See Chalmers v. Tulon Co., 101 F.3d 1012 (4th Cir. 1996) (involving such a claim, but concluding that in this particular case the employee didn't have a religious obligation to offer such criticisms and, even if the employee did have such an obligation, the employer wouldn't have to accommodate an obligation to make criticisms that "impose personally and directly on fellow employees, invading their privacy and criticizing their personal lives"); cf. Venters v. City of Delphi, 123 F.3d 956, 977 (7th Cir. 1997) (describing free speech claim to make such criticisms, and concluding that the Free Speech Clause "did not grant [the claimant] license to make highly personal remarks about the status of [a coworker's] soul when informed that these remarks were unwelcome"); Chalmers, 101 F.3d at 1021-27 (Niemeyer, J., dissenting) (concluding that the employee did have the religious obligation to make such criticisms, and that under certain circumstances the employer would have to accommodate this obligation). 

149 Sometimes the government might specifically bar religious speech or garb by its employees on the grounds that such a ban is needed to "preserve[e] . . . an atmosphere of religious neutrality" or even to avoid a violation of the Establishment Clause. United States v. Board of Educ., 911 F.2d 882, 893 (3d Cir. 1990); see also id. at 898 (Ackerman, J., dissenting); Cooper v. Eugene Sch. Dist. No. 4J, 723 P.2d 298 (Or. 1986). This, though, involves a law targeting religion and not a neutral law that incidentally burdens religion; such a law must therefore face strict scrutiny under the Free Exercise Clause, without regard to what might happen under RFRA. See Board of Educ., 911 F.2d at 888 n.3. 

150 Frances Welch, A Woman's Right to Jewishness, Sunday Telegraph, Apr. 28, 1996, at 6. 

151 Customer or coworker preference for employees of a particular sex or religion is generally not seen as enough to justify facial discrimination under the "bona fide occupational qualification" exception. See Eugene Volokh, The California Civil Rights Initiative: An Interpretive Guide, 44 UCLA L. Rev. 1335, 1370-74 & n.113 (1997) (discussing the issue and collecting cases). One might therefore argue that customer or coworker antipathy to a religious practice likewise can't count towards a finding of undue hardship under the accommodation proviso. 

152 Cf. Will Haynie, Agreement Lets Student Wear Wiccan Symbol: Settlement Ends Lawsuit Brought by Lincoln Park Senior, Detroit News, Mar. 23, 1999, at D1 (discussing a high school policy--in this instance aimed at controlling gangs--banning the wearing of pentagrams). 

153 Cf. Martin Dyckman, Wine, Women, and Spirited Debate, St. Petersburg Times, Mar. 3, 1998, at 19A (describing a case where a person claimed a right to wear a swastika as an Aryan Nations religious symbol, albeit in the prisoner context, where the RFRA rules are generally seen as different); Abdon M. Pallasch, Hindu Files Suit to Challenge Swastika Firing, Chi. Tribune, Aug. 6, 1998, at 4 (describing a workplace incident involving a swastika as a Hindu religious symbol); Falun Dafa (visited Oct. 13, 1999) < (featuring two prominent swastikas as part of the website of the Chinese Falung Gong spiritual movement). Some of the Asian swastikas may be the mirror image of the Nazi swastika, but I suspect that many Western observers who saw them without seeing any context--for instance, on a pendant or a pin--would interpret them as Nazi swastikas. 

154 See James Brooke, Terror in Littleton, N.Y. Times, Apr. 21, 1999, at A1 (discussing the latter example). Presumably the latter button would be worn out of antireligious conviction rather than religious conviction, but a RFRA may well be interpreted to cover such sentiments. See, e.g., Tooley v. Martin-Marietta Corp., 476 F. Supp. 1027, 1030 (D. Or. 1979) (interpreting Title VII as requiring accommodation of atheist practices related to religiosity as well as religious practices); Volokh, supra note 6, at 1493 n.80 (citing cases that read Title VII to apply to deeply held conscientious nonreligious beliefs as well as religious beliefs). 

155 See supra note 144. 

156 Cf. Wilson v. U.S. West Communications, 58 F.3d 1337 (8th Cir. 1995) (involving a similar, but slightly different, claim). 

157 See supra note 145. 

158 See supra note 147. 

159 See supra note 152. 

160 See supra note 153. 

161 See supra note 153. 

162 See supra note 156. 

163 Shahar v. Bowers, 114 F.3d 1097, 1110-11 & n.27 (11th Cir. 1997) (applying Pickering rather than strict scrutiny, and holding that the government action passes muster); see also Potter v. Murray City, 760 F.2d 1065, 1069 (10th Cir. 1985) (applying strict scrutiny and holding that the government action passes muster). 

164 Lumpkin v. Brown, 109 F.3d 1498, 1501 (9th Cir. 1997) (upholding firing under RFRA, applying strict scrutiny). Lumpkin involved not just a high-ranking employee but a member of an appointed commission; an argument can be made that RFRA shouldn't be read as applying to such policymaking positions, though the argument is not an open-and-shut winner. 

165 This is the only pure hypothetical in this group, but it's drawn from the furor surrounding Gov. George W. Bush's statement--made in response to a reporter's question--that only Christians can get into Heaven. See Ken Herman, Bush Trip Steeped in History, Austin American-Statesman, Dec. 1, 1998, at A1. 

166 Rodriguez v. City of Chicago, 156 F.3d 771 (7th Cir. 1998) (Posner, J., concurring). The majority opinion concluded that the city had adequately accommodated the officer by offering him a transfer to a district that didn't have an abortion clinic. See id.; see also Rodriguez v. City of Chicago, 975 F. Supp. 1055 (N.D. Ill. 1997) (rejecting such a claim under the Title VII undue hardship test, without reaching a strict scrutiny analysis). 

167 Cf. Employment Div. v. Smith, 494 U.S. 872, 874 (1990) (involving such a situation, albeit where the rehabilitation program was privately run). 

168 McMullen v. Carson, 754 F.2d 936, 940 (11th Cir. 1985) (accepting this argument as to a Free Speech Clause challenge; the organization there was not a religious group but the KKK, so no religious freedom argument was available); see also Shahar,114 F.3d at 1108-09 (citing McMullen favorably). 

169 See ISKCON v. Lee, 505 U.S. 672 (1992). 

170 See id

171 485 U.S. 439 (1988). 

172 Bowen v. Roy, 476 U.S. 693 (1986); see also Native Am. Heritage Comm'n v. Board of Trustees, 59 Cal. Rptr. 2d 402 (Ct. App. 1996) (discussing a claim, brought under a special state statute, that an urban site owned by a government entity was sacred to an American Indian religious group and could not be developed). 

173 For whatever it's worth--and I think it's worth virtually nothing--here's what the Senate Judiciary Committee's report on the federal RFRA suggests on this point: "[W]hile the committee expresses neither approval nor disapproval of that case law, pre-Smith case law makes it clear that strict scrutiny does not apply to government actions involving only management of internal Government affairs or the use of the Government's own property or resources." S. Rep. No. 111, at 9 (1993), reprinted in 1993 U.S.C.C.A.N. 1892, 1898. This question is also complicated by the fact that some state RFRAs specifically call for a rather low burden threshold, likely lower than the substantial burden threshold used in the pre-Smith cases. See Ala. Const. amend. No. 622 (approved by referendum Nov., 1998) (covering any action that "burdens" religious exercise, and omitting the "substantial" qualifier); Ariz. Rev. Stat. § 41-1493.01 (1999) (stating that the "substantial" qualifier is meant to exclude only "de minimis," trivial, or technical burdens). 

174 Cf. Siegel v. Blair Hall, Inc., 615 N.Y.S.2d 937 (App. Div. 1994) (describing such a case in a private apartment building). 

175 Cf. George W. Dent, Jr., Of God and Caesar: The Free Exercise Rights of Public School Students, 43 Case W. Res. L. Rev. 707, 729-30 (1993) (suggesting that accommodation might be constitutionally required as to hospital food services for patients). 

176 Cf. Lambert v. Condor Mfg., Inc., 768 F. Supp. 600 (E.D. Mich. 1991). Lambert held that an employer had a duty to accommodate an employee's religious objection to working around "nude photographs of women" by ordering its employees to take down such photographs. Id. at 604. I think such a holding violates the First Amendment rights of the coworkers; I don't believe a court is entitled to order people to take down offensive but nonobscene pictures, or to pressure an employer into making such an order to its employees. Nonetheless, the court disagreed, see id., and in any event in the hypothetical described in the text the objects are put up by a local government agency, which may not have First Amendment rights of its own against state laws such as state RFRAs. See Roderick M. Hills, Jr., Back to the Future? How the Bill of Rights Might Be About Structure After All, 93 N.W. U. L. Rev. 977, 1004 & n.98 (1999) (discussing this issue and persuasively arguing that state and local agencies should have free speech rights vis-à-vis the federal government, but probably not vis-à-vis their own state governments). 

      Note that there's nothing preposterous in assuming that the government might put up nude works in public spaces; I walk every day to work through the UCLA Sculpture Garden--an open space between buildings on the campus--where there are several sculptures of nudes. Cf. Henderson v. City of Murfreesboro, 960 F. Supp. 1292 (M.D. Tenn. 1997). 

177 See, e.g., Nomi Maya Stolzenberg, "He Drew a Circle That Shut Me Out": Assimilation, Indoctrination, and the Paradox of a Liberal Education, 106 Harv. L. Rev. 581, 585, 591, 595 (1993) (discussing some such religious objections in Mozert v. Hawkins County Board of Education, 827 F.2d 1058 (6th Cir. 1987)); see also Dent, supra note 175, at 708-09 (discussing other such objections); George W. Dent, Jr., Religious Children, Secular Schools, 61 S. Cal. L. Rev. 863, 867-69 (1988) (same); Richard F. Duncan, Public Schools and the Inevitability of Religious Inequality, 1996 BYU L. Rev. 569, 578, 581 (same); Rosemary C. Salomone, Common Schools, Uncommon Values: Listening to the Voices of Dissent, 14 Yale L. & Pol'y Rev. 169, 181-82, 213 (1996) (same); David Bernstein, Why Johnny Can't Pray, Reason, Feb. 1992, at 56 (same). 

178 See Heffron v. ISKCON, 452 U.S. 640 (1981). 

179 See ISKCON v. Lee, 505 U.S. 672 (1992) (O'Connor, J., concurring). 

180 Cf., e.g., Paul Brest & Sanford Levinson, Processes of Constitutional Decisionmaking 1247-48 (1992). 

181 Cf. Singh v. Workmen's Compensation Bd. Hosp. and Rehabilitation Centre (Ontario Bd. of Inquiry 1981) (holding that Sikh patients should be allowed to wear kirpans "of a reasonable length" in a hospital despite a hospital policy barring the possession of offensive weapons). 

182 Cf. Quaring v. Peterson, 728 F.2d 1121, 1127 (8th Cir. 1984) (granting such an exemption in a government-as-sovereign context), aff'd by an equally divided court, 472 U.S. 478 (1985); Penn High to Force Students to Wear Identification Badges, South Bend Trib., Aug. 31, 1999, at A2. 

183 491 U.S. 781 (1989). 

184 See ISKCON v. Lee, 505 U.S. 672 (1992) (O'Connor, J., concurring in the judgment) (treating the reasonableness requirement as stronger than the rational basis requirement, and using it to strike down part of the speech regulation). 

185 See supra note 174. 

186 See supra note 175. 

187 See supra note 176. 

188 Cf. Heffron v. ISKCON, 452 U.S. 640, 654 (1981) (discussing such a claim, but holding for the government under intermediate scrutiny); ISKCON v. Barber, 650 F.2d 430, 444 (2d Cir. 1981) (holding for the objector under strict scrutiny pre-Heffron). 

189 See supra note 93. 

190 See supra note 182. 

191 Tom Berg has a fine article on this subject, see State Religious Freedom Statutes in Private and Public Education, 32 U.C. Davis L. Rev. 531 (1999), but there definitely remains ample room for more work to be done in this area. 

192 393 U.S. 503 (1969). 

193 Id. at 513. 

194 319 U.S. 624 (1943). 

195 See, e.g., Settle v. Dickson County Sch. Bd. 53 F.3d 152 (6th Cir. 1995). 

196 Compare Spence v. Bailey, 465 F.2d 797 (6th Cir. 1972) (applying strict scrutiny), and Moody v. Cronin, 484 F. Supp. 270 (C.D. Ill. 1979) (same), and Mozert v. Hawkins County Bd. of Educ., 827 F.2d 1058, 1070 (6th Cir. 1987) (seeming to assume that strict scrutiny was the right test, and citing a district court decision which took the same view), with Mozert, 827 F.2d at 1073 (6th Cir. 1987) (Boggs, J., concurring in the judgment) (arguing that when parents object that a school's curriculum burdens their children's religious beliefs by exposing them to material to which their religion forbids exposure, the standard of review should be even more deferential than under Tinker v. Des Moines Independent Community School District, 393 U.S. 503 (1969)). 

197 Cf. Spence, 465 F.2d at 800 (granting an exemption request in a similar case). 

198 See Hardwick v. Board of Sch. Trustees, 205 P. 49 (Cal. Ct. App. 1921). 

199 See Moody v. Cronin, 484 F. Supp. 270, 277 (C.D. Ill. 1979) (granting an exemption request in such a case); Op. Att'y Gen., 1979 WL 43163 (S.C. Nov. 13, 1979) (taking the same view). 

200 Compare Ware v. Valley Stream High Sch. Dist., 75 N.Y.2d 114, 131 (1989) (remanding for consideration of whether the compulsory AIDS education in fact constituted a substantial burden and whether an exemption would in fact frustrate the compelling government interest), with id. at 131-38 (Titone, J., dissenting) (arguing that the exemption had to be granted), and id. at 138-40 (Bellacosa, J., dissenting) (arguing that the exemption had to be denied). 

201 See Davis v. Page, 385 F. Supp. 395, 405-06 (D.N.H. 1974) (rejecting an exemption request as to most classes, but granting it as to noneducational entertainment). 

202 Cf. Citizens for Parental Rights v. San Mateo Bd. of Educ., 51 Cal. App. 3d 1, 15-16 (1975) (suggesting that Free Exercise Clause might provide a right to exemption from sex education, but not needing to reach the question because the statute already provided an adequate exemption). 

203 See Mozert v. Hawkins County Bd. of Ed., 827 F.2d 1058, 1062 (6th Cir. 1987). 

204 See id. at 1074-76 (Boggs, J., concurring in the judgment). 

205 See id. (explaining why such exposure may qualify as a substantial burden, so long as the claimant's religion does prohibit such exposure); Berg, supra note 191 (same); Stolzenberg, supra note 177, at 643 (same). 

206 See Mozert, 827 F.2d at 1071 (Kennedy, J., concurring) (suggesting that the government has a compelling interest in all forms of "[t]eaching students about complex and controversial social and moral issues" because it is "essential for preparing public school students for citizenship and self-government"). 

207 See id. at 1072-73. 

208 See David Southwell, School Boots 70 Students on 1st Day Over Dress Code, Chi. Sun-Times, Aug. 29, 1999, at 4 (discussing uniform policy that on its face exempted religious objectors); Miss. Dad, ACLU Fight School's Ban on 'Jesus Loves Me' Shirts, Commercial Appeal (Memphis), Aug. 25, 1999, available in 1999 WL 22123004. 

209 See Rourke v. New York State Dep't of Correctional Servs., 915 F. Supp. 525 (N.D.N.Y. 1995) (dealing with religious obligation to wear hair long, albeit in the context of employment rather than education). 

210 See Haynie, supra note 152 (discussing a high school policy--in this instance aimed at controlling gangs--banning the wearing of pentagrams). 

211 See supra notes 152-53 and accompanying text. 

212 See Cheema v. Thompson, 67 F.3d 883, 886 (9th Cir. 1995) (holding that Sikh boys had a right to wear blunt daggers sown in their sheaths, despite the school's arguments that even such weapons could be seen as threatening by classmates). 

213 See, e.g., Berg, supra note 13, at 32 (acknowledging that "[s]ome [Supreme Court religious exemption cases] had applied the 'compelling interest' test but only half-heartedly" and that the test "seems to promise more than it can realistically deliver"); Ira C. Lupu, The Trouble with Accommodation, 60 Geo. Wash. L. Rev. 743, 756 (1992) (calling the strict scrutiny test in religious exemption cases "strict in theory, but ever-so-gentle in fact"); McConnell, supra note 139, at 170 ("The [strict scrutiny] test . . . provided little guidance to legislatures or lower courts . . . ."); McConnell, supra note 14, at 1127 (saying that "[t]he 'compelling interest' standard is a misnomer" because the test the Court actually applied was more lenient); James E. Ryan, Note, Smith and the Religious Freedom Restoration Act: An Iconoclastic Assessment, 78 Va. L. Rev. 1407 (1992) (arguing that the supposedly strict scrutiny in religious exemption cases was in fact far from strict). 

214 476 U.S. 693 (1986).