The evidence is controverted concerning precisely what the First Congress intended the two religion provisions to mean. At a minimum, the *Establishment Clause prevented the newly created federal government from granting to any denomination the official privileges and status that the Anglican church had in England. On the other hand, clearly the clause was not intended to eliminate religious establishments then existing in several states. And the new federal government gave various supports to religion such as legislative prayers, presidential Thanksgiving proclamations, and religious missions to Native Americans.

There are likewise debates concerning the historical understanding of the Free Exercise Clause. Most of the framers intended the clause to prevent the governmental persecution of dissenting religions that occurred in England under the Anglican establishment. One strain within the founding generation also understood “free exercise” to require that religious believers be exempted from the application of laws that would force them to violate their religious beliefs. But another strain saw free exercise as simply preventing government from becoming intentionally involved in religious matters and emphasized the government’s power to legislate on secular matters even if this conflicted with religious tenets.

The range of founding-era arguments for religious freedom and disestablishment appears in James *Madison’s “Memorial and Remonstrance against Religious Assessments,” written in 1785 in opposition to Virginia’s proposed tax to support clergy. Madison’s arguments included the preeminence of duties to God, as the individual perceives them, over the demands of society; (p. 837) the unequal treatment that state favoritism in religion gives to dissenting citizens; the harms that establishment causes to the favored religion, including dilution of its independence and vigor; and the societal discord caused by government involvement in religious disputes.

De Facto Establishment.

State religious establishments, in the sense of tax support for clergy, died a natural political death early in the 1800s. The Constitution was understood to require separation of church and state at the institutional level; an establishment of religion existed when the government coerced support for a particular denomination or sect. However, this did not mean that government or politics should be secular or separate from religious influence. Indeed, as Madison’s arguments above suggest, Americans’ understanding of religious freedom itself rested in significant part on theological claims. Nineteenth-century America generally affirmed that Protestant values formed an important part of the foundation on which society was built.

Accordingly, church-state relations into the 1900s reflected what the legal historian Mark De Wolfe Howe called the “de facto Protestant establishment.” Public schools had a Protestant flavor, with teachers leading prayers and reading passages from the King James Bible without clerical comment. Legislative prayers became widespread; Thanksgiving, Good Friday, and Christmas became official holidays; and political rhetoric made frequent references to the Almighty. States prohibited blasphemy, enforced the Christian Sabbath, and forbade atheists and sometimes non-Christians to hold public offices. Protestant evangelicals rallied to pass laws prohibiting Mormon polygamy, enforcing temperance, and forbidding the teaching of evolution (see evolution and creation science). Protestant activism also fueled movements to abolish *slavery, give women the vote, and ameliorate conditions for industrial workers.

The First Amendment’s religion provisions played little direct role in these developments because the federal government’s activities were limited and, under *Barron v. Baltimore (1833), the Bill of Rights did not apply to the states. (The Court applied the Free Exercise Clause to states in *Cantwell v. Connecticut [1940] and the Establishment Clause in *Everson v. Board of Education [1947].) The only decision from this period that retains significance in contemporary constitutional law is *Reynolds v. United States (1879), which rejected a free exercise challenge by Mormons to the federal law prohibiting polygamy.

The de facto establishment melded government with religion in many ways, but usually in a diffuse and nonsectarian sense. Because Americans thought that the de jure Anglican establishment in England exemplified the kind of arrangement prohibited by the Constitution, they did not take seriously the possibility that more subtle involvements of government with religion could create similar evils. They honored institutional separation—forbidding a national church, abandoning taxes to support clergy and churches—but beyond that, they saw little reason to cabin the public influence of religion.

Modern Pluralism, Secularism, and Expanded Government.

The assumptions of the de facto establishment about the legitimacy and importance of public religious influence came under serious pressure in the twentieth century. The growth in numbers and legitimacy of non-Protestant faiths made any explicit government religious statement seem too partial toward one group of citizens. Challenges also arose from explicitly secularist ideas, such as the growth of naturalistic approaches to science spawned by Charles Darwin and others. The professionalization of American higher education replaced theologians and clergymen with secular degree-holders as university faculty and administrators. In law itself, the legal realists rejected theologically resonant *natural law as the source of optimal legal rules, looking instead to pragmatic reasoning and empirical investigation. A number of intellectuals elaborated a “secularization hypothesis,” which posited that religion would progressively lose all of its public influence.

However, religion has retained strong influence in the lives of millions of individuals, many of whom—whether they are “conservative” or “liberal” theologically—continue to insist that religion is relevant to society and public life. Religious activism in the civil rights movement of the 1960s, largely by liberals, served as precedent for religious activism by conservatives and liberals on issues such as *abortion and *homosexuality. If government had a strictly limited scope, many disputed religious-moral issues could be confined to cultural debates and would not implicate government action and the First Amendment. But at least since the *New Deal, government has grown to affect most areas of life, from education to family to economic relations, by regulating various activities and by subsidizing them. The welfare state complicates the task of keeping government out of religious matters without privatizing religion to a degree that itself violates religious liberty.

Many of these cross-cutting influences are visible in Everson, the Supreme Court’s first modern Establishment Clause decision. The question was whether a township could pay for the bus transportation of children to parochial as well as to public schools. The Court first gave an account of the history behind the First Amendment, concluding (p. 838) that the framers meant to eliminate the civil discord and violent persecution that had accompanied the establishment of a sect in both the Old World and the New. The majority relied heavily on the 1785 episode in which Virginia rejected taxes to support clergy. Citing the arguments of Madison and Jefferson in that debate, the Court stated that the Establishment Clause broadly prohibited the government from aiding not only one denomination or religion over another, but also religion over nonreligion. Quoting Thomas *Jefferson’s 1802 letter to the Danbury Baptists, the majority said that the Establishment Clause “was intended to erect a ‘wall of separation’ between church and state.”

The above language signaled that the de facto establishment would be abandoned as a guide to church-state relations. The de facto establishment called for government neutrality among different Protestant sects, but not between Christianity and other faiths, or religion and nonreligion. The de facto establishment affirmed the institutional separation of church and state, but not the more decisive division implied by the “wall of separation” metaphor. Indeed, the far-reaching language of separation and “no aid” indicated that the bus reimbursements would be forbidden.

The majority, however, turned around in the last half of the opinion and approved the reimbursements. The Court emphasized that denying religious believers the benefits of general welfare legislation would “handicap” them and make the state their “adversary” rather than “a neutral” (pp. 16, 18). Everson dramatized that in the context of welfare-state programs, church-state separation, if understood as “no aid” of any sort, could conflict with neutrality if the latter is understood as equal treatment between religious and nonreligious activities. Differing conceptions of separation and neutrality continue to underlie Religion Clause disputes.

Because direct financial support of churches and clergy by government was characteristic of the Anglican establishment, government grants of financial or other tangible aid to churches or other religious organizations have always raised significant constitutional issues. Virginia’s defeat of clergy assessments was, according to the Everson Court, central to the meaning of the Establishment Clause. Modern establishment disputes have most often involved state assistance to religiously affiliated (p. 839) schools, whether at the elementary, secondary, or college level. Although aid to such schools could similarly be said to promote their religious mission and entangle government with religious institutions, the aid also has the secular purpose and effect of promoting educational quality and, arguably, equalizing their treatment with that of secular schools that receive government aid.

Aid to Religious Elementary and Secondary Schools.

Returning to the subject two decades after Everson, the Burger Court in the 1970s struck down a number of programs providing financial aid to religious elementary and secondary education (grades K-12). The decisions relied on two separate arguments, one applying the effect prong of Lemon and the other the entanglement prong. The first argument was that religious K-12 schools are “pervasively sectarian”: they seek throughout their programs to inculcate their students with the doctrines of the sponsoring religion. Thus, state aid given directly to the school’s educational programs has the primary effect of advancing religion.

In response to this argument, many states tailored their financial grants so that they could not be used directly to teach religious values. In Lemon, for example, the state provided supplements to the salaries of private-school teachers but required that they not teach religion in their classes. But the Court found that these controls themselves created excessive entanglement between church and state, in violation of the third prong of the test. The restrictions impeded the religious schools’ ability to carry out their mission of inculcating religious values. They also required state authorities regularly to monitor how schools spent their funds.

The Court’s application of the effect and entanglement prongs ensnared religious K-12 schools in a constitutional trap. If the grants were unrestricted, the Court found a primary effect that advanced religion. If strings were attached to prevent this effect, the Court found that they created unacceptable entanglement in the religious school’s affairs. This “Catch 22,” as one decision put it, doomed every program of direct financial or other tangible assistance to religious K-12 schools for a more than a decade.

By contrast, the Court has been far more lenient toward state aid given to individual students and families who then choose whether to use it at religious schools, public schools, or secular private schools. A lengthy series of decisions approves such programs of “true private choice.” The first, Mueller v. Allen (1983) upheld a state law permitting taxpayers to deduct from their income the cost of tuition and other expenses for educating their children. Parents of public and private school children were equally eligible for the deduction. However, because tuition was the largest part of the available deduction, and public school families paid no tuition, the bulk of the tax benefit assisted families using private schools, 95 percent of which were religiously sponsored. The challengers therefore argued that the primary effect of the deduction was religious.

The Court, however, found the statistical evidence unimportant, holding that the program’s primary effect was permissible because the benefit was neutral between religious schools and other schools, and the decision of where the benefit went was “controlled by the private choices of individual parents” and was not attributable to the state. Accordingly, the state did not need to disqualify pervasively sectarian schools, or engage in any entangling surveillance to ensure that the uses of aid were properly restricted.

Subsequent decisions extended Mueller from tax deductions to affirmative government aid, and the “private choice” rationale culminated in *Zelman v. Simmons-Harris (2002). Zelman approved the inclusion of religious schools in a program offering vouchers to families of children in Cleveland’s failing public schools. The vouchers could be used at a range of private schools, religious or secular, and at participating suburban public schools; but more than 80 percent of the private schools were religious, and the suburban districts chose not to participate. The Court held, 5 to 4, that the program satisfied Lemon’s effect prong because its terms were neutral toward religion, it left to individual families the choice to use the voucher at a religious or secular school, and there were “genuine secular options” available to parents. The secular options included secular private schools and magnet schools, charter schools, and supplemental tutoring in the Cleveland public system. Because parents’ choices included these public alternatives, the majority found it irrelevant that the vouchers themselves were used mostly at religious schools. Because the percentage of vouchers used at religious schools was similar to the percentage of Ohio private schools that were religious, the Court again found the religious uses attributable to private choices by families, not to any skewing or incentives created by state rules.

Programs of direct aid to religious schools still are analyzed differently than private-choice programs such as vouchers. Mitchell v. Helms (2000) upheld the federal government’s provision of computers and other instructional equipment directly to religious schools, but five justices also reaffirmed that the equipment could not be used for religious purposes. By contrast, when aid flows through the choices of individuals, the Court requires no restrictions on its use. Mitchell upheld the equipment assistance only because two justices, O’Connor and Stephen G. *Breyer, (p. 840) who forbid religious uses of direct aid, found that the federal program had sufficient safeguards against such uses. On this narrow basis, the two cast the deciding votes in Mitchell to overrule 1970s decisions that had forbidden the provision of instructional materials. Similarly, *Agostini v. Felton (1997) overruled a 1985 decision barring the use of public-school teachers to teach remedial classes in secular subjects in religious schools. Thus, controls on the use of direct aid are still required, but they no longer necessarily create forbidden entanglement between the state and the religious school.

The distinction between direct aid and private-choice programs may not deserve the weight it has. Only two justices see it as important, and even they are more willing than their predecessors to uphold direct-aid programs. More fundamentally, the two lines of decisions rest on the two different understandings of government neutrality. The 1970s decisions striking down direct aid defined the neutral baseline in terms of secular education—especially education in the public schools, which in theory neither explicitly teaches religious values nor explicitly denigrates them. In this view, aid for religious education constitutes a special subsidy to religious-school parents “in addition to the right that they have to send their children to public schools totally at state expense” (PEARL v. Nyquist, 1973 (p. 782). By contrast, the more recent private-choice decisions treat secular education, including public education, as an alternative to religious education, which means that the provision of equal aid for religious and secular schools is neutral and permissible. Thus Zelman emphasized that the analysis of whether parents had genuine choice should count “all options Ohio provides Cleveland schoolchildren” (p. 656), including public magnet and charter schools and tutoring.

The shift in baselines corresponds to a shift in the dominant value in financial-aid cases. The 1970s decisions promoted strong church-state separation, both in forbidding the use of aid for religious teaching and in ruling that controls on the use of the aid created unconstitutional church-state entanglement. The recent decisions place more emphasis on equality between religious and nonreligious schools in the program’s formal terms, and on whether citizens have a reasonable choice among religious and secular alternatives.

Other Contexts: Higher Education and Social Services.

Outside the context of religious K-12 schools, the Court was always more willing to uphold aid. Even the 1970s Court upheld several programs of direct financial grants to religiously sponsored colleges and universities. The Court assumed that, in contrast to K-12 schools, religious colleges and universities seek to teach critical thinking skills in addition to (if not instead of) inculcating religious values. Moreover, college students are more mature than their younger counterparts and are less likely to be pressured into conforming to particular religious beliefs and practices. The Court therefore found the primary effect of grants to religious colleges was to support secular education.

Because the Court viewed the colleges as at least partly secularized, the approval of direct aid may still have rested on an assumption that secular education was the neutral baseline. The Court has never addressed a statute providing direct grants to a college where religious teaching pervades the curriculum and campus. Benefits to such a college might be constitutionally limited to assistance, such as federal Pell grants, that is provided to the students who choose to attend there. A unanimous Court in Witters v. Department of Services for the Blind (1986) confirmed that the Establishment Clause did not bar the provision of state aid to a student who chose to use it for pastoral studies at a pervasively religious Bible college.

The Court’s one decision concerning religious social services, Bowen v. Kendrick (1988), was also more lenient toward aid. The Adolescent Family Life Act of 1982 gave grants to social service organizations, including religiously affiliated ones, that conducted programs aimed at reducing teenage pregnancy. In an opinion that echoed its analysis of aid to colleges, the Court held that giving grants directly to religious social service organizations did not violate the Establishment Clause so long as the recipients were not pervasively sectarian. The Court may soon revisit the issue of aid to social services because of the recent federal initiative to include “faith-based” services on an equal footing in aid programs. The Court seems likely to follow the same analytical distinction between direct-aid and private-choice programs that it has followed concerning education.

Constitutional Rights to Equal Aid.

As Establishment Clause limits on aid have receded, the question has arisen whether the equal inclusion of religious recipients in aid programs is not just constitutionally permitted, but constitutionally required. Aid proponents argued that to single out religious education for exclusion from benefit programs discriminates against religious activity and unconstitutionally requires families or schools to forego religious exercise as a condition of receiving educational benefits. The issue pits the Rehnquist Court’s increasing tilt toward equal treatment of religion and nonreligion against its concern to preserve states’ discretion in how to allocate funds.

*Rosenberger v. Rector and Visitors of University of Virginia (1995) held that when a state university (p. 841) paid printing and other expenses for a wide range of student publications, it could not refuse to do the same for a student magazine written from an evangelical Christian perspective. The refusal to assist the magazine unconstitutionally discriminated against its religious viewpoint by excluding it from a “limited *public forum” (see freedom of speech). But recently, in Locke v. Davey (2004), the Court held that a state that provided scholarships to college students based on grades and family income could exclude an otherwise qualified student because he was majoring in theology “from a devotional perspective” in preparation for the ministry.

Locke limited Rosenberger’s holding to cases of funding solely to facilitate student speech activities as opposed to funding for educational programs. Locke also treated the training of clergy as a “distinct category of instruction” from subjects such as literature or math. Much of the opinion indicates that the Court will defer to the constitutional provisions in many states that explicitly forbid aid to support “sectarian” schools or instruction—although several passages suggest that the training of ministers is a special case where the state’s interest in denying aid is especially strong. A broad reading of Locke, combined with other recent decisions, would mean that states usually may include religious schools in aid programs but are not required to do so.

American government has long taken steps to promote or recognize religion in ways not involving financial aid: prayers in public schools and in legislatures, official Thanksgiving proclamations, and so forth. Such actions were pervasive in the de facto establishment, and the Court did not begin to scrutinize them until after Everson’s articulation of neutrality and separation principles. By that time, most of the overtly coercive practices had ceased, and relatively few cases of rank coercion have reached the Court. The more common argument against nonfinancial support of religion is that such actions depart from neutrality and officially endorse one religious position over another, or religion over nonreligion. As noted earlier, such endorsement is assertedly wrong not because it directly deprives anyone of liberty, but because—no matter how nondenominational and generic it is intended to be—it creates unwarranted discord and alienates those citizens with views on religion different from the government’s.

At the same time, even in today’s pluralistic and secular context, many citizens believe that eliminating all these practices would artificially and unwisely secularize American public life. That such sentiments are widespread became apparent in the angry reaction to Newdow v. U.S. Congress (Ninth Circuit, 2002), the court of appeals decision that invalidated the words “under God” in the Pledge of Allegiance, as recited in a public school. The Supreme Court ultimately overturned this ruling, in Elk Grove Unified School District v. Newdow (2004), on the ground that the noncustodial parent challenging the Pledge lacked standing to sue to raise the challenge. And three concurring justices wrote that they would have upheld the “under God” phrase on the merits, showing their discomfort with removing all religious elements from official school activities.

Another countervailing concern is to ensure that, even if all government-sponsored religious activity is eliminated under the Establishment Clause, religious activity by private individuals and groups in public settings retains protection under the Free Speech and Free Exercise clauses.

Government-Sponsored Religion in Public Schools.

As with financial aid to religion, a principal battleground concerning official religious activities has been the public schools. Since the early 1960s, the Court has consistently found a violation of the Establishment Clause when public elementary or secondary schools have institutionally sponsored or favored a religious activity, including organized classroom prayers and Bible readings, prayers at commencement or other school exercises, mandatory moments of silence in the classroom to encourage personal prayers, posting religious texts like the Ten Commandments, and teaching religious doctrines such as the Biblical creation account (see school prayer and bible reading). As noted above, “under God” remains in the Pledge of Allegiance as recited in schools because the majority in Newdow rejected the plaintiff’s standing to sue, not because it upheld the practice on the merits.

The decisions striking down school-sponsored religion rest on broad concerns about endorsement of religion—the alienation of dissenting students and families from the school they support with their taxes and their energy. But the Court has also mentioned that dangers of actual coercion are significant in K-12 schools, whose young students are most vulnerable to overt or subtle pressures. In Lee v. Weisman (1992), the Rehnquist Court, to the surprise of many, held that a school violated the Establishment Clause by inviting a clergyman to give a brief prayer at graduation exercises. Far from overruling the earlier school prayer decisions, Weisman extended their principles from the classroom to other school events. The majority opinion avoided the nonendorsement test and instead employed the noncoercion principle, which many observers expected would lead to upholding school-sponsored prayer so long as no one was formally required to join in it. But Weisman’s concept of coercion was quite (p. 842) broad: dissenters did not have to join in the prayer, but were coerced simply by having to stand silently during it. Eventually, the Rehnquist Court majority explicitly embraced the stiffer non-endorsement test in its 2000 decision invalidating officially sponsored prayers at football games.

The only decision approving on the merits a program in which public schools arguably engaged in official promotion of religion was *Zorach v. Clauson (1952), which upheld the practice of releasing students from their normal classes once a week to travel to religious instruction classes held off campus (see released time). Just four years earlier, the Court in *Illinois ex rel. McCollum v. Board of Education (1948) had found a violation of the Establishment Clause in a program under which priests, ministers, and rabbis came into school classrooms to give religious instruction to students affiliated with their respective congregations. The Court found that formal integration of on-campus religious instruction into the public school curriculum gave unconstitutional aid to religion.

The Court distinguished the two cases based on the on- versus off-campus location of the instruction, but that distinction has been widely criticized. A more promising approach might be to note that in McCollum students had to act affirmatively to be excused from the religious classes, whereas in Zorach students had to act affirmatively in order to attend them. Thus in McCollum the school was implicated directly in the effort to provide religious instruction, whereas in Zorach it merely responded to the independent choices of students and parents to receive religious instruction.

The decisions invalidating school-sponsored religion rest on a baseline in which secular education is viewed as neutral; the Court has repeatedly rejected contentions that a public education devoid of religious exercises establishes a “religion of secularism”(*Abington School District v. Schempp, 1963, p. 225). The case of “under God” in the pledge may test how staunchly the Court believes that removing references to religion in public exercises is the neutral course. In any event, it is noteworthy that the Rehnquist Court has retained, indeed strengthened, the secular-education baseline in cases of official religious exercises, even as it has departed from that baseline in cases involving aid to individuals attending private religious schools (see previous section). The difference can be explained by the fact that, while aid programs leave it to individuals to choose whatever school (religious or secular) they prefer, officially sponsored religion must always express a limited view or set of views, most likely the views favored by the majority.

Voluntary Religious Activity in Public Schools.

By contrast, when religious activity in public schools results from choices by students, parents, or other private individuals, the Court has repeatedly held that the activity can or even must be permitted. The progression began with Widmar v. Vincent (1981), which held by an 8-to-1 vote that religiously oriented student organizations at a state university were entitled to the same access to meeting rooms as were secular student groups. This rule was effectively applied to high school students in Board of Education v. Mergens (1990), which held that the Establishment Clause did not require a school to keep a student prayer group from meeting in classrooms during an after-class “activity period” on the same terms as other student clubs. Mergens therefore upheld the constitutionality of the Equal Access Act of 1984, which gave student groups a statutory right to meet. The sessions were student initiated; faculty members took no active part and attended only as safety monitors; and many other clubs met during the activity period, including intramural sports, music, academic, and social clubs. Based on these facts, the Court held that any advancement or endorsement of religion came from the individual students, not from the school.

Later decisions applied the equal access principle to increasingly controversial situations, including the provision of affirmative funds (not just facilities) for religious groups at a state university (Rosenberger), and the meeting of a religious club for elementary school students as part of a “limited public forum” of clubs immediately after school (*Good News Club v. Milford Central School, 2001). The Court repeatedly held that exclusion of religious clubs was discrimination by viewpoint, the clearest category of free speech violation, and that religious worship and instruction could not be treated worse in this context than speech addressing moral issues in nonreligious terms. Thus, in cases about access to facilities, the Court treats religious and secular viewpoints as alternatives, and equal access for religion as the neutral course—not merely permitted, but required.

Nonschool Contexts.

Outside of public schools, the Court has been considerably more lenient in permitting actions that endorse religion so long as the actions suggest no denominational preference. In Lynch v. Donnelly (1984), the Court upheld a city’s display of a nativity scene on city property near a display of Santa and his reindeer. In Allegheny County v. American Civil Liberties Union (1989), a majority held that a menorah displayed next to a large Christmas tree outside a county office building was constitutional—although a different majority struck down the placement of a Christmas creche alone in a prominent place in the county courthouse because, without any nearby (p. 843) displays representing other religions or secular images, the creche conveyed an endorsement of Christianity.

The principle of equal access for privately initiated speech and symbols applies outside the public schools as well. In Capitol Square Review and Advisory Board v. Pinette (1995), a divided Court held that the Ku Klux Klan had the right to display a large, freestanding cross in front of the state capitol, in a square used for speech by other groups.

It has long been widely accepted that religious freedom prohibits the government from directly punishing or regulating religious beliefs. Moreover, as the previous section shows, speech with religious content or motivation receives strong protection under the Free Speech as well as the Free Exercise clauses. And although conduct may be regulated more than belief, the government unquestionably may not single out religiously motivated conduct for punishment. For example, Church of Lukumi Babalu Aye v. City of Hialeah (1993) unanimously struck down a series of ordinances that criminalized killing animals only as part of a ritual sacrifice engaged in by Santeria worshipers.

Mandated Exemptions for Religious Conduct.

Litigation under the Free Exercise Clause, therefore, has largely concerned whether the government must exempt religiously motivated conduct from the application of a general law that conflicts with religious conscience. The question is whether the clause guarantees believers or churches the right to refuse to perform a legally required act that violates their religious beliefs (for example, employing women clergy on a nondiscriminatory basis) or the right to perform a legally prohibited act required by their beliefs (for example, ingesting an illegal drug as a sacrament). If such a guarantee exists, it cannot be absolute—consider the case of ritual human sacrifice—and thus the precise scope of mandated exemptions poses challenging questions.

Reynolds, the 1879 decision on Mormon polygamy, rejected a constitutionally compelled exemption and explicitly limited free exercise protection to beliefs alone. But the belief-conduct distinction was overridden in two modern decisions. In Sherbert v. Verner (1963) the Court ordered a state to pay unemployment benefits to a Seventh-Day Adventist even though she would not accept an available job, as the unemployment compensation law required, because it included work on Saturday, her Sabbath. The state asserted a compelling interest in protecting the integrity of the compensation fund against depletion by those not truly looking for work. But the Court held that government could burden a fundamental right like the free exercise of religion only if it was protecting a compelling interest by the least restrictive means possible, and it found that the state had not met this burden. Several later decisions have affirmed this holding concerning unemployment benefits.

The second case, *Wisconsin v. Yoder (1972), held that Amish parents were not required to send their children to school past the eighth grade in violation of their religious beliefs because the state failed to show that its compelling interests in education—ensuring self-sufficiency and training for citizenship—would be significantly undermined by granting the Amish an exemption from compulsory attendance laws. Yoder reasoned that the state must offer a compelling reason not just for the law in question, but also for denying an exemption to the particular religious objector. This substantially raised the government’s burden of justifying any law that incidentally burdened religious exercise.

The Sherbert-Yoder doctrine of exemptions can be said to protect the value of liberty or private choice in religious matters, since even a nondiscriminatory law can inflict serious burdens on religion in application. Exemption also arguably comports more with strong church-state separation by blocking government regulation of religion. The implications of equality for religious exemptions are more ambiguous. Mandatory exemptions can serve to equalize religions in practice, since minority religions are more likely to conflict with general legal norms than are majority or acculturated faiths. But if equality means how a law on its face treats religious versus nonreligious activities—what law professor Douglas Laycock calls “formal” equality—then exemptions limited to religiously motivated conduct are not required and may even be improper.

After Yoder, the justices soon began retreating from a strong exemptions position, partly because some were attracted to formal equality, partly because of practical challenges that the Sherbert-Yoder doctrine posed when combined with other holdings of the Court. Religious objectors usually constitute a small minority whose disobedience of general legislation would rarely undermine its general effectiveness. Moreover, United States v. Ballard (1944) effectively foreclosed judicial inquiry into the sincerity or reasonableness of religious beliefs. Thus, when read with Ballard, Sherbert and Yoder appeared to require that an exemption from law be granted to nearly any religious objector who asked for one— a scope of freedom that many justices began to find too broad.

Constitutionally mandated exemptions pose less of a difficulty when the benefit to be gained from exemption is something few people would want, such as freedom to keep children home from school. In United States v. Lee (1982), however, the (p. 844) Amish asked for a free exercise exemption from paying social security taxes. Perhaps fearing a tidal wave of exemption lawsuits by people claiming that their religious beliefs forbade paying various taxes, the Court found the government’s interest in denying the exemption to be compelling. Lee was just the first step away from exemptions. In quick succession, the Court denied the claims of a Native American who objected to his daughter receiving a social security number, an Orthodox Jew who sought to wear a yarmulke in violation of Air Force uniform regulations, and a Native American tribe that objected to construction of a federal highway that would prevent them from worshiping on lands they viewed as sacred. These three decisions refused to find a constitutionally relevant burden on religion in the first place, so that government need offer no compelling reason for its action.

In Employment Division v. Smith (1990), the Court did away with mandatory exemptions in a wide range of cases. Two members of the Native American Church were denied unemployment compensation after they were dismissed from their jobs for ingesting the drug peyote at a church ritual. The state argued that because the use of peyote was criminal under state law, the Native Americans had committed “work-related misconduct” and could be denied benefits. The Native Americans argued that the Free Exercise Clause as interpreted in Sherbert prevented the state from applying the misconduct standard to their religious practice. But the Court held that the denial of compensation was constitutional simply because it rested on a “neutral law of general applicability,” the criminal prohibition against peyote use. Under this rule, the importance of the state’s interest or the degree of imposition on religious practice were irrelevant.

The majority opinion by Justice Antonin *Scalia emphasized primarily that judges could not draw principled lines for balancing religious practices against competing governmental interests. In dealing with precedent, Smith recast Yoder as a case involving a “hybrid” of free exercise combined with substantive due process rights of parental and family autonomy. Sherbert was distinguished on the ground that it involved not an across-the-board prohibition, but the general standard of “good cause” that allowed refusals of work for some secular reasons and therefore must recognize religious reasons as well. If taken broadly, Smith makes the Free Exercise Clause essentially redundant to other constitutional rights such as free speech.

The terms of free exercise law are now set by reactions to and interpretations of Smith. Congress reacted negatively by passing the Religious Freedom Restoration Act of 1993 (RFRA), which reinstituted the Sherbert-Yoder “compelling interest” test as a statutory protection against “substantial burdens” on religious exercise. However, in *City of Boerne v. Flores (1997) the Court struck down RFRA as applied to state and local actions on the ground that it exceeded Congress’s powers to enforce *Fourteenth Amendment rights. Congress, the Court said, could not legislate based on a broader understanding of free exercise than the Court had enunciated in Smith.

Boerne leaves open several questions about the governing test. RFRA may remain valid as applied to actions of the federal government, and several states have passed their own versions of RFRA. A newer statute, the Religious Land Use and Institutionalized Persons Act of 2000 (RLUIPA), imposes strict judicial scrutiny on state and local laws concerning zoning, architectural preservation, and prison conditions. All of these statutes have been challenged on the ground that exemptions limited to religious conduct, without exemptions for comparable secular conduct, violate the Establishment Clause. The Court granted certiorari in Cutter v. Wilkinson in 2004 to address the Establishment Clause challenge to RLUIPA in the prison context. The case may also raise questions about congressional authority to pass the statute under the *Commerce Power and the *Spending Power.

Meanwhile, constitutional free exercise claims themselves continue to be governed by Employment Division v. Smith. Interpretation of Smith’s rule has centered on what makes a law “neutral and generally applicable.” Although most lower courts have read the rule to eliminate exemptions, some have held that when a law makes exceptions for secular conduct, it is not generally applicable and religious practice must be exempt as well. Since most laws include some exceptions, this approach would strip Smith of much of its effect. The Court likely will have to resolve this question soon.

Smith also approved the legislature’s ability to write religious exemptions into laws, a practice that the Court had upheld against an Establishment Clause challenge in Corporation of Presiding Bishop v. Amos (1987). Discretion for legislative exemptions when combined with the elimination of constitutionally mandated exemptions may mean that politically powerful religions will be able to lobby successfully for protection while smaller or powerless faiths will be dependent on the goodwill of political majorities. Sometimes legislatures protect minority practices—sacramental peyote use is exempted in many states—but other times they do not, as the relentless federal campaign against Mormon polygamy showed.

A final area of free exercise doctrine relates to adjudication by secular courts of disputes (p. 845) within a religious organization. Situations include contests between factions in a church over who controls the property or offices of the church and claims by individual employees or members that the church has discriminated against them or otherwise treated them illegally. The Court has suggested two approaches to such cases. Under the first, enunciated in Watson v. Jones (1872), the court defers to the resolution of the matter by the church’s governing authority, whether that is a hierarchical tribunal or the majority of the individual congregation. Under the second approach, approved in Jones v. Wolf (1979), the court may apply secular legal principles even if these override the church authority’s decision as long as the secular principles involve no interpretation of religious doctrine. If doctrine must be interpreted, the court must defer to the interpretation by the church’s governing structure.

State courts resolving church property disputes may choose either of these approaches, and many now apply principles of secular property and trust law. By contrast, in disputes directly involving clergy, the Court, and lower courts even after Employment Division v. Smith, have deferred to church authorities rather than apply secular employment laws. The reasoning is that judicial intervention would place the court on one side of a religious controversy and would likely require the court to resolve theological questions such as whether a clergyman performed his job competently. But as to less central religious positions, church autonomy is less likely to prevail over secular laws, unless the case requires determinations of religious doctrine.

Christopher L. Eisgruber and Lawrence G. Sager, “Congressional Power and Religious Liberty after City of Boerne v. Flores,” Supreme Court Review (1997): 79–139. Frederick M. Gedicks, The Rhetoric of Church and State: A Critical Analysis of Religion Clause Jurisprudence (1995). Philip A. Hamburger, “A Constitutional Right of Religious Exemption: An Historical Analysis,” George Washington Law Review 60 (1992): 915–948. John C. Jeffries, Jr., and James E. Ryan, “A Political History of the Establishment Clause,” Michigan Law Review 100 (2001): 279–370. Douglas Laycock, “Formal, Substantive, and Disaggregated Neutrality Toward Religion,” DePaul Law Review 39 (1990): 993–1022. Ira C. Lupu, “Government Messages and Government Money: Santa Fe, Mitchell v. Helms, and the Arc of the Establishment Clause,” William and Mary Law Review 42 (2001): 771–822. Michael W. McConnell, “Religious Freedom at a Crossroads,” University of Chicago Law Review 59 (1992): 115–194. Michael W. McConnell, “The Origins and Historical Understanding of Free Exercise of Religion,” Harvard Law Review 103 (1990): 1409–1517. Steven D. Smith, Foreordained Failure: The Quest for a Constitutional Principle of Religious Freedom (1995). John Witte, Jr., “The Essential Rights and Liberties of Religion in the American Constitutional Experiment,” Notre Dame Law Review 71 (1996): 371–445.

Frederick Mark Gedicks; revised by Thomas C. Berg