Establishing freedom of religion as both constitutional principle and social reality is among America’s greatest contributions to the world. Nevertheless, the concept of free exercise of religion is not self-defining. The boundaries of free exercise, like those of other rights, must be delineated as against the claims of society and of other individuals. The history of the Free Exercise of Religion Clause, in both its original understanding and modern interpretations, reveals two recurring impulses, one giving free exercise a broad, the other a narrow, scope. The narrower view sometimes collapses free exercise into other constitutional rights, for example treating religious activity as no more than a variety of speech or expression. The broader view sees the right of choice in religious practice as independently valuable. The tension between broad and narrow rights has played out in five sets of issues under the Free Exercise of Religion Clause: belief vs. conduct; discriminatory vs. generally applicable laws; institutional free exercise and internal vs. outward acts; religion vs. secular conscience; and the scope of “prohibiting” vs. “burdening.”

The first key issue concerns the meaning of the protected “exercise” of religion: does it encompass only the belief and profession of a religion, or does it also protect conduct that stems from religious tenets or motivations; for example, wearing a head covering or religious garb, or refusing to accept blood transfusions or other medical treatment?

The great weight of the original understanding controverts the narrowest interpretation of the text, that is, that belief alone is protected. At the Founding, as today, dictionaries defined “exercise” to include action, not just internal belief. Thomas Jefferson, in his famous 1802 “wall of separation” letter to the Danbury Baptist Association, drew a sharp distinction between protected belief and unprotected action: “the legitimate powers of government reach actions only, and not opinions” and “[man] has no natural right in opposition to his social duties.” But a number of statements from other leading figures support the broader view—from James Madison’s statement that religion includes “the manner of discharging” duties to God, to William Penn’s statement that “liberty of conscience [means] not only a meer liberty of the mind, in believing or disbelieving . . . but the exercise of ourselves in a visible way of worship.”

In its first interpretation of the Free Exercise of Religion Clause, Reynolds v. United States (1879), the Supreme Court drew a sharp line between belief and action, relying on Jefferson’s letter to the Danbury Baptists, in holding that the Mormon practice of polygamy was not protected. Since then, however, the Court has ruled more frequently in line with the original meaning, protecting certain religiously motivated actions such as proselytization, Cantwell v. Connecticut (1940), refusing work on one’s Sabbath, Sherbert v. Verner (1963), choosing the education of one’s children, Wisconsin v. Yoder (1972), and sacrificing animals at a worship service, Church of Lukumi Babalu Aye v. City of Hialeah (1993). Action inevitably receives less protection than belief, but it is not unprotected.

Reynolds, which involved federal and territorial laws against polygamy, was the Court’s only free exercise decision on the merits until the middle of the twentieth century, since the clause applied only to acts by the federal government. During that time, the significance of the Free Exercise of Religion Clause lay less in its legal effect than in its affirmation of the value of religion in American culture. In 1940, however, in Cantwell, the Court “incorporated” the Free Exercise of Religion Clause into the Due Process Clause of the Fourteenth Amendment and applied it to the states. Subsequently, most contests over free exercise have involved state laws.

A second key issue involves discriminatory vs. generally applicable laws. Because it is now accepted that the Free Exercise of Religion Clause protects religiously motivated conduct as well as belief, the most important modern issue has been whether the clause only prohibits laws that target religion itself for restriction, or more broadly requires an exemption in some cases even from a generally applicable law that happens to conflict with a particular religious practice. To take just one of many examples, must an Orthodox Jewish military officer, who is religiously obligated to wear a yarmulke, be exempted from a general rule forbidding all servicemen to wear anything other than official headgear?

On this issue, the text of the clause can support either the narrow or the broad reading. A law could well be said to be “prohibiting the free exercise [of religion]” if it in fact prohibits a religious practice, even if it does so incidentally, rather than overtly or intentionally. On the other hand, one might argue that the legislature does not “make [a] law prohibiting the free exercise” unless the prohibition or restriction on religion is part of the law’s very terms or is the legislature’s intent, as opposed to simply the effect of the law in a particular application.

This issue therefore requires examination of the legal background and the Founding generation’s attitude toward conflicts between law and religious conscience. By 1789, all but one of the states had free-exercise–type provisions in their constitutions. Many of these state guarantees included provisos that such freedom would not justify, or could be denied for, practices that “disturb[ed] the public peace” or were “inconsistent with the peace and safety of the State.” Michael McConnell has argued that the provisos reflect the broader, pro-exemptions conception of free exercise, because if religious practices were subject to all general laws, there would be no reason to identify a subset of laws that protected the peace of the state. In response, Philip Hamburger has asserted that in eighteenth-century legal terminology, “every breach of law [was] against the peace [of the state],” so that the provisos would have been triggered by any secular law of general applicability.

The legal background also includes accommodations made by colonial and state legislatures for specific religious practices. Virtually all states by 1789 allowed Quakers to testify or vote by an affirmation rather than an oath; several colonies had exempted Quakers and Mennonites from service in the militia; and there was a patchwork of other exemptions throughout the states. Supporters of the narrower view of the Free Exercise of Religion Clause, such as Professor Hamburger, argue that these examples imply only that specific statutory exemptions may be granted by legislative grace. But advocates of the broader interpretation, such as Professor McConnell, infer that the Founding generation thought that exemption from the law was the appropriate response to conflicts between legal and religious duties, that is, an exemption was part of the meaning of “free exercise” so long as the religious activity did not harm public peace or others’ rights.

More deeply, the question of exemptions from generally applicable laws implicates ideological differences over the relationship between civil government and religion. One important philosophical influence on the Founders, the Enlightenment liberalism stemming from the writings of John Locke, does not lend itself easily to exempting religious practice from general secular laws. In his famous Letter Concerning Toleration (1689), Locke argued that the proper domains of government and religion were largely separate; “the power of civil government . . . is confined to the care of the things of this world,” whereas “churches have [no] jurisdiction in worldly matters.” Although this limit on government control over belief and doctrine was liberal for its time, just as central to Locke’s understanding was the limit on religion’s role in worldly matters. And in those cases where both religion and government claimed jurisdiction—that is where religious duties clashed with general laws, and an exemption is sought—Locke gave the nod to the government on the ground that “the private judgment of any person concerning a law enacted in political matters . . . does not take away the obligation of that law, nor deserve a dispensation.”

The Enlightenment view, however, was hardly the dominant impetus for religious freedom in America. Popular support for religious freedom came most heavily from the newer evangelical Protestant sects, especially the Baptists and Presbyterians. These religious “enthusiasts,” who helped defeat religious taxes in Virginia and elect James Madison to Congress, began from a different premise: that religion was a matter of duties to God, and that God, in the words of Massachusetts Baptist leader Isaac Backus, “is to be obeyed rather than any man.” Madison echoed these ideas in his Memorial and Remonstrance Against Religious Assessments (1785), arguing that the duty to the Creator “is precedent, both in order of time and in degree of obligation, to the claims of Civil Society”; thus, everyone who joins a civil society must “do it with a saving of his allegiance to the Universal Sovereign.” This view logically suggests that the proper governmental response to conflicts between legal and religious duties is, at least sometimes, exemption from legal duties.

Whether religious exemptions from generally applicable laws are ever constitutionally mandated has been the central question in this area for many years. After rejecting mandated exemptions for many years, the Supreme Court switched course and exempted religious claimants from laws in Sherbert v. Verner and Wisconsin v. Yoder. In Sherbert, the Court struck down a state law that denied unemployment benefits to a Seventh-Day Adventist on the basis that she had refused “available” work, when her religion forbade her from working or being available for work on Saturday. Because a different provision of state law specifically barred employers from firing or penalizing employees who objected to Sunday work, the state’s laws overall discriminated against Saturday sabbatarians. However, much of the Sherbert opinion’s language cut more broadly, and subsequent decisions interpreted Sherbert as a protection for religiously based objections to laws that were clearly generally applicable. Thus, in Yoder the Court held that the Free Exercise of Religion Clause protected members of the Amish faith from having to abide by a compulsory school attendance law.

The pro-exemptions approach, however, was often applied half-heartedly in the next two decades, and in Employment Division, Department of Human Resources of Oregon v. Smith (1990) the Court declared that the Free Exercise of Religion Clause did not grant an exemption from generally applicable drug law to members of a Native American religion who used peyote in its religious services. The Court abandoned the pro-exemptions approach in most cases, holding that exemptions are not required from a “neutral law of general applicability.” Because most restrictions on religious conduct today come from the application of general laws rather than from laws targeting religion, Smith greatly limited the protections accorded religiously motivated actions.

In response to Smith, Congress passed the Religious Freedom Restoration Act of 1993 (RFRA), reinstating the Sherbert-Yoder test that laws that “substantially burden” religion, even if they are neutral and generally applicable, must be justified as the “least restrictive means” of achieving a “compelling governmental interest.” Nonetheless, in City of Boerne v. Flores (1997), the Supreme Court struck down RFRA as applied to state and local laws, on the ground that Congress exceeded its power to enforce the Fourteenth Amendment in attempting to define the constitutional parameters of the (incorporated) Free Exercise of Religion Clause. RFRA, however, remains applicable to federal laws and regulations, and a number of states have passed their own versions of RFRA. Thus, the rule concerning exemptions from general laws remains divided under modern law, just as there is division and ambivalence in the original understanding of the Free Exercise of Religion Clause.

Questions also remain over what makes a law fail the “neutral, generally applicable” test. In Lukumi, the Court held that ordinances prohibiting the killing of animals fell “far below” the standard because they contained so many exceptions that they effectively targeted only the ritual sacrifices of the Santeria religion. Many laws, however, contain some secular exceptions but not so many as to target religion. In Fraternal Order of Police v. City of Newark (1999), the Third Circuit, through then-Judge Samuel Alito, held that a police department’s rule against officers wearing beards violated a Muslim officer’s free exercise rights because the rule contained an exception for officers with medical conditions requiring beards. In the court’s view, the department’s willingness to accommodate a secular interest but not a religious need “devalued” religion, triggering strict scrutiny. But other lower courts have required that a law contain more than one secular exception in order to find it not “generally applicable.”

The question whether the provision of secular accommodations triggers a duty to accommodate religious objectors also arises in lawsuits over the Obama administration’s rule requiring employers to include contraception, including some “emergency contraception” medicines that may act as abortifacients, in their employees’ health-insurance coverage. The rule exempts churches, but many other religious organizations (social services, hospitals, and schools), along with commercial businesses run by religious individuals, have objected that it would force them to support acts they believe are sinful, including abortions of new embryos. Suing under both the Free Exercise of Religion Clause and RFRA, the organizations argue that the government must exempt them because it has already recognized numerous exceptions for small businesses, “grandfathered” plans, and businesses receiving case-by-case waivers on ground of hardship.

A third key issue regards the institutional free exercise of religion and internal vs. outward acts. Although many Free Exercise of Religion Clause cases involve the religious practices of individuals, questions also arise whether religious institutions enjoy distinctive protection, especially for their internal governance. Several Supreme Court decisions affirm institutional rights, beginning with Watson v. Jones (1872), the first in a line of disputes over property ownership in the wake of schisms within denominations. Watson held that courts should resolve such disputes by accepting the decision of the body’s highest authority: for a hierarchically organized church, the highest tribunal, and for a congregationally organized church, the congregation in question. Watson’s principles rested on general common law, but the Court adopted them for the Free Exercise of Religion Clause in Kedroff v. St. Nicholas Cathedral (1952). Kedroff affirmed a religious organization’s “power to decide for themselves, free from state interference, matters of church government as well as those of faith and doctrine.” It held that a state statute transferring control of church property from one Russian Orthodox bishop to another intruded on what was “strictly a matter of ecclesiastical government”: the power of the church’s highest authority to appoint the ruling bishop for North America.

However, in Jones v. Wolf (1979), the Court determined that courts deciding church-property disputes could apply “neutral principles” of property, trust, or contract law applicable to organizations in general. Jones, coupled with the ruling a few years later in Employment Division v. Smith, led many to question whether special free exercise protection for religious organizations survived.

Even after Jones and Smith, however, lower courts continued to recognize the “ministerial exception” to Title VII and other anti-discrimination laws. Under that exception, a court cannot hear a lawsuit by a minister challenging a religious employer’s decision to dismiss him, refuse to hire him, or control the terms of his employment. The lower courts pointed out that Smith had continued to prohibit the government from intervening in “controversies over religious dogma or authority.” And in Hosanna-Tabor Evangelical Lutheran Church and School v. EEOC (2012), the Court unanimously affirmed the ministerial exception, holding that the Free Exercise of Religion Clause guarantees an organization “the authority to select and control who will minister to the faithful.”

In Hosanna-Tabor, the federal government argued that Smith’s rejection of exemptions had undermined the ministerial exception, and that protection for churches’ decisions concerning ministers rested largely on the First Amendment’s freedom of expressive association. The government’s position in Hosanna-Tabor thus epitomized the narrow view of religious freedom, all but reducing it to expressive rights held equally by secular organizations. Hosanna-Tabor resoundingly rejected the narrow position, calling it “hard to square with the text of the First Amendment itself, which gives special­ solicitude to the rights of religious organizations”; it rejected “the remarkable view that the Religion Clauses have nothing to say about a religious organization’s­ freedom to select its own ministers.”

The Court held that Smith’s rejection of exemptions extended only to “outward physical acts,” such as the ingestion of peyote in that case, and not to “internal decisions” concerning a religious organization’s governance or doctrine. Future cases will surely explore this line. One principle should be clear: a religious organization does not lose distinctive free-exercise protection simply because it turns “outward” to the broader society by running schools or social services. This principle is implicated in several recent controversies; the contraception mandate, for example, initially denied protection entirely to a religious organization if it served persons outside its own faith, or if it served anyone through health or social services, rather than preaching or teaching (“inculcat[ing] religious values”). Such a provision confines free-exercise protection to the narrowest category of the house of worship— reflecting the view of many proponents that once religion enters any area that might be possibly be described as “public,” it must follow all the rules of the regulatory state. But Hosanna-Tabor rejects this narrow approach, confirming that a non-church organization such as a school has institutional religious-freedom rights. The question remains, however, what acts of a school, hospital, or social service count as “internal” matters of doctrine and governance.

The original understanding supports the freedom of religious organizations to decide certain matters of internal governance without state interference—especially to select and control religious leaders. The clergy-selection question has been sensitive in Western history for almost a thousand years, at least since the medieval controversy between popes and Holy Roman Emperors over who had the power to appoint bishops. Although that dispute involved a fundamental clash between two authorities seeking predominance, the rough compromise that ultimately emerged—popes appointed bishops, emperors appointed civil officials—reflected a solution of separating certain core powers of the church and of the state.

Early Protestantism, in struggling with the Catholic Church, often sought assistance from civil rulers, sometimes to the point of letting them control clergy selection and other important religious functions. In the Church of England, the most familiar example of an establishment to the American Founders, the government appointed leading clerics, the monarch was official head of the church, and Parliament approved the Thirty-Nine Articles (the church’s doctrinal tenets) and the Book of Common Prayer.

The First Amendment, in its non-establishment as well as its free exercise provision, confirms that Americans rejected any role for the federal government in choosing church leaders. In 1783 the Vatican proposed an agreement with Congress to approve a Bishop-Apostolic for America now that the new states were outside English authority; but Congress responded that it had “no authority to permit or refuse” the appointment, and the Pope could appoint whomever he wished, because “[t]he subject . . . being purely spiritual, it is without the jurisdiction and powers of Congress.” James Madison, as secretary of state in 1806, reaffirmed that the civil government had no power over the purely “ecclesiastical” matters of choosing a Catholic leader for the Louisiana Territory.

Leading Founding-era proponents of the First Amendment understood it to protect religious institutions’ autonomy, especially concerning clergy selection. Isaac Backus, the leader of Massachusetts Baptists, wrote in 1773 that, “God has appointed two kinds of government in the world which are distinct in their nature and ought never to be confounded together”—civil and ecclesiastical government—and therefore it was solely a church prerogative to determine what [God’s] worship shall be, who shall minister in it, and how they shall be supported.” The same sentiments were expressed by Enlightenment states-men such as Madison, who as President vetoed a bill incorporating a church in the District of Columbia on the ground, in part, that the bill enacted rules “relative purely to the organization and polity of the church incorporated, and comprehending even the election and removal of the Minister of the same, so that no change could be made therein by” the congregation or the denomination. After the U.S. took control of the Louisiana Territory, President Jefferson, responding to a letter from Ursuline nuns who ran a school in New Orleans, assured them that “the principles of the Constitution of the United States” guarantee that “your institution will be permitted to govern itself according to its own voluntary rules, without interference from the civil authority.”

Civil involvement in clergy selection, a feature of several colonial arrangements, disappeared during the Founding and early republic. In New England and the Southern colonies, civil authorities regulated the conduct of clergy in the established church and at first prohibited, then licensed, religious teachers from dissenting sects. Such measures triggered massive resistance from Baptists and other dissenters, and they disappeared as part of the commitment both to disestablishment and to free exercise. Clergy-selection issues also led to the demise in 1833 of the Massachusetts religious establishment under which a majority of a town’s voters determined which clergyman would presumptively receive tax-funded subsidies and occupy the town’s “First Church.” When Unitarians began to control some towns, and courts awarded them control of the First Church and of tax funds, support for the tax system collapsed. This was disestablishment, but it was also a victory for a congregation’s free exercise of religion against civil (town) interference.

In Hosanna-Tabor, the Court refused to limit the ministerial exception to cases where the government explicitly second-guessed a church’s religious criteria for selecting its minister. The original understanding supports this holding; it shows that government can trigger the harms that the Religion clauses seek to avoid—coercion, divisiveness, government overreaching—simply by overriding a church’s decision on who is suitable to be its minister. Under the Massachusetts scheme for clergy taxes, the majority of town voters needed no theological rationale for selecting one clergyman for the “First” church; they could choose based on any factor. Other restrictions— Virginia’s limit on the number of places where a minister could be licensed to speak, and Massachusetts’ requirement that all ministers have college degrees—were formally neutral among theologies and called for no doctrinal determination by civil authorities. Nevertheless these restrictions provoked intense resistance and eventually were repealed.

A fourth key issue addresses religion vs. secular conscience. Related to the question whether religious exercise should be exempted from generally applicable laws is the question whether the exercise “of religion” extends to behavior motivated by norms of secular conscience, as opposed to beliefs in God or other traditional features of religion. For example, should the exemption from school-attendance laws for the Amish in Yoder extend to followers of Henry Thoreau who rejected traditional schooling for their children? (Yoder itself answered no.)

The word “religion” might be understood in direct contrast to a broader idea of “conscience” that includes secular-based norms. Both terms were used during the Founding period—indeed, during the debates on the language of the First Amendment, which began with Madison’s proposal to protect “the full and equal rights of conscience” but eventually changed to “the free exercise of religion.” The change may have meant little substantively, because during the Founding period “conscience” was often used as synonymous with “religion.” Or possibly the change may have meant a narrowing from all deep moral convictions to theistic ones.

In a pair of cases involving challenges to military conscription during the Vietnam War, the Supreme Court read the statutory phrase “religious training and belief” to encompass objections based on any secular conscientious belief “which occupies in the life of its possessor a place parallel to that filled by the God of those” who are traditionally religious. United States v. Seeger (1965); Welsh v. United States (1970). Those expansive definitions, however, were adopted under the language of the draft-exemption statute. For the Free Exercise of Religion Clause itself, the Court has been more cautious in construing “religion” in cases like Yoder. And Hosanna-Tabor confirms that the clause gives “special solicitude to the rights of religious organizations” in distinction to secular ideological organizations.

The fifth key issue involves the Scope of “prohibiting” vs. “burdening.” Another question bedeviling courts in Free Exercise of Religion Clause cases has been just what sort of effects on religious exercise trigger protection. Are Free Exercise rights violated only when one is put in jail or fined for religious practice, or are some less serious burdens also unconstitutional?

The term “prohibiting” in the Free Exercise of Religion Clause may suggest the narrower scope of the right, covering only the affirmative imposition of sanctions such as imprisonment or a fine. Indeed, “prohibiting” might be contrasted directly with “infringing,” the term used in an earlier draft, and with its broader counterpart in other First Amendment Clauses: “no law abridging” the freedom of speech, press, assembly, or petition. Madison rejected a parallel argument during the 1798 debate over the Alien and Sedition Acts. In response to the claim that Congress could regulate freedom of the press without “abridging” it, he argued against such a semantic distinction because “the liberty of conscience and the freedom of the press were equally and completely exempted from all [congressional] authority whatever.”

In Sherbert, the Court adopted a broad understanding of unconstitutional “burdens” on religion, holding that the state violated Free Exercise by withholding unemployment benefits on the basis of the claimant’s religiously motivated refusal to work on Saturdays. Later, however, the Court took a more narrow approach, pointing to the term “prohibiting” in holding that the government did not violate Free Exercise by building a road that disrupted forest areas sacred to Native American believers, because the project did not “coerce individuals into acting contrary to their religious beliefs.” Lyng v. Northwest Indian Cemetery Protective Ass’n (1988). Sherbert, however, though now limited in its application, has never been directly overruled by the Court. The Court has never questioned Sherbert’s holding that the government can “prohibit” free exercise by withholding important benefits from the individual because of a religious practice, not only by imprisoning or fining him.