Why do we protect freedom of religion? The commonsense answer, which I think hits close to the truth, is that we protect it because religion is important. That simple answer creates serious problems for liberal theory, however, so it is seldom discussed or defended by legal writers.

Some say freedom of religion is important because it is one way (though only one) of exercising our autonomy as human beings interested in making our own choices and shaping our own lives. Religious devotees create lives for themselves around certain kinds of beliefs and values. They will probably join a community of like-minded people (a church). They typically have ideas about their relationship to God that orient them in their daily life. And so on. In doing these things they are protected by what Laurence Tribe calls “rights of religious autonomy.”

There is nothing unique about religious autonomy. It is a name for one set of choices people make about how to live, but there are other sets of choices within the field of autonomy: choices about reproduction, risk taking, vocation, travel, education, appearance, and sexual behavior. For Tribe, religious autonomy is just one aspect of the larger “rights of privacy and personhood.”

Moreover, within the set of religious choices we attach value to the act of choosing, not to particular outcomes. A decision to reject God is entitled to the same protection as a decision to follow him. As Gail Merel has put it, “Individual choice in matters of religion should remain free: individual decisions are to be protected whether they operate for or against the validity of any or all religious views . . . . The individual is freed from . . . the oppressive effects of government regulation in order to believe or disbelieve as he chooses.”

The Supreme Court has given some support to the idea that autonomy is the value underlying religious freedom. It held in Torcaso v. Watkins (1961) that Maryland had violated religious freedom by requiring state officeholders to declare their belief in God. This suggests that the Constitution attaches equal value to belief and disbelief: the important thing is the choice, not the outcome.

This conclusion is hard to square with the language of the First Amendment, which protects only the free exercise “of religion.” Rejecting religion is an exercise of freedom, but it is not an exercise of religion. (Amputation is not a way of exercising my foot.) The free exercise clause by its terms seems inconsistent with the idea of autonomy. It seems to favor choices for religion over choices against religion.

One way to avoid this textual limitation is to define “religion” very broadly”so broadly that even disbelief is a kind of religion. This is what the Court did in interpreting the draft law ( U.S. v. Seeger, 1965). When I was a boy people were exempt from military service if their “religious training and belief” made them oppose war. Federal law defined religious belief as “an individual’s belief in a relation to a Supreme Being involving duties superior to those arising from any human relation.” The Supreme Court interpreted the law with an eye on the free exercise clause, and said that the question was whether “the claimed belief occup[ies] the same place in the life of the objector as an orthodox belief in God holds in the life of one clearly qualified for exemption.” The idea of “God” is “more of a hindrance than a help.” We should think of God “not as a projection ‘out there’ or beyond the skies but as the ground of our very being.” And “religion” is nothing more than “the devotion of man to the highest ideal that he can conceive.”

The autonomy theory views religious freedom from an agnostic standpoint. That is hardly surprising. It follows rather naturally from the “autonomous” view of human nature. If you scratch a person deep enough, the theory holds, you will find a kind of free-floating self. If you looked at the surface of my life you might say that I was a middle-class Irish Catholic, husband, father of five children, law professor, part-time musician, Celtics fan, and so on. I have naturally inherited a variety of moral convictions (those typical of bourgeois Catholics, or lawyers). I am also moved by various desires that arise from and act upon the details of my life (I want prestigious publishers for my books, money for my children’s education, time with my wife).

But my essential self is able to rise above these details. It is unencumbered, unsituated. It can step back from my habitual convictions and desires (my first-order preferences), reflect critically on them, and change them to suit its own plan (second-order preferences) for what my life should be like. Exactly where I get my second-order preferences is a matter of some dispute. Some say that I am guided by reason to universally applicable principles. Others say that I just make them up. But everyone agrees that it’s up to me”to my unencumbered self”to choose them, however I might find them.

This view of human nature is the basis for a powerful argument in favor of freedom. A just political order has to take account of the way people really are. It must, in other words, respect their freedom to act as unencumbered selves on their second-order preferences. In the case of religion this means that it must view them as persons choosing, from a detached position, a theological orientation. This is what I mean by saying that the autonomy theory assumes the agnostic viewpoint. I might be a Catholic in my daily life, but that is a first-order preference. The real me is able to step back from it, assume an agnostic stance, and make a fresh start. I might then renew my religious commitment; but I might reject it. It doesn’t really matter. The important thing is that the real me should organize my life along lines it freely chooses. The law protects religious freedom in order to facilitate that choice.

The autonomy theory is in one sense too powerful. It holds that a just society must let its citizens choose how to live their own lives. Some of the relevant choices are religious, so it follows that the government must not interfere with them. But there is nothing special about religious choices in this argument. They are on a par with promiscuous sex, cigarette smoking, and the practice of optometry. Our instincts and the language of our Constitution tell us, though, that there is a difference. The Bill of Rights protects the free exercise of religion. It says nothing about free love, free trade, or excise taxes on tobacco. What we need is an argument that is capable of protecting religion without protecting a lot of other activities that we don’t feel strongly about.

Having said this much, I will not pursue the point further. I want to turn instead to a second problem with the autonomy theory. It concerns its assumptions about human nature. One is the factual assumption that we are capable of stepping back from our convictions and desires and reorganizing them according to second-order preferences that we freely choose. Another is the more value-laden assumption that we should do this in order to live “authentic” lives. There are those who would dispute both assumptions, and people who want religious freedom are among those most likely to do so.

Consider first the factual assumption. It is inconsistent in several ways with recurring ideas in Christian theology. The notion of original sin is meant to suggest the inherent imperfection of human nature. In the strongest statements of this idea”Augustine’s is a good example”it entails our inability to master sinful desires and to freely will doing good. In the common phrase, human nature is the slave of sin. The counterpoint to this unhappy view of human nature is the idea of grace. It is a kind of sharing in divine life, a power that enables us to control sinful desire, live good lives, and win salvation. But grace is given to us by God gratuitously. We can’t call it down with a rain dance, and we can’t behave as we should without it. It is out of our control. This aspect of grace, followed to its logical conclusion, leads to the Calvinist notion of predestination: our salvation is entirely in God’s hands, and some are not saved.

This view of human nature affects the way many religious people look at the idea of choice. The individual does not have complete control over choosing the religious option. It is God who makes the choice. I might have to accept God’s choice and cooperate in carrying it out, but I am cast as a supporting actor. Thus the Jews understand themselves as the chosen people. Their stories tell of people pursued by God and brought back to do his work. Jonah, called by God to be his prophet, tried to escape on a boat for Tarshish but was brought back by miraculous means. In the New Testament, Jesus himself set the example by praying before his death, “Father, if it be thy will, take this cup from me. Yet not my will but thine be done.” God converted the apostle Paul by striking him to the ground and blinding him. The scholar Alan Simpson has argued that a similar experience of conversion is the essence of Puritanism.

Those who take this view of human nature will also disagree with the autonomy theory about the value of running our own lives according to our second-order preferences. That is not the basis of real freedom. Augustine claims, for example, that real freedom is freedom from the bondage of sin. And it is out of the question for free will to realize this freedom through its own power; this it can do only through the grace of God. It sounds paradoxical, but it is accurate to say that Christian freedom consists not in making our own choices but in obeying the law of God.

The autonomy theory, then, bases religious freedom on a view of human nature that many religious people would reject. This need not be a fatal defect. We also justify freedom of speech on grounds that some speakers would reject”we say that it promotes democracy, and yet we grant it to Nazis who don’t believe in democracy. But religious believers play a crucial role in free exercise law. They are not like the Nazis. They are like the New England town meeting”the paradigm around which the theory is built. If the theory does not work for them there is probably something wrong with it.

There is an additional doctrinal point. Free exercise law has a tendency to assume a kind of split-level character. In addition to its other shortcomings, the autonomy theory fails to explain this tendency.

In some areas of the law believers and unbelievers get equal protection. This is how it is with compelled worship and belief. Agnostics as well as Quakers can object to a test oath. But there are other areas where the law protects only believers. The free exercise clause sometimes requires states to pay money to religious believers”as in paying unemployment compensation to those whose religion requires them to abstain from work on certain days. It does not require payment to agnostics. It excuses Amish children from public school, but not agnostics. It protects the internal affairs of churches but not other associations against government interference.

In recent years the Court has shown an inclination to even out these differences. It has held that the Constitution does not, as a general rule, require special treatment for religiously required behavior. But this principle is still subject to a number of qualifications. First and most obvious is the fact that the free exercise clause still forbids discrimination against religion. That in itself is a kind of special treatment. There is no comparable rule protecting nonreligious action. The army can’t have a special rule against yarmulkes. But it can have a role preferring yarmulkes to other nonuniform garb. Second, the Court has left standing all the old cases requiring religious exemptions (the ones about government benefits, school attendance, church affairs, and so on). It has even enlarged upon some of these.

These different levels of protection suggest that there are several principles at work in the law of religious freedom, not just one. The autonomy theory is appealing in its simplicity. But it is too simple to explain the actual complexity of the law.

The second standard argument for freedom of religion is political rather than ethical. The argument is that the denial of freedom causes strife that leaves everyone worse off. We can find both comparative and historical evidence for this conclusion. Lebanon, Iran, India, and the Sudan have recently seen violent struggles for religious supremacy. England and much of Western Europe did so in the sixteenth and seventeenth centuries. The Supreme Court has found parallels in early American history. As Justice Felix Frankfurter argued in McGowan v. Maryland (1961), “In assuring the free exercise of religion, the Framers of the First Amendment were sensitive to the then recent history of those persecutions and impositions of civil disability with which sectarian majorities in virtually all of the Colonies had visited deviation in the matter of conscience.”

Like the autonomy theory, the political theory tries to justify religious freedom from an agnostic viewpoint. It stresses two kinds of harm that affect unbelievers. One is civil war. Even noncombatants get killed in a civil war, and everyone suffers from the collapse of the government and the economy. The other harm is persecution. Unbelievers cannot be prevented from practicing their faith. (They have none.) But if the government wants to compel a particular form of religious observance it might have to “torture, maim, and kill . . . ‘atheists’ or ‘agnostics’“ along with nonconforming believers ( Zorach v. Clauson , 1952).

This theory, like the autonomy theory, makes freedom universally available. But here the value of freedom is instrumental, not intrinsic. It leads to peace. If it didn’t, we would take another approach. The autonomy argument, by contrast, said that freedom was intrinsically good for people like us. Of course it had to make some controversial assumptions about what kind of people we were. It is a virtue of the political argument that it dispenses with those assumptions. And there is much else to recommend it. It is realistic and practical, and goes some way toward justifying a special place for religious freedom. It does, though, have some weaknesses. The most important one is that it is incomplete.

Consider first the case of fringe groups. In American society there are, depending on how you count, hundreds or thousands of them. They include small but well-known sects (Hare Krishnas, members of Reverend Moon’s Unification Church) and smaller, little-known local cults. For my purposes they also include unchurched believers”religious individualists who seek God in their own way. The political defense of freedom gives no protection to these people. If a group is sufficiently small the government can simply stamp it out without running the risk of civil war. Of course civil war is only one kind of strife. Stamping out fringe groups is persecution, and the political argument is designed to avoid that too. But what exactly is wrong with fining, jailing, medicating, religious eccentrics? These forms of punishment and cure are not in themselves objectionable, the way cruel and unusual punishment would be. We routinely apply them to drug offenders and think that we’re doing the right thing. The obvious answer is that there is a difference between religious activity and drug dealing: one is good and the other is bad. But that argument, which I will develop further below, takes us beyond our concern with political strife.

So the political strife argument doesn’t protect groups who can’t fight because they’re too small. Neither does it protect groups (some of them large) who are unwilling to fight. The Amish on principle flee from controversy and eschew politics. Quakers are well known for their pacifism. Groups that are far larger engage in many practices that they see as desirable but not essential, and that they would not defend with violence. Consider the employment practices of Catholic schools. The political defense of freedom gives no shelter to these groups if they pose no threat to peace.

I think we can state these objections in an even more general form. The political explanation tells us that freedom is good because it brings peace. It does not tell us why we should prefer freedom to other means of bringing peace. It gives us no reason to object to the suppression or the establishment of religion, provided the job is done ruthlessly enough to prevent civil war. Religion was an insignificant cause of strife in the Soviet Union from Stalin’s time until very recently. There was little freedom, but that is no objection if all that matters is peace. The obvious advantage of freedom is that it respects piety as well as peace. But we need an argument that will tell us why it is good to respect piety.

The best reasons for protecting religious freedom rest on the assumption that religion is a good thing. Our Constitution guarantees religious freedom because religious people want to practice their faith. As Mark DeWolfe Howe said in The Garden and the Wilderness (1965): “Though it would be possible . . . that men who were deeply skeptical in religious matters should demand a constitutional prohibition against abridgments of religious liberty, surely it is more probable that the demand should come from those who themselves were believers.”

One form of distinctively religious action is the performance of ritual acts. These include prayer and other kinds of worship; compliance with sumptuary rules governing dress, diet, the use of property; the observance of sacred times (feasts and holy days) and places (pilgrimages to shrines); rites connected with important events in the believer’s life (birth, death, maturity, marriage); and so on.

Acts like these make sense only in the context of an entire religious tradition. Acts of worship presuppose a belief in a supreme being. In Christianity, Judaism, and Islam, God is described as the Creator, Lord, and Judge of the world we live in. The believer thinks that acts of worship are good because they please God or harmonize with the order of nature. Other forms of prayer presuppose a belief in a transcendent reality”a kind of life outside our world that is more real than our own, and that affects us in miraculous and mundane ways. We may go on living there (in heaven or hell) when we die. Prayer is the means by which we communicate with and try to influence this reality. Still other rites presuppose the existence of a social organization that enables like-minded people to act together. Some are led by a specialized class of functionaries who teach, supervise, and minister to ordinary members.

There is in our traditions a religious argument for religious freedom that is peculiarly associated with ritual acts. It is, simply, that it is futile to coerce people to perform ceremonies (prayer, worship, declarations of belief) they don’t believe in. This idea has ancient roots, but it was most fully developed by English Protestants during the seventeenth century. Locke appeals to it in his Letter Concerning Toleration : “True and saving religion consists in the inward persuasion of the mind, without which nothing can be acceptable to God. And such is the nature of the understanding, that it cannot be compelled to the belief of anything by outward force.” Coercion can be worse than futile-it can be counterproductive. In Milton’s phrase, to force a ritual performance is “to compel hypocrisy, not to advance religion.”

In modern free speech law we often run across the idea that the mind is a private domain which the government should not, and as a practical matter cannot, enter. Locke’s and Milton’s claims are different. They rest on a religious idea about our relations with God. Coerced ritual is futile because it cannot put the soul in touch with God. The individual cannot hear God unless he has faith. And faith does not come to people just because they go through the ritual motions. God gives it to whom he wills. It is an idea characteristic of Protestantism that this happens in a very individual way. The most effective medium is Scripture, through which God may speak to the pious reader.

This distinctively Protestant “right of private judgment” began as a protest against the Catholic Church’s claim to mediate between God and individual souls, but it served equally well as an objection against state mediation. Roger Williams underlines the connection in The Bloody Tenent : “In vain have English Parliaments permitted English Bibles in the poorest English houses, and the simplest man or woman to search the Scriptures, if yet against their souls’ persuasion from the Scripture, they should be forced (as if they lived in Spain or Rome itself without the sight of a Bible) to believe as the Church believes.”

Let me turn now to a second form of religious action, and a different argument for religious freedom. Members of a religious tradition typically want to acquire and spread knowledge about the esoterica of their beliefs, ritual forms, ceremonial duties, and so on. These special kinds of religious truth are often set down in sacred texts (Bible, Torah, Qur’an) and elaborated upon in written and oral commentaries (Talmud, Sunna). Believers like to study these texts and commentaries, to discuss them with others, and in some traditions to bring them to the attention of unbelievers.

Those who feel this way sometimes argue that the freedom to acquire and spread religious knowledge leads us to the truth. We inherit this idea, like the last, from seventeenth”century English Protestantism. It is the message of Milton’s Areopagitica ”an expression of Puritan faith published in the same year as Roger Williams’ Bloody Tenent . Milton offered several reasons why unlicensed printing would promote the discovery of religious truth. One was the now familiar claim that truth will prevail over falsehood in any free encounter. A less familiar but more radical idea was that God’s revelation is progressive. This makes free inquiry not only safe but actually desirable. Individual thinkers may wander astray, but the net social effect of freedom is to bring us closer to God. “To be still searching what we know not by what we know, still closing up truth to truth as we find it . . . this is the golden rule in Theology as well as in Arithmetic, and makes up the best harmony in a Church.”

Let me reemphasize that these two arguments (futility, truth) rest on religious premises (faith is a gift; revelation is progressive). They will convince only religious believers. But within that group they have carried the day. Consider the current positions of the Catholic Church and the Presbyterian Church”the two chief targets of Puritan polemicists like Milton and Williams.

The Catholic Church’s Declaration on Religious Freedom was promulgated during the Second Vatican Council in 1965. The Declaration offers several reasons for protecting religious freedom. Prominent among them is the idea that faith arises through internal communication between God and the individual: “For of its very nature, the exercise of religion consists before all else in those internal, voluntary, and free acts whereby man sets the course of his life directly toward God. No merely human power can either command or prohibit acts of this kind.” Equally prominent is the notion that freedom assists the search for religious truth: Truth “is to be sought after in a manner proper to the dignity of the human person and his social nature. The inquiry is to be free, carried on with the aid of teaching or instruction, communication, and dialogue. In the course of these, men explain to one another the truth they have discovered, or think they have discovered, in order thus to assist one another in the quest for truth.”

The 200th General Assembly of the Presbyterian Church (U.S.A.) adopted a Policy Statement on religious liberty in 1988. It too offers a variety of reasons for securing freedom. One is that faith cannot be coerced: “Religious insight and faith come from God, who exists over and beyond the powers and principalities of earth. Such insight and faith are recognized and received; they can neither be commanded or controlled by civil authority, military power, or religious piety.” The other is that control of religion inevitably leads to suppression of the truth, because “there are no foolproof human mechanisms by which to test the authenticity of insight claimed to be from God.”

Let me turn now to a third variety of religious action, and a different argument for religious freedom. Religious believers are often bound by special moral obligations. These come from a moral code that has some supernatural sanction (the halakhah in Judaism, the shari’a in Islam). Such a code often demands forms of behavior that the rest of society views as supererogratory, morally neutral, or even (occasionally) wrong. A violation of the moral code may be seen as something worse than a breach of dutya kind of personal harm or insult to the author of the code, which calls for repentance and might be punished or forgiven on a transcendent level.

About these kinds of actions we might say that the government should not force people to violate moral duties if (in their system of belief) they will face transcendent consequences. Otherwise the believer might have to choose between violating the law and risking damnation. This is how it was with the early Mormons who were convicted of practicing polygamy. Or the believer might be forced to forgo a great good. In one of the Supreme Court’s recent cases an American Indian complained that the government’s use of a social security number for his daughter would “rob [her] spirit.”

Of course the government often causes great harm to unbelievers as well. A religious pacifist fears for his salvation when he is drafted, but the average Marine also suffers at the thought of leaving his family and going into combat. From a religious point of view, though, the cases are not comparable. The harm threatening the believer is more serious (loss of heavenly comforts, not domestic ones) and more lasting (eternal, not temporary). That is what justifies this special kind of freedom to religious claimants alone.

This is a consequentialist argument for freedom (though the consequences it relies on are religious). But we could also make a nonconsequentialist argument. Moral codes impose religious duties, and there is something uniquely wrong with forcing people to violate a religious duty even if they are not primarily concerned about final rewards and punishments. Strict Calvinists, for example, see no connection between the performance of religious duties and election to heaven. But they can still demand religious freedom. The focus of their claim is not their own destiny. They are concerned instead with the effect on God, as it were”they have to disappoint him in order to comply with the law. The individual places great value in keeping faith with such duties, and it is this value that religious liberty protects.

These arguments about suffering and duty differ from the earlier arguments about futility and truth. Claims about suffering and duty focus on the personal interests of religious believers. They are an appeal to rights in the modern sense”a form of protection for people who are losers in the political process. Claims about the futility of coercion and the discovery of truth focus on a larger social interest. “We will all be better off,” the believer says, “if we allow religious freedom.”

Along with this difference in focus goes a difference in coverage. The earlier arguments apply universally. Coercion is futile no less for atheists than for Catholics and Jews. God may give them faith or he may not, but the government can’t help him out. So too with the discovery of truth. It’s no use letting only right-thinking Christians search, because we’re talking about revelation and God can reveal himself to anyone. As the Gospel says, “The spirit blows where it wills.” The arguments about suffering and duty, by contrast, offer protection only to religious believers. The sufferings of the faithful are special precisely because they believe in heaven, hell, eternal life, and so on. The believers’ duties are more compelling just because they arise from God’s commands.

This explains what I have called the split-level character of free exercise law. In some areas the clause protects everyone. This happens when we are dealing with ritual acts and the pursuit of knowledge. Atheists and Quakers alike can object to laws prescribing forms of faith (test oaths) and worship (school prayers). Anyone can object to a law that forbids inquiry (the teaching of evolution) for religious reasons. Likewise anybody can object”on free exercise as well as free speech grounds”when the government tries to limit communication about religiously significant questions. These matters are all covered by the first set of principles: compelled belief is futile; revelation is progressive.

In other areas the free exercise clause protects only religious believers. The cases where this happens are cases about compliance with a moral code. The believer’s faith might require him to leave his job, or school, or the army. The Supreme Court used to give serious consideration to all such claims. Today it is harder to get special treatment, but it is not impossible. And it remains true now, as before, that “to have the protection of the Religion Clauses, the claims must be rooted in religious belief” ( Wisconsin v. Yoder , 1972). This disparity is explained by the second set of principles: believers face a special kind of suffering; they are subject to a higher kind of duty.

The draft cases do not fit this picture. None of the reasons I have given seems to cover the conscientious objection of nonreligious young men to service in the armed forces. But we should not generalize from these cases, any more than we should make death penalty cases the pattern for rules of criminal procedure. Killing is an extreme act, and the feeling of dread that attends it can give even nonreligious duties an absolute cast.

I have argued that we should take the believer’s viewpoint rather than the agnostic’s viewpoint in thinking about religious freedom. But my argument seems incomplete. It relies upon reasons that only some people find convincing. And sometimes it protects freedom only for those who are convinced. How can such a lopsided idea justify one of our basic constitutional rights?

I admit that my argument for religious freedom is lopsided, but I want to stress that this is not as serious a problem as it might appear. This is so, first of all, because it does not require agnostics to give up something for nothing. Free exercise law has a split-level character. On one level it gives special protection to religious believers. But on another level it treats everyone alike. The government can’t force anyone to perform ritual acts, and it can’t interfere with the pursuit of religious knowledge. This means that everyone has a reason to support some degree of religious freedom.

The standard arguments assume the agnostic point of view because they want a rule of religious freedom that is fair to everyone. Fairness here has two dimensions. One is consent. Universal consent is a good indication that a rule treats everyone fairly. (That is why social contracts are always adopted unanimously.) The agnostic point of view tries to base freedom on principles everyone can agree with. Revelation is out because it’s hidden from some people. We are asked to look instead at facts about human nature and our social situation: the autonomy argument refers to the unencumbered self; the political strife argument refers to the causes of war and peace.

The other dimension of fairness is reciprocity. A contract is fair in this sense if the parties share equally in the benefits of the bargain. The autonomy argument satisfies this condition by making religion just one of many protected choices, and by offering equal religious freedom to believers and unbelievers. The political argument says that freedom results in a public good (peace) that everyone enjoys.

The religious defense of free exercise is lopsided because it violates these conditions of fairness. The principles that it relies on to justify freedom-futility, truth, suffering, and duty-all refer in some way to religious beliefs that many people do not hold. This makes it hard for some people to consent. The religious defense also gives special protection to some kinds of religious action. That is, it excuses religious actors from some generally applicable laws when their moral code requires some other course of conduct. This violates the condition of reciprocity. Why should we prefer an argument of this kind to arguments that seem to satisfy the canons of fairness?

For several reasons. First, the standard arguments themselves fail the test of fairness. The autonomy theory, like my own theory, appeals to assumptions about human nature (the unencumbered self, the value of authenticity) that are inconsistent with convictions that many religious people hold about original sin, grace, faith, and revelation. In the real world these people would not consent to a social contract based on autonomy. To avoid this problem the theory asks everyone to assume the agnostic viewpoint. But why not ask agnostics to assume the religious viewpoint? That too would produce unanimous consent. It would not be consistent with liberal theory, because it commits us to a view of the good before we have resolved the issue of rights. But we can’t assume the correctness of liberal theory. That is the very question we’re debating. If my theory won’t get universal consent, neither will the autonomy theory.

The political strife theory is flawed in the same way. It asks us to make the empirical assumption that we can only have civil peace through religious freedom. But there are other ways of avoiding strife; repression is one of them. Unless freedom has some other good points, there is no reason to prefer it over repression.

There is reason to doubt, then, that the standard arguments are more fair than mine. My own approach, on the other hand, has some real strengths that they lack. It is the most convincing explanation for why our society adopted the right to religious freedom in the first place. It is possible to imagine a society of skeptics insisting on a free exercise clause, but the idea is far-fetched.

The religious justification is also the reason why many, perhaps most, religious believers claim the right to freedom today. It enables them to perform their religious duties and to avoid religious sanctions. It allows them to pursue the truth, as God gives them to know the truth. And no other course could bring them closer to God.

Finally, the religious justification is the only convincing explanation for the split-level character of free exercise law. Sometimes religious believers and nonbelievers are treated alike; but sometimes the law protects only religious believers. This is not something that we can explain by appeals to consent and fairness. It violates the canon of reciprocity. The only convincing explanation for such a rule is that the law thinks religion is a good thing.



John H. Garvey is Professor of Law at the University of Notre Dame Law School. This article is excerpted from his book What Are Freedoms For? , just published by Harvard University Press.