Beyond Neutrality

by Stephen L. Carter

Stephen L. Carter is professor of law at Yale University and author of The Culture of Disbelief.

This article is excerpted from God’s Name in Vain: The Wrongs and Rights of Religion in Politics. © 2000 by Stephen L. Carter Reprinted by permission of Basic Books. All rights reserved.

This article appeared in The Christian Century, October 11, 2000, pp. 996-1001. Copyright by The Christian Century Foundation; used by permission. Current articles and subscription information can be found at This material was prepared for Religion Online by Ted and Winnie Brock.


A democracy that believes in religious freedom should be willing to live with tension between the two; so should a religion that believes in democracy.

Over the years, the Supreme Court has had trouble deciding between two competing theories on how the courts should interpret the religion clause of the First Amendment. One strand, which survives in popular discourse and in the courts but is in serious danger of vanishing from the academy, is the view that the state must be neutral, neither choosing among religions nor choosing between religion and nonreligion. Unsurprisingly, this is known in the literature as neutrality. The second strand, thriving in academic discourse but remarkably difficult to express in popular discourse, is that the state must take steps to accommodate religious believers whose practices are burdened by otherwise neutral state laws. This strand, predictably, is known as accommodationism.

I confess that I have somewhat oversimplified a vast and often intricate literature. There are any number of carefully nuanced middle positions. Yet, in the end, nearly all theorists lean toward one of these two positions, simply because legal theory (unlike, say, literary theory) must finally guide the decisions by actual judges of actual cases. Consequently, legal theorists quite understandably feel impelled to tell us in the end how the cases should come out. And, in the end, just about everybody who presents a theory — no matter what the theory is called — winds up favoring something that looks an awful lot like either neutrality or accommodationism.

The problem is this: Neutrality is a theory about freedom of religion in a world that does not and cannot actually exist, whereas accommodationism, although a theory about the real world, is not really a theory about freedom of religion. Accommodationism is certainly to be preferred, at least if one takes religion seriously. But one must also be careful. Although a theory of religious freedom is plainly necessary to the proper function of the state (because its courts must decide actual cases), a theory of religious freedom may be unnecessary to, and might even prove dangerous to, religion itself.

Why do I say that neutrality is a theory about a world that does not and cannot exist? Because neutrality supposes that it is possible for the state to act without taking any account of religion generally, or of the specific religious beliefs of constituent groups. No religion is to be favored over any other; nor is religion generally to be favored over nonreligion. The metaphorical wall of separation of church and state (which is only a metaphor, although we sometimes pretend is a part of our constitutional law) seeks to capture this idea: there is the sphere of religion and the sphere of the state, and a mighty wall protecting each from the other.

But none of this is possible.

In the first place, what we are bold to call neutrality means in practice that big religions win and small religions lose. When the Supreme Court decreed in 1988 in Lyng v. Northwest Indian Cemetery Protective Association that three Indian tribes in California could not prevent the Forest Service from allowing road building and logging on their sacred lands, the justices believed, I suppose, that they were acting neutrally: the tribes had no more right than anybody else to protect the lands. But from the point of view of the tribes, whose religious tradition, as the justices admitted, would be “devastated” by the government’s action, there was nothing neutral about the destruction of the forest. With the forest gone, their religion also would be gone: a neutral result without any remotely neutral consequences.

A more powerful religion would not suffer so from neutrality. To take what might seem a painfully obvious example, nobody proposes to build a road through the Cathedral of St. John the Divine in Manhattan. My own Episcopal Church, although our numbers are dwindling, remains sufficiently powerful that nobody would even dream of so absurd a notion. (And, lest you object that the federal government “owns” the national forests, and not the cathedral, let me add that nobody would dare try to take the cathedral through eminent domain either.) To be sure, neutrality governs this result: the cathedral is not safe because it is a religious building; it is safe because it is a building valued by a politically powerful constituent group. But that only illustrates the point. Neutrality is a blueprint for the accidental destruction of religions that lack power.

The reason neutrality fails is that it imagines an impossible world. No true wall of separation is possible. Religion and the state, the two great sources of control all through human history, will never be fully separate from each other. Each will always shade into the other’s sphere. Schoolchildren learn this truth in their science classes: All containers leak. The only interesting question is how fast. In the case of religion and state, the leakage is rapid, and constant. How could matters be otherwise? Religion, by focusing the attention of the believer on the idea of transcendent truth, necessarily changes the person the believer is; which in turn changes the way the believer interacts with the world; which in turn changes political outcomes. Although there have been some clever moves in political philosophy to explain why the religious voice should not be a part of our public debates, such theories wind up describing debates from which deeply religious people are simply absent.

Besides, in a nation in which the great majority of voters describe themselves as religious, religious belief will usually be the background — even if frequently unstated — of our policy debates. A widespread religious conviction that we must aid the poor will inevitably find its way into legislation, and so the nation will create welfare programs. A widespread religious conviction that long-term help is no substitute for hard work will inevitably find its way into legislation, and so welfare will evolve into workfare.

Once one recognizes that the wall of separation leaks, and leaks badly, the case for neutrality disappears. It seeks to impose an artificial order on a naturally chaotic relationship. By so doing, it cloaks hard truths about our society. Some of the truths are pleasant — for example, the happy truth the religion matters deeply to most people, even in politics. And some of the truths are unpleasant — for example, the tragic fact that other people’s religions often matter less to people than their own. (One recalls Ambrose Bierce’s definition of a Christian: “One who believes that the New Testament is a splendid guide for the life of his neighbor.”) This truth, too, works its way into policy, which is why the “neutral” result in lying is able to stand. The state is permitted to devastate a religion, as long as it doesn’t do so on purpose.

The bureaucrats who decided to build the road and level the sacred forest were not wicked people but fallible humans, engaged in the fallible human task of governance. Yet none of them, I suspect, would have considered for a moment allowing the destruction of whatever physical symbol their own religions held to be sacred. Neutrality’s ideal world, in which all of us rigorously separate our religious and secular selves, simply is not the world in which most Americans live. Nor would many people want to. So although neutrality is certainly a theory about religion and law, it is not a theory about a world that exists.

The strongest argument in favor of the accommodation of religion is the preservation of genuine diversity — not simply people who look different, but people who in deep ways are different. A religion that makes no difference in the lives of its adherents is not a religion at all. The accommodationist believes in religion as something that actually changes the way people are; nurturing religion, then, also nurtures a plurality of communities, communities that assign to existence meanings different from those of the dominant culture. The accommodationist, therefore, argues that the state should back off whenever possible — formally, whenever a compelling state interest is not served by the effort to rein in religions that would otherwise press for meanings of reality that the state abhors.

These are already dangerous words, because they can invite oppression: the state, after all, might reasonably conclude that many different ends are compelling. Indeed, the immediate difficulty for the accommodationist is that it is the state that is doing the accommodating. The problem facing the religionist is deeper than the struggle between the ideal of neutrality and the reality of accommodation. For the accommodationist, at the very beginning of the analysis, is faced with two crucial problems that necessarily limit religious freedom. First, the accommodationist must define religion, which already narrows the universe of what counts. One saw evidence of this, for example, when the Massachusetts Supreme Judicial Court decreed that when religious counseling is informed even in part by secular psychology, it ceases to be religious and is entitled to no free-exercise protection. The problem is plain: If the religion seeking an accommodation must first prove to the state (through its courts) that it is a religion, it is already under pressure to meet a test that might have nothing to do with the religion’s teaching.

The courts might try to avoid this difficulty, by adopting, for example, Kent Greenawalt’s proposal that we define religion analogically. Greenawalt suggests that we begin with what we know to be religions, search out their common elements, and then compare other claimants by looking for similar (but perhaps not identical) elements. This approach (which is, I think, more impressionistic than analogical) would have the advantage of not forcing all religions into a single, narrow mold — a point to which I shall return. But it would have the disadvantage of beginning with the religions with which our culture is most familiar, which at once makes potential outsiders of religions that may look very different. Besides, like all definitional approaches, this one still says to people of faith that they must ensure that they come within the approved definition if they want, in a legal sense, to be free.

Second, the accommodationist must indicate a limit beyond which no claim of religious freedom will be recognized — to resolve, for example, the problem of religiously mandated murder. Most accommodationists place the limit at “compelling state interest”; but even setting compellingness as the standard, and handling it correctly, the courts in the end will be centering their concern on the needs of the state, not the needs of the religionist. This became painfully clear in 1996 when the Supreme Court refused to hear an appeal of the Alaska Supreme Court’s Swanner decision, which held that the state’s interest in preventing discrimination against unmarried heterosexual couples is sufficiently great that it trumps the objections of landlords who believe they are forbidden by God to permit “fornication” on their property. It is not easy to discern what made the interest compelling, except that the state had decreed it to be so.

The simple problem is that the accommodationist must ultimately draw some lines; and every line divides those religions that will be privileged from those that will be demoted. The mere existence of these dividing lines creates pressures on religions to twist themselves into shapes that the state (through its courts) will recognize. No religion picks a fight with the state if it can be avoided, and few religionists are able to resist the lure of state assistance. And so, precisely because the rules of constitutional law create some faint possibility of gaining special consideration, the question the religion must ask itself becomes (to take Needham as our example) not What form of counseling does God require? but What form of counseling will the state allow? Thus does constitutional protection of religion threaten to undo the specialness of religion itself.

Suppose the state says no? Is the follower of R then to decide that B is not required after all — because the state will punish him if he does it? In principle, this outcome is no more appealing than the notion that the state should decide to allow the follower of R to do B because he is going to do it anyway, even if punishment follows. But no such symmetry exists. We seem to expect that the sensible religionist will take into account state response in making a decision whether to break the law or not; yet we evidently suppose it to be absurd that the sensible state should take into account religious response in deciding whether to enforce a law or not. The religionist who persists in doing B after it is deemed to carry no constitutional protection is a fanatic, but the state that puts him in jail is only doing its job. As a formal matter, this result is possible because the state possesses the legal authority to enforce its decrees and the religionist does not. As a practical matter, the result is possible because the state, not the religionist, has the guns.

In a nation that properly prides itself on religious pluralism, we certainly do not want to put the guns in the hands of the competing religions instead. Yet the same pride should help us to recognize that the asymmetry of religious freedom is mischievous. Although the compelling-ness test (which only accommodationists believe in anyway) gives us a standard for determining which state interests are so vital that a claim of religious requirement cannot overcome them, we have no test for determining which religious interests are so vital that no state interest — not even a compelling one — can overcome them. In other words, the balance is struck in a way that creates only two sets of cases: those that are hard and those in which the state automatically wins. There may be no set of cases in which the religion automatically wins.

This asymmetry is also unstable. The neat separation assumed by the model does not survive. The result is not chaos; the result, rather, is a steady leakage in our separation cylinder, but leakage in only one direction, so that the state wins more and more and the religionist wins less and less — precisely the trend line of the cases over the past 20 years. This does not mean that there is a point at which accommodation and neutrality converge (although there may be); it means only that as time and cases go by, the state-centeredness of what we are bold to call religious freedom is more and more apparent, and the pressure on religionists to conform to the expectations of others grows and grows, with the inevitable complicity of the judges who supposedly protect religion’s “right” to be free.

But the phenomenon I have been describing, the reader might object, is simply the way the world works. For me to accuse the courts of being state-centered on religion issues, the reader might argue, smacks of sophistry. After all, the judges must decide the cases properly before them. To decide cases they must draw lines. That the lines judges draw are occasionally a little bit arbitrary is a very old critique of judging. Unless we are to adopt a standard under which the religious freedom claimant always wins, then we must pick a point, by some method, at which the claimant loses. Whatever that point turns out to be, it will invariably be the point at which the losing claimant will accuse the judges of statism.

And besides (so the counterargument might continue) — besides, trade-offs are a necessary part of the life of a society. Of course religions will sometimes feel themselves pressured to change; everybody is pressured to conform; that is a fact of life, not of law. If a religious belief is sufficiently strong, it will survive the pressure. Recall that Bob Jones University refused to yield to state pressure to abandon its religious belief in racial separation after losing in the Supreme Court in 1983. Imagine, however, that the Internal Revenue Service had been after Bob Jones to abandon faith in Jesus Christ as Son of God and Savior, the bedrock on which Christianity rests. In that case, nobody would have been surprised by the school’s refusal to back down.

Neither one of these objections disturbs the theory. Yes, judges must decide cases and draw lines; but it is the fact that the judges must draw lines — not the place where the lines happen to be drawn — that creates the risk for religions that go to court to seek protection. And, yes, religionists should ideally be strong enough to resist state pressure; but the state itself has created layer on layer of purportedly neutral pressures that combine to weaken the ability of the religions (or, rather, the religious people) to resist. Among these layers are a variety of tax rules, which encourage religions to reshape themselves so as to be eligible for tax benefits, and the recent legislative efforts, certainly constitutional but perhaps of dubious value to religion, to allow religious groups to share in the rather substantial largesse of the programmatic side of the welfare state.

A principle of accommodation will not eliminate the tension that pressures the religious to change for the sake of society, but it will at least create a space in which the religions can organize their resistance. Moreover, if combined with a healthy respect for the role of the religious voice in politics, the accommodation principle might at least create some pressures for the society to change for the sake of religion. A democracy that believes in religious freedom should be willing to live with the tension; so should a religion that believes in democracy.