1. Rawlsian Objections to Special Treatment

Accommodation of religious objectors is the most familiar instance of American law's special treatment of religion. Quakers’ and Mennonites’ objections to participation in war have been accommodated since Colonial times. Sacramental wine was permitted during Prohibition. Today the Catholic Church is exempted from antidiscrimination laws when it denies ordination to women. Jewish and Muslim prisoners are entitled to Kosher or halal food.

The prohibition of establishment of religion likewise gives religion a privileged status. A central justification for disestablishment, since Colonial times, has been that religion can be corrupted and degraded by state support.Footnote ⁴ When the Court invalidated school prayers, it declared (quoting James Madison) that “religion is too personal, too sacred, too holy, to permit its ‘unhallowed perversion’ by a civil magistrate.”Footnote ⁵

The object of this special treatment, “religion,” is vague. It is not confined to theistic religion. When an agnostic with a conscientious objection to war claimed a draft exemption, the Supreme Court deemed his claim to be “religious” in the pertinent sense: the law “exempts from military service all those whose consciences, spurred by deeply held moral, ethical, or religious beliefs, would give them no rest or peace if they allowed themselves to become a part of an instrument of war.”Footnote ⁶ A review of the cases by the US Court of Appeals for the Seventh Circuit concluded that “the Supreme Court has recognized atheism as equivalent to a ‘religion’ on numerous occasions.”Footnote ⁷

These results are facilitated by the undeniable fact that the boundaries of the category of “religion” are so fuzzy.Footnote ⁸ Those boundaries are not, however, infinitely elastic. Brian Leiter points out that a Sikh will have a colorable claim to be allowed to carry a ceremonial dagger, while someone whose family traditions value the practice will be summarily rejected.Footnote ⁹ A claim's status as “religious” will entitle it to special treatment in American law.

Are these practices of privileging religion acceptable within the terms of Rawls's theory? Two prominent arguments point to a negative answer.

Rawls argued that the basic structure of society would be fair if its terms were those that would be agreed to in a hypothetical “original position,” in which a “veil of ignorance” prevents any of the parties from knowing such morally irrelevant facts as their position in society. Pertinently here, in the original position,

Rawls thus attempts to articulate, at a level abstract enough to make sense to those behind the veil of ignorance, the demand for religious liberty. The parties are evidently aware of the general phenomenon of religious persecution, and want to protect themselves from it. Like many other theorists, Rawls recharacterizes the object of protection as conscience. The structure of his argument is, however, not conscience specific. In this context, “conscience” is a misleading term.

Rawls thinks that, given that the parties know the general facts about human psychology,

The passage just quoted is the basis of Leiter's claim that Rawls's philosophy provides no basis for giving religion special treatment. Rawls's argument for liberty of conscience, Leiter thinks, does not support any special treatment for religion as such:

The objection to special treatment is simple and elegant. Whatever reasons the parties have for safeguarding liberty of conscience cannot involve “anything specific to religion.”Footnote ¹³ It follows that “the Rawlsian perspective cannot help us evaluate the principled case for toleration of religion qua religion.”Footnote ¹⁴

A second Rawlsian objection to special treatment for religion is based on the constraints of public reason.

When we exercise political power over others, Rawls argues, we should seek “to be able to justify our actions to others on grounds they could not reasonably reject.” In justifying the use of political power we are “to appeal only to those presently accepted general beliefs and forms of reasoning found in common sense, and the methods and conclusions of science when these are not controversial.”Footnote ¹⁵ These are not the same thing—a belief could be presently accepted and widely uncontroversial without being of a character that could not reasonably be rejected—but both formulations suggest a problem with the claim that religion deserves special treatment. Some citizens do not think that religion is at all good. They deplore it and think that we would be better off without it. That is one reasonable view among many, and it is not ruled out by common sense or the methods of science. So perhaps special treatment of religion is ruled out by public reason.Footnote ¹⁶

2. Exigency and Opacity

If Rawls's arguments that support liberty of conscience are tested to determine what weight they will bear, we will find that the scope of the liberty is radically underspecified in the original position.

The term “conscience,” Thomas Hill has shown, is a general concept that is capable of being specified in various particular conceptions. The general concept is “the idea of a capacity, commonly attributed to most human beings, to sense or immediately discern that what he or she has done, or is about to do (or not do) is wrong, bad, and worthy of disapproval.”Footnote ¹⁷ Rawls writes of “obligations,” and seems to have in mind conscientious beliefs about moral obligations. If this is right, the parties in the original position are giving too much weight to the will to be moral. Rawls may be relying on the dubious assumption that the will to be moral trumps all our other projects and commitments when these conflict, and that no other exigency has comparable weight.Footnote ¹⁸ If he is not, then it is unclear what “freedom of conscience” refers to.

In American law, religion is protected even when it does not fall under the description of conscience. The most recent federal religious liberty statute, the Religious Land Use and Institutionalized Persons Act of 2000, declares that “the term ‘religious exercise’ [which is protected] includes any exercise of religion, whether or not compelled by, or central to, a system of religious belief.”Footnote ¹⁹ Rawls does not say enough about “liberty of conscience” for us to tell whether he agrees or disagrees.

What the parties should really care about is the prospect of being denied some liberty that is urgently important to them. The liberty to follow one's conscience is one such liberty. There are others. What is protected is, however, something less than liberty in general; “no priority is assigned to liberty as such.”Footnote ²⁰

What is salient, for the parties, about moral and religious obligations is their exigency. They do not know their substantive religious, philosophical, or moral views, but they “do know the general structure of rational persons’ plans of life (given the general facts about human psychology and the workings of social institutions) and hence the main elements in a conception of the good.”Footnote ²¹ That knowledge leads them to specify “forms of belief and conduct the protection of which we cannot properly abandon or be persuaded to jeopardize for the kinds of considerations covered by the second principle of justice [which remedies economic inequalities].”Footnote ²²

Those forms of belief and conduct are not exhausted by religion. They are not exhausted by conscience, either. The scope of liberty of conscience is unclear in Rawls: he never defines it, and what he does say is so vague that its religious liberty protection might only bar deliberate persecution.Footnote ²³ His discussion of conscientious objection is confined to resistance to military conscription.Footnote ²⁴ He must have envisioned other extensions of the principle, but does not say what these are.

This gap in the argument leaves Rawls vulnerable to what we will call the Hobbesian Objection, which holds that private conscience is too capricious to be an appropriate basis for exemption from legal obligations. Hobbes thought human beings were impenetrable, even to themselves, their happiness consisting in “a continuall progresse of the desire, from one object to another; the attaining of the former, being still but the way to the later”;Footnote ²⁵ their agency consisting of (as Thomas Pfau puts it) “an agglomeration of disjointed volitional states (themselves the outward projection of so many random desires).”Footnote ²⁶ There is no common good for men to orient themselves toward: “since different men desire and shun different things, there must need be many things that are good to some and evil to others. … Therefore one cannot speak of something as being simply good; since whatsoever is good, is good for someone or other.”Footnote ²⁷ No appeal to “such diversity, as there is of private Consciences”Footnote ²⁸ is possible in public life for Hobbes.Footnote ²⁹

Rawls wants to authorize such an appeal, but he also thinks that we must regard one another with a model of agency as opaque as that of Hobbes, in which for all we can tell the man who compulsively counts blades of grass is pursuing what is good for him.Footnote ³⁰ If people are thus incommensurable, then it is not apparent how some of their desires can legitimately be privileged over others. Conscience is the same black box that it was in Hobbes. Rawls does not explain how his conception of “liberty of conscience” can answer the Hobbesian Objection.

Michael Sandel observes that among the “circumstances of justice” that motivate Rawls's liberalism is an “epistemic deficit” in “our cognitive access to others.” “Where for Hume we need justice because we do not love each other well enough, for Rawls we need justice because we cannot know each other well enough for even love to serve alone.”Footnote ³¹ This is why it is hard for Rawls to answer the Hobbesian objection. Sandel responds that Rawls exaggerates our mutual unknowability. Sometimes people are fortunate enough to live within “a common vocabulary of discourse and a background of implicit practices and understandings within which the opacity of the participants is reduced if never finally dissolved.”Footnote ³²

Some limitation on our mutual opacity must be stipulated if the liberty that Rawls posits is to be institutionally realized. There will have to be some way the state can discern whether “conscience” is really implicated in any particular objection to a law.

In the original position, the parties do not have enough information about one another's ends to penetrate the opacity. A Theory of Justice relied on a thin moral psychology based on the “Aristotelian principle” that people want to develop their talents and skills.Footnote ³³ Political Liberalism retreats to an even more parsimonious psychology. In the original position as revised, the parties understand their highest-order interests in an exceedingly abstract way, one that seems to preclude privileging any specific, contestable conception of the good. Persons are regarded as free and equal in virtue of their possessing to a sufficient degree the two powers of moral personality, the capacity for a sense of justice and the capacity for a conception of the good.Footnote ³⁴ The two moral powers are derived analytically from the minimal requirements of human agency and collective self-government. No one can act without a conception of the good; no collectivity can fairly govern itself without some sense of justice. The conception of justice is built up from this political conception of the person. Rawls thinks that any conception of the good that is not analytically derivable from these thin premises cannot be the basis of social unity, because of the inevitable plurality of comprehensive conceptions. The moral powers, as he describes them, are at least consistent with—perhaps they are inferences from—maximal opacity, in which all we know about our fellow citizens is that they are agents with whom we must somehow cooperate. “A liberty is more or less significant depending on whether it is more or less essentially involved with, or is a more or less necessary institutional means to protect, the full and informed and effective exercise of the moral powers in one (or both) of the two fundamental cases.”Footnote ³⁵ But in some sense all human conduct involves one's conception of the good. People do things, even trivial things, for reasons. So what basis could there be for singling out some determinate subset of liberty and placing it under the umbrella of “liberty of conscience”?

3. The Second Stage

Rawls envisions a sequence of steps by which the principles of justice are to be institutionalized. Once those principles, including “liberty of conscience,” are in place, a second stage of deliberation designs a constitution. At this stage, the parties know “the relevant general facts about their society, that is, its natural circumstances and resources, its level of economic advance and political culture, and so on.”Footnote ³⁶ The designers of the constitution also have “knowledge of the beliefs and interests that men in the system are liable to have.”Footnote ³⁷ Samuel Freeman explains: “This information is relevant since societies with different histories, cultures, resources, and levels of development might require different kinds of constitutions to enable them to best realize the requirements of justice.”Footnote ³⁸

Rawls also says that, at the constitutional stage, the parties “do not know their own social position, their place in the distribution of natural attributes, or their conception of the good.”Footnote ³⁹ They do not know their own personal conceptions of the good, because “any knowledge that is likely to give rise to bias and distortion and to set men against one another is ruled out.”Footnote ⁴⁰ Their task is to “weigh the justice of procedures for coping with diverse political views,”Footnote ⁴¹ and differing conceptions of the good are among those diverse views. But this does not mean that they do not know the conceptions of the good that are prevalent in their society, “the beliefs and interests that men in the system are liable to have.” The interests that the veil keeps them from knowing, at this stage, are “particular facts about individuals such as their social position, natural attributes, and peculiar interests.”Footnote ⁴²

At the constitutional stage, then, it is possible for the parties to take account of which “forms of belief and conduct,” in this culture, are particularly likely to be important to the natives. The constitutional convention, aiming to institutionalize “liberty of conscience,” should try to discern which interests have that degree of urgency.

Discerning which set of actions should be especially protected must begin with one's knowledge of the general facts about human psychology. One fact about human psychology is that there is substantial—not perfect, but substantial—overlap in what humans regard as good. That overlap is greater within most nation-states than it is on the planet as a whole. That makes it possible to answer Hobbes.

In the United States, for example, local circumstances make “religion” an attractive candidate for protection. The population is unusually religious, and the term “religion” is widely used to denote a definite set of deeply held values. Americans’ religious beliefs often motivate socially valuable conduct. Hardly any of the various religious groups seek to violate others’ rights or install an oppressive government. All religions are minorities and so distrust government authority over religious dogma. Although “religion” is a term that resists definition, courts have had little difficulty determining which claims are religious, and the question is rarely even litigated. The rules described at the beginning of section 1 have thus been objects of broad consensus. Even if the growth of unbelief raises new difficulties for this approach, it has worked for a long time.Footnote ⁴³ Michael Walzer has famously objected to the original position's blindness to the particular needs of people in particular times and places: the parties’ reasoning

At the constitutional stage, however, the parties must be aware of just the local values that Walzer thinks are necessary to undergird a fair distribution. They might well embrace the complex equality that Walzer advocates. Indeed, the reading of “liberty of conscience” that we have just offered implies that the parties need to consider those values in order to reach determinate answers to the questions that Walzer asks. The deliberators at the constitutional stage could, for example, legitimately single out religion, and so end up with something like the American law of religious liberty.Footnote ⁴⁵ In the original position, “systems of ends are not ranked in value.”Footnote ⁴⁶ But this is no longer true at the constitutional stage. Communitarians and perfectionists thus need have no quarrel with Rawls, so long as they are willing to respect the basic liberties and a minimum guarantee of resources.

Local values are never unanimous, of course. Specific measures to protect particular instantiations of liberty of conscience will always leave some people out. Many Americans do not think religion is valuable, or think that other ends are equally or more valuable. (Rawls, of course, never took a position on that question. Nor do I.)

Since the absence of unanimity is inevitable, it cannot be an objection to any particular political arrangement. Rather, it is a problem to be managed: there are always minorities, and minorities within minorities. The question of which locally valued allegiances ought to be the object of majoritarian recognition is not different in kind from the question of which locally valued allegiances ought to be the object of rights protection. Both are contingent on what matters urgently to a large proportion of the locals. The state can recognize categories of deep and valuable human concerns, but any such categories will inevitably be overinclusive and underinclusive. The only remedy is supplementation by additional categories which will themselves be similarly imperfect. The remainder can be reduced but never eliminated.

4. The Awful Situation

The parties in the original position aim to prevent what we will call the Awful Situation.Footnote ⁴⁷ The maximin rule in the original position entails that they must agree to avoid “outcomes that one can hardly accept.”Footnote ⁴⁸ But what can one hardly accept?

The answer to this question can help to remedy some of the indeterminacy that plagues Rawls's late work. Rawls eventually acknowledged that there is “a family of reasonable though differing liberal political conceptions.”Footnote ⁴⁹ Even if Rawls's basic framework is accepted, “there are indefinitely many considerations that may be appealed to in the original position and each alternative conception of justice is favored by some considerations and disfavored by others.”Footnote ⁵⁰ Freeman observes that the concession that there will not be general agreement on justice as fairness “must have been an enormous disappointment to him, for he had worked for nearly forty years trying to show how a well-ordered society where everyone accepts justice as fairness as its public charter is a realistic possibility.”Footnote ⁵¹

One approach to social contract theory, already present in Locke, is what Jeremy Waldron calls “negative hypothetical contractarianism”: we can rule out some principles of justice using a contractarian approach. One can “show that a suggested principle of justice is unacceptable by showing that there is no remotely plausible or coherent counterfactual hypothesis under which the principle would command the universal consent of citizens.” Footnote ⁵² The deployment of maximin is a similar move. The Awful Situation is unacceptable because (the deliberator in the original position anticipates that) the person subjected to it will find it to be so. Its awfulness is a matter of psychological fact.

On one possible reading, the Awful Situation would simply refer to a state of affairs in which some “experience their condition as so miserable, or their needs so unmet, that they reject society's conceptions of justice and are ready to resort to violence to improve their condition.”Footnote ⁵³ The most obvious response would be to guarantee the minimal income necessary for a decent life, and the first principle of justice already provides for at least this.Footnote ⁵⁴ Unmet needs might however not be material. Some idiosyncratic disutility monstersFootnote ⁵⁵ —of whom religious conscientious objectors are only a subset—might need special accommodation in order to avoid their own personal Awful Situation. If they are entirely idiosyncratic, however, it is doubtful that their predicament can be remedied.

The standard answer to the familiar utility monster is that she can educate herself to have less expensive tastes. If the regime does not adapt to people's unintelligible and chaotic utility curves, most will learn to adapt to their situation, which consequently would become less Awful. The standard welfarist response addresses such issues by offering each person a reasonable share of resources, protecting property, facilitating contracts, and letting each pursue happiness in their own way. The same response might reasonably be made to the disutility monsters. Special provision can be made for situations likely to cause distress for anyone, such as disability and disease, but unique personal disutilities are generally ignored by the polity and left for the individual to work out. Given the difficulty of interpersonal comparisons of utility, it is doubtful that, within a utilitarian framework, one can do much better than that.

The choice would be different if there were some reason to think that the goals blocked in the Awful Situation are goals that have some independent value. Then the frustration would have a weight that is both interpersonally intelligible and a valid basis for interpersonal claims. What you are frustrating is not a blind brute urge of mine, but my access to something that is genuinely good. If the parties in the original position know that there are such goods, then they have reason to avoid blocking those goods, even in cases where their understandings of those goods are minority views.

We are in our depths opaque to one another. But we are similar enough to know where the deep places are likely to be.

Those deep places consist, in large part, in goods toward which we are drawn. The valorization of choice itself makes sense only if the objects of choice have independent significance, so that some choices are especially weighty.Footnote ⁵⁶ These choices are the “fundamental religious, moral, and philosophical interests” that the parties in the original position “must keep themselves free to honor.”Footnote ⁵⁷

The goods are contestable. Some people reasonably reject them. Many are indifferent to religion. Some have never felt sexual desire. Some find the demands of morality alienating.Footnote ⁵⁸ The exigency of these goods is nonetheless a general fact about human psychology, at least in American society. Around here, one—not the only!—locus of depth is the nebula of practices and longings that cluster around the loose term “religion.”Footnote ⁵⁹ If people were radically idiosyncratic in the needs that they assigned such weight, then even at the constitutional stage, the parties would have no basis upon which to discern “forms of belief and conduct the protection of which we cannot properly abandon or be persuaded to jeopardize.”Footnote ⁶⁰

We share recognition of the value of these goods, at least at an abstract level. That fact illuminates our individual perspectives on substantive religious beliefs that we find preposterous. Your specific religious beliefs and rituals strike me as weird and repellent. I am amazed that anyone can find transcendent meaning in that. But I know that religion falls within a field of human activity in which many of us deem our own beliefs and rituals good and worthy of respect, and in which our religious commitments are often unintelligible to one another. I can appreciate the urgency of your demand for a space in which to pursue your idiosyncratic religious needs.

This structure of argument supporting toleration and accommodation is not unique to conscience or religion. That is why the religion analogy—and the objection to singling out religion as if it were uniquely important—are both sometimes powerful. Consider sex. Your specific desires strike me as weird and repellent. I am amazed that anyone can be turned on by that. But I know that sex falls within a field of human activity in which many of us deem our own desires good and worthy of respect, and in which our desires are often unintelligible to one another. The situation of gay Americans in the 1950s is another variant of the Awful Situation. I can appreciate the urgency of your demand for a space in which to pursue your idiosyncratic needs.

Even the accommodation of conscience has this form. Your specific understanding of what morality requires of you strikes me as weird and repellent. And so forth. The will to act morally has interpersonally intelligible value, and this too may be a reason to accommodate conscience as such.Footnote ⁶¹ But conscience cannot completely substitute for the other categories of accommodation.

In each of these categories, the case for toleration rests on a distinctive interlocking pattern of mutual transparency and opacity. Were there no transparency, we would not have devised these categories, which transcend our own specific orientations toward the good as we apprehend it. Were there no opacity, we would not be impelled to institutionalize our appreciation of the good under such intentionally vague descriptions as “conscience” or “religion” or “sexuality.”

None of these categories can fully capture the Awful Situation in any society. All are somewhat overinclusive and underinclusive. It would be a mistake to rely solely on any of them. Religion isn't that special. But there is no alternative to such imperfect proxies.

Because it is possible, to some extent, to penetrate this opacity, the case for doing so is compelling. Parties in the original position would not agree to terms of cooperation that forbid reliance on contestable conceptions of the good, if such reliance is necessary in order to avoid the Awful Situation. They would deem that state of affairs as unacceptable as destitution or marginalization in an unregulated market.

Of course, many accommodation claims are not based on an Awful Situation. The claimant simply values his religion and wants to exercise it. Religious accommodations are sometimes defended with the claim that the law ought to defer to a believer's fear of extratemporal consequences. The burden of obeying a law, so the argument goes, is greater for a person who believes that doing so will put his soul in jeopardy for eternity.Footnote ⁶² This claim has a narrower extension than its proponents think. Many doings that have been protected under the description of “freedom of religion” are not responses to the threat of damnation. Such claims are less exigent, but they still have weight, and it is permissible for the state to cognize and accommodate the relevant goods.

Here at last we come to the present American regime. American law gives religion special treatment in many cases in which the adherent's need for accommodation is not urgent. The adherent simply is pursuing goods associated with his religion, and the state is facilitating that pursuit.

If however it is permissible for the state to take account of those goods in order to prevent the Awful Situation, then it must also be permissible to take account of those goods in other circumstances.Footnote ⁶³

The justification of American law's special treatment of religion must rest on this point. Neither the protection of religion from corruption via disestablishment, nor most accommodations of religion (which are not responses to religious duties), respond to the Awful Situation. Our exploration of that Situation, however, shows why they are permissible measures at the constitutional stage.

5. Answering the Hobbesian Objection

Return to the Hobbesian objection. It is hard to construct an intersubjective anchor for “liberty of conscience.” One might, perhaps, interrogate individual conscientious objectors, in order to determine whether their claim is sufficiently deep to demand respect. That is what draft boards used to do. But that is itself a highly fallible method of detection, and would rule out a lot of accommodations that are familiar and uncontroversial.

During Prohibition, the Volstead Act exempted sacramental wine.Footnote ⁶⁴ No attempt was made to examine individual Catholic priests and parishioners to determine the depth of their conviction. If “religion” is not cognizable, it is hard to imagine how that could have been done. Conscience is at best a complement, not a substitute, for teleologically loaded terms such as religion.

Hobbes's skepticism can be avoided because our agency consists in the pursuit of ends outside ourselves. Those ends can provide the intersubjectively intelligible basis for singling out certain choices as especially important. That may approach the teleological conception of agency we find in Aristotle or Aquinas,Footnote ⁶⁵ which, of course, accompanies a politics that is not particularly liberal. But in recognizing the value of religion, the liberalism in question here is not committed to the idea that a life with religion is better than one without it. Rather, it merely cognizes this as one of many distinctive goods whose value is not (experienced as) merely an artifact of human choice, and which therefore may legitimately be privileged over other choices. Charles Taylor refers to such goods as “hypergoods,” “goods which not only are incomparably more important than others but provide the standpoint from which these must be weighed, judged, decided about.”Footnote ⁶⁶ If the Awful Situation consists in having one's access to such goods blocked, then the neutrality that Rawls's theory of justice entails is neutrality among hypergoods, at least to the extent of removing obstacles to them without privileging any.Footnote ⁶⁷ It is because justice as fairness, properly understood, is responsive to such hypergoods and accommodates them where possible that Rawls is entitled to claim that it gives “adequate protection to the so-called positive liberties (those involving the absence of obstacles to possible choices and activities, leading to self-realization).”Footnote ⁶⁸

The Awful Situation is too vaguely specified in the original position to generate many of the rights that we regard as important. It is, however, a potent right-generating device. At the constitutional stage, the deliberators must (if they are to be faithful to the decisions already made in the original position) consider what counts as an Awful Situation in their society and take whatever steps are necessary to prevent it from coming about. That means knowing which hypergoods happen to be valued around here. If there is broad overlap across human cultures in what is locally regarded as a hypergood, then there may even be a basis for a fuller understanding of universal human rights than that imagined in a thin conception of agency.Footnote ⁶⁹

The parties in the original position, anticipating the constitutional stage to follow, must make decisions that will only be able to be operationalized later on the basis of information not yet available. Instructions are often contingent on unknown future developments. A general can say, “Engage with the enemy whenever you make contact,” without knowing where and when that will happen. The parties in the original position declare, “Prevent the Awful Situation.” They do not know when it will threaten to occur.

They can specify a bit more. The Awful Situation that is pertinent in the case of religious oppression is not mere disutility. It is a blockage in the attempt to realize an urgently valuable end. Gerald MacCallum pointed out long ago that any claim about liberty necessarily refers to a triadic relation between agents, restraining conditions, and action. One should always ask, who is free, from what restraint to perform which action?Footnote ⁷⁰ The Awful Situation of the person who is religiously persecuted takes this form: the persecution is an obstacle that stands between the person and the hypergood that she is aiming to pursue.Footnote ⁷¹ It is the exigency of that hypergood that makes it ridiculous to say, “OK, so the law won't let you perform a Catholic Mass. Why don't you just find something else to do with your Sunday mornings?”

Rawlsian “liberty of conscience,” then, would be a presumptive right to pursue hypergoods. Which hypergoods are salient is not knowable in the original position. That depends on local cultural conditions. But these limitations of the original position do not prevent the state from pursuing the good, any more than the general's ignorance prevents his troops from pursuing the enemy. He is ignorant of some salient facts, but by the time they do their job they had better not be blindfolded. At the constitutional stage, the parties carry out the marching orders formulated in the original position by specifying and protecting the hypergoods that are salient in their own societies. They need to know what those hypergoods are. They necessarily reject neutralitarian liberalism, the idea that the state must be neutral among contested conceptions of the good. That understanding of liberalism would prevent them from implementing their already-formulated commitment to “liberty of conscience.”

Return to the objection from public reason. We saw that “liberty of conscience” was capable of implementation only because the veil of ignorance became somewhat more transparent at the constitutional stage. Public reason, however, imposes a kind of veil of ignorance that persists even at the stage of individual political behavior. It limits the legitimate use of political power even in the voting booth.Footnote ⁷²

Nonetheless, if public reason is understood to exclude any notion that religion is valuable, then it is impossible to protect “liberty of conscience.” Public reason will constrain, not only the deployment of coercive state power against nonbelievers, but also the decision to selectively accommodate believers. We are back in the world of Hobbes. There would be no basis for singling out any particular desire of any particular citizen for special treatment. All such desires would be equally opaque, because the conceptions of the good that could penetrate the opacity would be ruled out as a basis for political action. By deciding to protect “liberty of conscience,” the parties in the original position must be presumed to have agreed to allow such local facts to enter into political deliberation at the constitutional or legislative stages: he who wills the end wills the means.

Another possibility might be to protect liberty of conscience indirectly, under the description of more general rights (so that heresy, for example, is protected as free speech),Footnote ⁷³ or to disaggregate freedom of religion into its component goods.Footnote ⁷⁴ In a way, this is Hobbes's approach: there are no individual rights against the state, but the sovereign's interests entail a broad field of liberty for the subjects.Footnote ⁷⁵ Even if one goes beyond Hobbes and adopts the familiar schedule of nonreligious rights, some familiar accommodations would be impossible. How could the sacramental wine question be resolved without reference to “religion”?Footnote ⁷⁶

Few philosophers are willing to abandon all individual accommodations. Those who do generally rely on arguments from political prudence that, within Rawls's framework, are appropriately considered at the constitutional stage. They do not rule out all religious accommodation as a matter of principle.

If the parties at the constitutional stage do intend to implement freedom of conscience, they must have some idea of what goods are salient, what counts as an Awful Situation, in their culture. They can do this because they can cognize conceptions of the good that are unavailable to the parties in the original position. Thus they can legitimately treat religion as a good. (Just how they understand “religion” might shift over time as the religious views of the society shift over time.)Footnote ⁷⁷ That can be the basis for religious accommodation (and similar logic can accommodate other exigencies, such as conscience). The same positive valuation of religion might justify a requirement that a law have a secular purpose, because the parties might reasonably think that laws that in effect endorse religious beliefs are likely to corrupt religion.Footnote ⁷⁸ Such corruption does not generate Awful Situations, but it obstructs the realization of a good that the parties can legitimately seek to facilitate.

The parties might conceivably agree on norms of public reason that should guide their exercise of political power. Whatever those terms might provide, they cannot prohibit voters or officials from relying on the very conceptions of the good that are the basis of freedom of conscience. So the aspiration to deduce, from a Rawlsian idea of public reason, a general antiperfectionism is barred by the idea of liberty of conscience that the parties would embrace in the original position.

The parties at the constitutional stage will be aware of the burdens of judgment that inevitably generate disagreement about fundamental moral issues. Those burdens are a fact of political life in a free society. But many public reason theorists claim, not only this, but also that we can confidently say where those burdens are salient—with respect to which propositions and which issues. In fact, the set of intractable disagreements is constantly in flux.

Christopher Eberle argues that the obligation of mutual respect permits the religious citizen to freely offer her religious reasons for proposed legislation so long as she continues to pursue a search for public reasons and thinks that it will eventually be possible to do so.Footnote ⁷⁹ There is a sense in which Eberle is proposing terms of public reason; his argument is intended to persuade religious and secular citizens alike, and to be the object of overlapping consensus.Footnote ⁸⁰ Civic friendship demands that we keep trying to bridge the gaps between us that are the inevitable consequences of the burdens of judgment. These gaps do not pertain only to comprehensive views. They are common with respect to medicine or economics, for example. They are ubiquitous in political discourse. Even when religious disagreements are not implicated, the stakes can be very high. The remedy had better not be anything as philosophically obscure as public reason. Theorists have erected an elaborate structure of duties and ideals that ordinary citizens are unlikely to comprehend, much less follow. Public reason, understood as stringently exclusionary, is thus self-defeating in that simple sense.

A kind of public reason comes into being whenever we exchange reasons with one another. This can be done without ever relying on universally acceptable premises. I can try to take seriously the point of view that each of my fellow citizens holds, addressing them one at a time. My discourse inevitably will often be secular, in that I will avoid reliance on religious premises that I know my interlocutors do not accept.Footnote ⁸¹ But this is a response to a rhetorical imperative, not a moral one.Footnote ⁸² The fundamentals of any reasonable political conception, Rawls writes, is a set of basic rights and liberties, which have priority over perfectionist values, and “measures ensuring for all citizens adequate all-purpose means to make effective use of their freedoms.”Footnote ⁸³ With these in place, a political structure in which my fellow citizens have a vote, and in which therefore I am compelled to take their views seriously if I am to hope to prevail in the face of political disagreement, may be all the public reason that is needed. “The idea of public reason specifies at the deepest level the basic moral and political values that are to determine a constitutional democratic government's relation to its citizens and their relation to one another.”Footnote ⁸⁴ The strictest constraint of public reason is the veil of ignorance in the original position. If a regime's practices are consistent with the basic rights set forth there, then there is agreement “at the deepest level.” The constraints of public reason are not so severe at the constitutional stage, or in ordinary life.Footnote ⁸⁵

6. Surplus Repression

The argument just laid out may be the correct Rawlsian resolution of the problem of free exercise.Footnote ⁸⁶ If it is correct, then any objection that “conscience” ought to be substituted for “religion,” as Leiter proposes, fails to grasp the institutional limitations of the abstract idea of “freedom of conscience”—perhaps we can now call it freedom of exigency—and does not distinguish the original position from the constitutional stage. The Awful Situation is not directly perceptible by third parties. The parties in the original position will know this, since “they know the general facts about human society.”Footnote ⁸⁷ The most promising device for detecting that Situation is reliance on local cultural proxies such as “religion.” (Conscience, in its ordinary semantic meaning, is another such proxy, a subset of Rawlsian “freedom of conscience.”)Footnote ⁸⁸

Many legal categories do not precisely track moral reality. We want to give licenses to “safe drivers,” but these are not directly detectible, so we use the somewhat overinclusive and underinclusive category of “those who have passed a driving test.” So even if “religion” is not a distinctive human good with unique value, it may be an appropriate legal category.

Religion is also singled out for special treatment by the Establishment Clause. School prayer is unconstitutional, but the public schools remain free to unfairly privilege nonreligious “comprehensive views,” such as the philosophy of Hegel. The unfairness is the same as Leiter's case of the Sikh who is unfairly permitted to carry a kirpan when the inheritor of a secular dagger-bearing tradition is not.

It is, however, relevant that both of these marginal cases, the schools’ indoctrination into Hegel and the secular dagger-bearer, are philosophers’ fictions. Concern about these cases is ubiquitous in religion clause scholarship.Footnote ⁸⁹ It is not, however, a response to real injustices.

The scholars, many of whom are religious skeptics, appear to be driven by status anxiety: the special treatment of religion appears to relegate the nonreligious to second-class status. That is why the nonreligious conscientious objection cases are so salient. But the law already deals with such cases by deeming the objections “religious.” The fuzziness of the category of “religion” could lead to unbounded and extravagant claims, but it hasn't happened yet.