The Yale Law Journal

April 2017

Scalia, J., Dissenting: A Fragment on Religion

Stephen L. Carter

[Author’s Note This unpublished opinion was supposedly found among the papers of the late Justice Antonin Scalia. I cannot say with authority that the supposition is true. Whatever its source, the opinion is plainly a draft. The argument contains certain doctrinal inconsistencies, and the prose does not entirely possess the stylishness for which the late Justice was known. Probably additional flourishes were to be added later. In addition, although there are references to the majority’s argument, there is no Supreme Court opinion that matches up with the dissent’s criticism. It is not clear whether the draft was even connected to any particular case–or, as I say, whether it is authentic at all. From internal evidence, the dissent appears to assume that the petitioners are challenging certain religious accommodations, that the government respondents are defending them, and that the intervenors argue that the accommodations do not go far enough. But I am only guessing. I am grateful for the research assistance of Sam Adkisson. Copyright 2017 by Stephen L. Carter.]

Justice Scalia, Dissenting.

The Court today continues its reckless assault on the principle of religious freedom enshrined in the First Amendment. Because I believe that the Constitution contains no more important clause, and because I consider every encroachment on religious liberty an encroachment on democracy itself, I dissent.


Our jurisprudence on the religion clauses of the First Amendment is famously opaque. But instead of clarifying the law, the Court today makes itself the obfuscator of last resort.1 Without ever offering a serious argument, the majority today holds unconstitutional a set of seemingly reasonable accommodations of religion that appear in the laws of a number of states. True, the accommodations before us are controversial, and they even provoke in some corners considerable ire, but neither of these concerns is of the mildest constitutional moment. Appointment to the Supreme Court of the United States is not the same as entrance in a popularity contest. No sane citizen would want things any other way.2

Today’s decision is but another in a long line of cases in which this Court fundamentally misunderstands the Establishment Clause.3 The clause, as I have argued in the past, was never intended to prevent the Government from conferring on religious persons and organizations a reasonable set of privileges not available to the irreligious.4 On the contrary: “Those who wrote the Constitution believed that morality was essential to the well-being of society and that encouragement of religion was the best way to foster morality.” McCreary County v. American Civil Liberties Union of Kentucky, 545 U.S. 844, 877 (2005) (Scalia, J., dissenting). Of course the Government is free in most cases to choose not to grant preferential treatment to religion. The question is whether a “mandatory choice . . . has been imposed by the United States Constitution.” Lee v. Weisman, 505 U.S. 577, 645 (1992) (Scalia, J., dissenting). The only sensible answer is no.

The Government can and does treat religion as special, because the American people can and do treat religion as special. The state may not discriminate among religions, and a favor available to one must be made available to all.5 Accommodations of the sort at issue today may be said to encourage religion in the sense that they reduce the cost to the believer of believing. But a state can make a rational judgment that it is better off with a larger rather than a smaller number of religious believers, as long as it manifests an indifference to which religion they believe.

The majority in its Establishment Clause jurisprudence has not quite become yet another “ad hoc nullification machine,” Madsen v. Women’s Health Center, Inc., 512 U.S. 753, 785 (1994) (Scalia, J., concurring in the judgment in part and dissenting in part), but that is the worrisome direction in which we may be moving. The Court has come to view with suspicion even the most moderate legislative effort to balance the desires of millions to live according to what they think God commands with the growing intrusiveness of the administrative state. Neither the text nor the history of the Establishment Clause supports such an approach. Certainly the Founders encouraged religious belief, believing it essential to the proper functioning of democracy.6 Not even those among the Founding Generation who demanded anti-establishment language in the federal Constitution doubted the authority of the state to support and encourage religion generally.7 The accommodations at issue here do little more. “Nothing stands behind the Court’s assertion that governmental affirmation of the society’s belief in God is unconstitutional except the Court’s own say-so, citing as support only the unsubstantiated say-so of earlier Courts going back no farther than the mid-20th century.” McCreary County v. ACLU of Kentucky, supra, at 889 (Scalia, J., dissenting).

Beyond that brief summary of the essence of my position, I will not swell the United States Reports with repetition of what I have said before.8 I should note, however, that I am perplexed by the constant blather to the effect that such a position would disadvantage minority religions. Oddly, no one says just how. What surely hurts minority religions more is a callous neutrality that leaves them to the political process to protect ways of life that may be under threat.9 Consider the case before us. Stripped of the frills, we are presented with

[Author’s Note: Here there is a gap. Presumably Justice Scalia (or whoever wrote these pages) intended to fill in the facts of the case later.]

It is true that the statutory accommodations at issue offer benefits to the religious that others do not receive. The majority warns that such accommodations must be viewed with skepticism because they grant an illegitimate preference to religion. Ante, at __. Alas, the Court has never offered anything but its own ipse dixit in support of the proposition that the government cannot prefer religion over non-religion. The parade of impressively scary adjectives with which the majority describes imagined dangers of the contrary rule, ante at __, adds nothing to the argument. This should be unsurprising, given that the contrary rule is in fact the right one: “there is nothing unconstitutional in a State’s favoring religion generally.” Van Orden v. Perry, 545 U.S. 677, 691 (2005) (Scalia, J., concurring). The accommodations at issue here are fully within the power of the Government. They pose no threat to any value protected by the Establishment Clause.10


Merely establishing that the accommodations at issue are permitted by the First Amendment does not dispose of the case. Intervenors argue that the exemptions they have been granted are too narrow. The right to free exercise, they insist, demands more. In effect they contend that the rules should be written around them, rather than demanding that they demonstrate their entitlement to the exception in order to receive it. The majority rejects both arguments. According to the Court, ante, at __, accommodation is properly the work of the political branches—an ironic proposition indeed, given that six Justices have today voted to strike down precisely the accommodations that the work of those political branches produced.



Before we can decide whether the exemptions at issue are narrower than what the Constitution commands, we must first determine whether the Constitution commands any exemptions at all. The Court answers, in effect, that accommodations are mandated only in rare circumstances left unspecified. But when the burden on religious practice is “incidental” and the law in question is one of “general application,” the burdened believer must seek relief from the regulator. Ante, at __. This, the majority tells us, is the only sensible reading of the Free Exercise Clause. Otherwise, we are told, the creation of “a prima facie” right to accommodation would allow challenges on religious freedom grounds to every conceivable statute or regulation. To illustrate the presumed absurdity of a broader requirement for accommodations, the Court, ante at __, invents a religion that objects to driving on the right-hand side of the road.11 Such tomfoolery might suffice in the law school classroom, but in a court of law we should worry only about issues likely to arise.

I once reasoned as the majority does now. I joined the Court’s opinions in, among others, Lyng v. Northwest Indian Cemetery Protective Ass’n, 485 U.S. 439 (1988), holding that there was no violation of Free Exercise when the Forest Service decided to develop public lands in a way that was “devastating” to a tribe’s religious practices; O’Lone v. Estate of Shabazz, 482 U.S. 342 (1987), holding that reasonable penal practice trumped a Muslim inmate’s need to attend a religious celebration; and Goldman v. Weinberger, 475 U.S. 503 (1986), allowing the Air Force to punish a Jewish officer for refusing to remove his yarmulke indoors. I also authored the Court’s opinion in Employment Division v. Smith, 494 U.S. 872 (1989), where we held that the use of peyote in a ritual of the Native American Church was not entitled to a constitutional exemption from a state’s drug laws. For disappointed religionists, the only source of relief in these cases was the legislature.

I am not prepared to conclude that all of these cases were wrongly decided. But we were mistaken in each of them to suggest that the very political branches that had chosen to ignore the religious practices now burdened were the only constitutionally permissible saviors of those same practices. Yet this has been our basic approach to Free Exercise claims for three decades. Absent evidence of patent discrimination by government against a particular religion (see, for example, Church of Lukumi Bababalu Aye v. City of Hialeah, 508 U.S. 520 (1993)), we remand religious minorities to politics. But a freedom of religion that must await the Government’s favorable response to a “Mother, may I?” is not a freedom of religion at all.12

We should admit our errors. I did indeed once believe the First Amendment right to religious freedom was as restricted as the Court today reaffirms. “More mature consideration has persuaded me that is wrong.” Tome v. United States, 513 U.S. 150, 167 (1995) (Scalia, J., concurring in part and concurring in the judgment). The trouble with Free Exercise understood as narrowly as the majority understands it is that the liberty the clause is intended to protect turns out to depend entirely on the whim of the passing majority. Thus we are presented with the rare and extraordinary case (I confess that I cannot recall another) in which our critics among the scholarly commentariat were actually right.13

We are asked in effect to trust government officials to make wise decisions about what religious behavior to prohibit and what religious behavior to permit. This is a peculiar way indeed to think about constitutional rights, because the liberty of the believer turns out to include only so much of his belief as the relevant bureaucrat can stomach. “Our Nation has preserved freedom of religion, not through trusting to the good faith of individual agencies of government alone, but through the constitutionally mandated vigilant oversight and checking authority of the judiciary.” Goldman v. Weinberger, 475 U.S. 503, 523 (1986) (Brennan, J., dissenting).


Consider what it means to “exercise” one’s religion freely. It is conceded on all sides that the word includes the ability to worship, preach, and study without the interference of the state.14 But is that all that the word encompasses? To the scholar, religion may be defined as referring properly to practices that give meaning to life and flow from convictions of a particular kind.15 That is also how we should understand the word religion in the First Amendment. In interpreting the Constitution we should look for “the original meaning of the text, not what the original draftsmen intended.”16 We should uncover this original meaning by learning how the words were used at the time of adoption.17 Undertaking that exercise here, we learn that the Founding Generation seems to have understood religion in much the same way. In the late eighteenth century, when the regulatory state was many orders of magnitude smaller, it was taken for granted that members of every sect could live their lives largely as they pleased. Many of the state constitutions of the eighteenth century, some by their language and some by interpretation, implicitly recognized this notion by protecting not only religious opinion and speech but religious practices.18 This approach in turn fostered a religious plurality valued by the Founders and demanded by dissenters.19

This is the vision that should guide our analysis. If exercising religion does not include the freedom of the believer to live and act in the world according to his understanding of what God commands, then the constitutional guarantee is a mere grimbly-gunk.20 Perhaps this simple truth helps explain why in Bob Jones University v. United States, 461 U.S. 574 (1983), numerous religious organizations passionately opposed to the university’s ban on interracial dating nevertheless filed amicus briefs on the school’s behalf, arguing that receipt of government benefits should not turn on whether the government approved of how those in a particular religious community lived.21 Their concern was that if Bob Jones University could be punished for engaging in religious practice that the government disliked, their own practices, under a different government, might be next. The amici were unanimous in the view that religious belief itself entails living in a particular way. Believers have duties.22 For that reason, belief and action are not as neatly disentangled as our cases pretend.

A professor at Yale Law School – not known to be a hotbed of conservative judicial thought – offers the following analogy:

Suppose you believe that the ice on your local pond is thinner than it looks. Very likely you will decide not to skate today. If you happen to be a humanitarian sort, you will warn others as well. You might even contact the local authorities and demand that they put up a sign forbidding anyone from going out there until the condition of the ice improves. Each of these actions would be entirely natural, and, in many circumstances, predictable.23

The analogy, borrowed from the philosopher Gilbert Ryle,24 is imperfect, and, like so much academic writing, considerably overdrawn. But it will do for present purposes. For one sees in the example that, given the skater’s inner convictions, it would be immoral to take no action.25

The point is that belief and action are deeply intertwined. It is not merely that our beliefs provide motivations for our actions; it is that the ability to act on our beliefs helps sustain them. Consider the realm not of religion but of politics. It is fine to believe in your candidate but it is finer to be able to act in accordance with your belief: to campaign, to seek to persuade others, ultimately to vote. If we believe a food is good for us we act rationally when we consume more; if we believe the opposite we act rationally when we consume less. Religious believers see their belief in much the same way.

Freedom of religion recognizes the fundamental historical truth that human beings all through history have believed in a God whose commands are both prior to and superior to the commands of the state.26 This truth holds for nearly every culture of which we have evidence.27 That many influential people today consider the notion of God to be so much bosh does not alter the facts. To the believer, following God’s dictate is not a choice but a duty.28 In Lyng, for example, we dealt with a group of tribes for whom belief and action were tightly integrated. We tend to think of religion as involving creeds and doctrines. Thus it is easy to say that the religion is sufficiently protected as long as it can teach and preach. But the tribes in Lyng handed down their traditions through ceremonies that were connected to specific places. “Rituals are performed in prescribed locations not merely as a matter of traditional orthodoxy, but because land, like all other living things, is unique, and specific sites possess different spiritual properties and significance. Within this belief system, therefore, land is not fungible . . . .” Lyng, supra, at 461 (Brennan, J., dissenting).

The tribes involved in Lyng are only the most obvious example. The religious believer, in order to be a religious believer, requires a certain scope of freedom of action.29 Practice and belief are mutually reinforcing—or, as the Bible puts it, faith without works is dead.30 Each time the state strips away the ability of the believer to act in some particular sphere according to his vision of God’s will, the underlying belief is itself necessarily undermined. Moreover, we put the believer to the unappetizing choice between God and Caesar. Of course there will be times when forcing the choice upon the believer is necessary; but if that choice becomes the ordinary consequence of all legislation and regulation, the result can hardly be called neutral.31 That is why, in Burwell v. Hobby Lobby, 134 S. Ct. 2751, 2783 (2014), we pointed out that to refuse respondents an exemption from a law of general application “would effectively exclude these people from full participation in the economic life of the Nation.”


Burwell, of course, interpreted not the First Amendment as such but rather the Congress’s effort to enact free exercise protections in the Religious Freedom Restoration Act, or RFRA.32 RFRA was adopted in 1993 in the hope of altering the course of Free Exercise jurisprudence. The passage of time and the explosion of regulation make it apparent that the Congress that passed RFRA might have been wiser than we. We should perhaps spend more time than we do paying attention to the constitutional opinions of our coordinate branches, not because they bind us but because they might hit upon arguments we have missed. I am not suggesting that we cede our proper constitutional role. Our emphatic province and duty to say what the law is will remain untouched. Nevertheless, as I pointed out when I was somewhat a younger man, nothing in the Constitution requires the Congress to agree with this Court on the constitutionality of any particular piece of legislation. A legislator, bound by oath, must in good conscience vote against what he deems unconstitutional, notwithstanding how many cases teach otherwise. And the same proposition applies when the disagreement goes in the other direction—when the opinions of this Court suggest that a measure may be unconstitutional, but the legislator is persuaded it is not.33 Most of the time these disagreements will amount to no more than the fool’s gold of political posturing that we can safely ignore, but now and then this endlessly renewed stream of ideas might include valuable nuggets of practical experience from which we can learn.34


The majority, ante at __, insists that its approach will nevertheless grant the occasional religious believer relief from a statute of general application, and points with evident pride to our unanimous decision in Hosanna-Tabor Evangelical Lutheran Church & School v. Equal Employment Opportunity Commission, 565 U.S. 171 (2012). But if the best we can do when challenged to prove that we protect Free Exercise is to trumpet as bold and courageous our ruling that religious groups are free to choose their own clergy without government interference, we must have few accomplishments indeed.35

True, the majority propounds a test of sorts to distinguish the case before us from Hosanna-Tabor, but like most of our Free Exercise jurisprudence, the test is essentially invented from thin air. It bears no relation to either the words of the constitutional text or the meaning those words would have had when they were written. Alas, there is nothing new about this Court’s manufacturing of principles as needed, but further furtive feeble finagling will not save a doctrine that has become incoherent. What goes by the name of neutrality nowadays is as neutral as a political convention. The doctrine takes sides, and religion is the loser.36 The modern administrative state has become our altar to Molech, demanding constant sacrifice.37 And its bureaucrats and supporters, like high priests everywhere, command the sacrifice only of others.

When the scope of the state was smaller, the religious could go about their business with little worry that they might run into a rule limiting their ability to live in accordance with their judgment about God’s will. But with each passing year the space in which the faithful are free to follow their traditions grows a little smaller. So intrusive has regulation become that it is impossible to protect religious liberty without carving out a constitutional sphere the state cannot enter. Otherwise the effort of the individual or community to live as God intends will be walled up by an endless parade of bureaucrats muttering “Yes, but” and pushing the religious back a bit more, a bit more, a bit more. Each restriction seems minor in itself but in sum they are overwhelming. In the end freedom itself will suffer the death by a thousand qualifications.38

I can no longer support the proposition that the liberty to exercise one’s religion can be constrained by the vicissitudes of ever-changing regulatory policy. Consider the believer’s point of view. One follows God in the way that one has always followed God, in the way that one’s ancestors have for hundreds or thousands of years, then is suddenly told by the state that what has gone before is no longer to be permitted for no better reason than that some powerful constituency has persuaded the legislature or an administrative agency to bar the conduct in question. The majority holds that absent extraordinary circumstance, the believer loses. But it is a peculiar constitutional freedom that can be waved away upon partisan whim. If indeed the Government’s bureaucrats are deserving of such deference, we might as well just say that they are characterized by maximal excellence and be done with it.39

But they are not. We owe the other branches deference and respect. We do not owe them fealty. The majority offers no workable standard for determining the narrow class of cases where the believer might prevail. Reason finds no refuge in this jurisprudence of confusion.40 It is true that legislatures have often been more sensitive than courts to the rights of the religious.41 The current era, however, does not seem to be one in which the trend is likely to continue. In any case, if legislative bodies are protecting rights that we are ignoring, that should be cause not for celebration but for widespread juridical embarrassment.

We should tear down the edifice of Free Exercise law built over the past several decades and begin anew. The message of today’s result is that the work will not be done swiftly. It appears that the mansion must be disassembled doorjamb by doorjamb, and perhaps never entirely brought down, no matter how wrongly constructed it might be.42


Even in a more expansive version of free exercise, the religious believer cannot always prevail. Intervenors and numerous amici urge the Court to adopt the “compelling state interest” standard of the Religious Freedom Restoration Act as the proper rule of interpretation when free exercise rights clash with the command of the state. Petitioners answer that the standard would be too difficult for Government to meet, and that in Madison’s multiplicity of sects43 we would swiftly find a multiplicity of demands for exemption. The result would be anarchy.

Of course every constitutional freedom has limits. Should the Court decide to change course and rule, as it should, that the Free Exercise Clause mandates religious exemptions, we will have to determine where the limits of that liberty lie. But I am skeptical that we will be able to develop a one-size-fits-all approach. Part of our task is to give clear guidance to the lower courts that must apply our rulings. In the cases where we cannot explain ourselves with clarity, there is good reason to think we have decided wrongly. The higgledy-piggledy state of religion jurisprudence suggests that we have indeed erred. I trust that should the majority come to the correct view – that the Constitution requires some nontrivial religious accommodations – we will pronounce a clear standard. For the nonce, I am content to reserve determining what that standard should be for the occasion when the Court actually proves willing to reconsider its free exercise jurisprudence.44 I am certain, however, that our current jurisprudence is mistaken.

I dissent.

[Author’s Note: Here the fragment ends. Like me, you have likely spotted any number of unanswered questions. The last section of the dissent strikes me as particularly frustrating, for we are told that accommodations are sometimes required but we are not offered a way of determining when. I append a few observations below.]


The trump card for supporters of neutrality has long been the view that widespread accommodations will lead to anarchy, although the trump may not be a trump if we accept, for example, the view of the legal scholar Abner S. Greene that there is no presumptive duty to obey the law.45

If we agree on the idea of prima facie accommodations but accept a more traditional doctrinal analysis, we would have to determine when the state’s interest is so compelling that a religious objection cannot be permitted. Some would put national security into this cubbyhole, although the draft cases46 then stand on a shaky footing. Others would argue for the primacy of antidiscrimination law, as for example in Green v. Connally.47 (Here Ronald Dworkin, in his final book, offers an intriguing partial dissent.48) But in discussing the question we tend to put the cart before the horse. We ask “Should the believer be able to violate such-and-such a law?” when what we should be asking is “What does the free exercise of religion actually mean?”

In answering this last question, we must be wary of the currently fashionable liberal authoritarianism that holds that every institution—and, in effect, every life—should ideally be regulated according to liberal values. My own support for accommodation of religion is very much tied up with resisting this notion. I believe in dissent resting on diversity of view. Diversity of view rests on diversity of the nomic communities in which children and adults alike are nurtured.49 Those spaces are fragile and must be protected, a proposition that Roger Williams understood when he introduced the metaphor of the wall of separation to American thought. Accommodation offers a vital means of protection.

Do I believe religion is different from other human practices, and different in ways that are constitutionally and democratically relevant? I do, and I have made the argument many times over the years.50 Certainly there are those who passionately disagree,51 but I will not reargue the point here. I do fear that the liberal state of the authoritarians would be an impoverished and intolerant one. The purposeful diminishment of religion seems to me a mark against the legitimacy of the state, which is why I cannot conceive of a situation likely to arise in which my moral and legal sympathies would be with the government rather than the Little Sisters of the Poor.

I mention all of that background in order to explain that although I have considerably more sympathy than the median law professor with Justice Scalia’s Establishment Clause jurisprudence (limited, of course, by the notes above), I am unalterably opposed to the approach that he took to the Free Exercise Clause for most of his time on the Court. We will never know whether joining the majority in City of Boerne was simply a matter of statutory interpretation or signaled a new openness to the claim of the religionist whose practice is burdened by a statute or regulation. Although Justice Scalia’s vote no longer “counts,” one hopes fervently that it was the latter.

I am well aware of the criticism that accommodations are wrong because religions are chosen, and it is thus unfair to place the burdens of that choice on innocent third parties.52 But although I have not the space here to go into detail, this concern seems to me thrice mistaken. In the first place, at least in orthodox versions of Islam, Judaism, and Christianity, no religionist imagines that he or she has chosen God. It is God and not the believer who makes the choice. I understand that many secularists would sharply disagree but on this point they are simply wrong. To insist that believers simply pick which religion to join is to “marginalize their major feature, belonging, which individuals most often experience as ascribed, not chosen, and understand as fixed, not changeable.”53

Second, if one’s priority is not the state but God, then the problem is not that the believer wishes to place a burden on third parties; it is that third parties, acting collectively, wish to place a burden on the believer. That the believer is a member of political society is not ipso facto a justification for that burden, for if that were so then the entire structure of constitutional rights would collapse in the face of a willful majority.

Finally, even assuming that the accommodation does burden innocent third parties, we might view that burden as a tax that is necessary if the otherwise overarching state is to make space for religious nurturance. Of course, as I have noted, the state may choose to weaken rather than nurture nomic communities that teach meanings different from its own. But this is a highly regressive project.

I want to believe that at some level Antonin Scalia understood all of this, and was moving in a more accommodating direction; in other words, I want to believe that the Scalia of the fragment was in the offing. And even if he saw the world quite differently, and even if many of the actual results he reached were troubling, his opinions on the religion clauses—as on other matters—were colorful, witty, and often very deep. He was smart and thoughtful and was perhaps more responsible than any other Justice in recent decades for creating the modern Court. Even the Justices who disagreed with him found themselves mimicking his method. Close attention to text and history are everywhere. I probably agreed with him more often than did many of my colleagues, but “more often” is not necessarily the same as “usually.” I always learned from his opinions, and although I frequently found them maddening, I will miss him all the same.