The tension between gay rights and religious liberty has become tediously familiar, as evidenced by the endless procession of court cases in which some wedding vendor—a cake baker, a florist, or a website designer—has a conscientious religious objection to facilitating a same-sex wedding. The Supreme Court’s latest decision on the subject has left the issue more confused than it was before. Yet some recent court developments suggest that progress is possible.

Tom Berg’s new book, Religious Liberty in a Polarized Age, makes a powerful case for deploying a broad principle of religious freedom as a strategy both to make American law fairer and to ameliorate the polarization of our politics. The book is a marvelously clear and concise overview of the legal, ethical, and political issues raised by today’s conflicts over law and religion. His conclusions broadly overlap with those that I have developed in my own work, notably my book, Gay Rights vs. Religious Liberty? The Unnecessary Conflict. Here I’m going to focus on an impasse that neither of us resolves—and note a possible way out of it.

We both address the question of whether religious people who conscientiously object to facilitating same-sex weddings, and who therefore decline to provide cakes, photography, or other services, should be exempted from antidiscrimination laws. Berg and I both think they should be accommodated when this can be done without defeating the purposes of the laws from which exemption is sought. With public accommodations laws, Berg observes, the pertinent state interests “include ensuring (1) that people have full access to goods and services and (2) that they not face regular refusals of service that inflict immediate humiliation and ongoing apprehension.” In order to avoid the latter harm, I argued that “businesses should be permitted to refuse to facilitate same-sex weddings if they publicly make their objections clear.” Berg responds, accurately, that “so far no legislature has shown interest,” and LGBTQ rights groups “probably . . . think the notices themselves would stigmatize.” Instead, he proposes that courts should protect “small businesses that provide personal services directly to facilitate a wedding or marriage, where other providers are readily available.”

I’m going to begin my response by repeating some points I made in my book on this subject:

Lawyers are trained to think about conflict resolution by devising abstract principles that should cover all future cases, and which incidentally entail that their side wins. But this is not the only way to think about conflict. Sometimes, the right thing to do is not to follow a principle, but to accurately discern the interests at stake and cobble together an approach that gives some weight to each of those interests. Ethics is not only about principles. There is a tradition in moral philosophy, going back to Aristotle, that holds that a good person does not necessarily rely on any abstract ideal, but rather makes sound judgments about the right thing to do in particular situations. Sometimes principles are overbroad generalizations from experience, and distract us from the moral imperatives of the situation at hand.

The nature of the accommodation that’s needed in this particular context is both specific and abstract. It may sound paradoxical, but the particularistic judgment at issue needs to take the form of a rule. The standard formulation of the federal and state Religious Freedom Restoration Acts—grant an exemption unless denial is necessary to a compelling interest—is fatally indeterminate as a way of resolving cultural conflict. I wrote:

It’s hard to compromise when we don’t know what we are agreeing to. Lawyers distinguish between rules and standards. A rule is clear and easy to apply: there’s no doubt about how you’re obligated to respond to a red traffic light. Standards, on the other hand, such as the injunction to drive with a reasonable degree of care, involve more discretion and unpredictability in the application. Rules are cruder, but one can be much more confident about how they will be applied. . . . 

That is the deepest problem with any proposed judicial resolution. All . . . involve the application of contestable standards. The principles with which judges work do not yield bright lines, but rather general considerations that must be worked out on a case by case basis. This is a job for legislatures, not for courts.

Any such rule would be an instance of what Aquinas called determinatio. A rule can be derived from the natural law, he observes in the Summa Theologiae, “in two ways: first, as a conclusion from premises, secondly, by way of determination of certain generalities.” The latter necessarily is somewhat arbitrary. We all need to drive our cars on the same side of the street, but one can’t deductively establish which side that should be. But it is not unconstrained. “Though the lawmakers’ determinatio is in a sense free,” John Finnis explains in his book, Aquinas, “it must also be made with due consideration for the circumstances which bear on the appropriateness of alternative laws.”

General principles of religious liberty can’t yield determinatio. The most familiar judicial attempt to deduce rules is the trimester framework for abortion rights that the Court crafted in Roe v. Wade. Few were persuaded. Justice Rehnquist, dissenting, complained that the framework was “judicial legislation,” and the Court, when it overruled Roe, declared that it had “concocted an elaborate set of rules, with different restrictions for each trimester of pregnancy, but it did not explain how this veritable code could be teased out of anything in the Constitution, the history of abortion laws, prior precedent, or any other cited source.”

So courts can’t enact the notice requirement that I proposed. Berg and I are both proposing determinatio that, we hope, will accommodate religious objectors without defeating the purposes of antidiscrimination law. He and other scholars have proposed a safe harbor for businesses of five employees or less, thus specifying his exemption for “small businesses.” But he understands that “decisions applying a general religious-freedom standard could not set an exact numerical ceiling.” That’s why legislatures have a comparative advantage here.

But legislatures today have no interest in crafting compromises. Red states won’t protect LGBT people from discrimination. Blue states have become increasingly hostile to religious accommodation. Berg optimistically writes: “The Supreme Court has stepped into the void and protected both sides.” The protection it offers is however crude and overbroad. A recent Court of Appeals decision shows that, but a footnote in its opinion offers a possible path to a better resolution.

The case presented the question of religious exemptions from Title VII of the Civil Rights Act of 1964—a statute which, like all of federal law, is subject to the Religious Freedom Restoration Act. A business owned by a conservative Christian claimed a religious right to discriminate on the basis of LGBT status. (There was no actual suit against it, but it sued the Equal Employment Opportunity Commission, claiming a reasonable fear of prosecution in the future.) The question was handled clumsily in the district court—though the judge, Reed O’Connor, could plausibly argue that the clumsiness was mandated by Supreme Court precedent. He addressed both statutory and constitutional issues, on the sensible basis that “the interests of judicial economy” dictated “that the Court of Appeals may be presented with all issues.” He declared that the government had not shown a compelling interest in preventing discrimination for RFRA purposes. He explained that conclusion by relying on a constitutional analysis, the “most favored nation” theory of religious liberty that has now been embraced by the Supreme Court. The Court now holds that the denial of a religious exemption is presumptively unconstitutional if the state “treats some comparable secular activities more favorably.” O’Connor observes that “for no apparent reason other than administrative convenience, Title VII exempts businesses with fewer than fifteen employees.” The Supreme Court had held that such an exception undermines the state’s claim that its interest is compelling. “Since Defendants extend these exemptions to nonreligious decisions, they must treat requests for religious exemptions the same.”

This is a path to exemptions, but it is way overbroad. The most favored nation theory is so capacious that it is hard to imagine any case in which it will not support exemptions, no matter how destructive they are to legitimate secular interests. Its tendency is to construe religious liberty as a right to violate any law that does not pursue its purpose with monomaniacal intensity. As an earlier, saner Supreme Court observed, “All laws are selective to some extent.” The effect is to make the religious a sort of aristocracy that is entitled to violate any law with impunity. The logic extends beyond antidiscrimination law, generating colorable claims on behalf of religiously motivated homicide: if you can kill people for secular reasons like self-defense, is it unfair to deny the same privilege to the Aztecs? Because the Court is not composed of anarchists, I have argued that the most favored nation theory “in practice will be necessarily selective and is likely to benefit claimants the judges like and to constrain laws the judges dislike.” Judge O’Connor, however, can reasonably say that it’s not his job to craft limits upon the rule laid down by the Supreme Court.

On appeal, Fifth Circuit Judge Jerry Smith reached the same conclusion by similar logic, but offered an escape hatch. It is not clear whether the Supreme Court will permit it, but it is worth trying. 

In RFRA cases, Smith observes, “the courts must ‘scrutinize[ ] the asserted harm of granting specific exemptions to particular religious claimants.’” In the instant case, the EEOC “does not show a compelling interest in denying [the religious claimant], individually, an exemption.” This logic has startling implications for any law that aims to prevent cumulative harms. What harm could possibly come from one person discriminating?

Once an exemption is granted, the cost of claiming it is low and so sincerity is almost impossible to determine. So anyone who wants to discriminate will get to do so if they cite a religious reason. There is a danger of opening the floodgates to so many claims that the underlying law is defeated. Courts have generally responded to this by being skeptical of exemption claims that have nonreligious value, such as claims to be free from the obligation to pay taxes. There’s no such skepticism here.

The district court entirely ignored this issue. But the Fifth Circuit did address it, albeit somewhat elliptically. Smith wrote in a footnote:

An example of a less restrictive means of furthering the government’s interest in preventing employment discrimination on the basis of sex under Title VII could involve the EEOC’s propagating guidance that provides a framework for employers, like Braidwood, that oppose homosexual or transgender behavior on religious grounds, to obtain an exemption. The lack of any guidance or method of gaining an exemption gives rise to the inference that the EEOC “has no intention in granting an exception” regardless of an employer’s religious exercise claim.

This could be read as offering government a lifeline: regulations from the EEOC can provide the determinatio that deductive legal reasoning cannot supply. It could say who qualifies for an exemption and, more importantly, who doesn’t.

Courts cannot provide determinatio, but they have determinatio-forcing techniques. Most obviously, statutes can be invalidated for vagueness or overbreadth. The footnote may be read as an instance of that technique. If legislatures or agencies can be compelled to draw lines, then they might incorporate my proposal or Berg’s.

The Fifth Circuit’s footnote is essentially an invitation to legislators and administrators to probe this territory by specifying the field of exemption. Legislatures that want to prohibit discrimination, which may have no interest in allowing any exemptions, are now on notice that they had better create some, or else they will not like the very broad ones that courts will force upon them.

In order for this to work, though, the Court would have to show some interest in clarifying how government can ever satisfy the requirement that a burden on religion be justified by a compelling interest. And that means that sooner or later it would have to rule against some religious claimant on the merits, and explain why. Until it does that, the conflict between religious liberty and gay rights is only a small part of the trouble we are in.