With all the U.S. Supreme Court’s other controversial decisions in its 2021–2022 term, it’s understandable that its decision not to intervene in the politics of vaccine mandates passed largely unnoticed. Over the dissent of Justices Samuel Alito, Neil Gorsuch, and Clarence Thomas, the Court declined to review a challenge to a New York regulation mandating that health-care workers be vaccinated against COVID-19. The petitioners had objected to the regulation on religious-liberty grounds. Justices Alito, Gorsuch, and Thomas had also dissented from another 2021 decision not to issue an injunction against a similar regulation in Maine.

Even though his view did not prevail, notice should be paid to Justice Gorsuch’s argument that such regulations violate the First Amendment. In both the New York case, Dr. A. v. Hochul, and the Maine case, Does v. Mills, Gorsuch cites the 1981 decision Thomas v. Review Board of the Indiana Employment Security Division, which has become the Court’s go-to precedent for cases in which people object either to having to contribute to an action they believe to be wrong (e.g. baking a cake for a same-sex wedding) or to having to make use of an action they believe to have been wrong (e.g. taking vaccines that were tested or developed using a cell line derived from an abortion). The first kind of case involves what ethicists call “cooperation”; the second involves “appropriation.” The problem is that the decision in Thomas v. Review Board of the Indiana Employment Security Division is fundamentally confused. And that is also a problem for the important cause of religious liberty.

Eddie Thomas was a Jehovah’s Witness. He worked at a roll foundry making sheet steel, which is used in all sorts of things. The roll foundry closed, however, and Thomas’s employer gave him a new job making tank turrets. For religious reasons, Thomas sought to be reassigned to work that did not directly serve the military, only to learn that all his employer’s remaining operations were in arms production. So he asked to be laid off. After his employer denied his request, he quit and applied for unemployment benefits in Indiana, his state of residence. His application was denied, despite the recognition that he had acted on his religious beliefs as a Jehovah’s Witness. Eventually, the case came before the Supreme Court.

In an 8-1 decision, the Court observed that Thomas “drew a line” between working in the roll foundry, which he found “sufficiently insulated from producing weapons of war,” and working on tank turrets, which he believed made him complicit in war-making. The Court went on to state that “it is not for us to say that the line he drew was an unreasonable one.” For “religious beliefs need not be acceptable, logical, consistent, or comprehensible in order to merit First Amendment protection.” Instead, the Court decided that the case turned on the question of whether Thomas “terminated his work because of an honest conviction that such work was forbidden by his religion.”

This is where the confusion starts. It is one thing sincerely to believe, on religious grounds, that war-making (or whatever else) is evil and that, therefore, complicity in war-making is also evil. The government, including the judiciary, has no business judging whether such beliefs are reasonable, or trying to arbitrate disagreements among religious believers about how to interpret religious scriptures and traditions on such questions. The Thomas decision got that much right. But it is quite a different thing to believe that this or that particular action makes one complicit with an activity forbidden by one’s religion. Here, government surely must judge the reasonableness of the belief, unless it is prepared to countenance objections that might seem frivolous or even absurd.

Imagine that Eddie Thomas had not lost his job in the roll foundry, but instead quit after coming to believe that his role in the production of sheet steel made him complicit in evil, since some of the steel he made might go to the military, just as some tax dollars do. Would the Supreme Court have insisted that Thomas was entitled to draw that line, too? What about the conscientious objector who refuses to pay taxes on the grounds that doing so makes her complicit in evil? Is the belief that paying taxes makes one complicit in war-making a religious belief beyond the scrutiny of the Court? If so, should it be accommodated under the terms of either the First Amendment or the Religious Freedom Restoration Act (RFRA)? The IRS currently rejects such arguments as frivolous.

The distinction that needs attention is between the sincere religious belief that, for example, war-making and complicity with war-making are evil and the sincere belief that a particular action constitutes complicity with war-making. No one would deny that the belief that war-making is evil may be a religious belief. Many Christian traditions take Jesus’ injunctions against violence quite literally. It is not the judiciary’s job to decide whether their interpretation of the biblical texts is correct. It may also be a religious teaching that complicity in war-making is evil, and the judiciary has no business assessing this teaching, either. But the court cannot consistently avoid judging whether it is reasonable to describe a particular action as constituting complicity without tying our whole legal system in knots.

If Eddie Thomas had quit his job at the roll factory after coming to the conclusion that even just producing sheet steel made him complicit in the evil of war-making, would we call that conclusion a “religious belief”? No doubt we could, and perhaps, in this hypothetical situation, Eddie Thomas would do so himself, following legal advice. But that description would obfuscate the fact that Thomas was making a judgment about a particular action. Assessing the reasonableness of that judgment does not require assessing the reasonableness of his religious beliefs about war and complicity with war-making.

The line Thomas did in fact draw between working in the roll foundry and working on tank turrets appears quite reasonable. He was aware that the sheet steel he made in the roll foundry could be used by others to make weapons of war, but he rightly understood that kind of cooperation to be acceptably remote. By contrast, he reasonably understood the making of tank turrets to be proximate to the making of war, since these are a probable instrument of what he believes to be wrong. So the Court could have decided in favor of Thomas without treating his judgment that a particular action was contrary to his religious beliefs as itself a religious belief and therefore beyond the Court’s scrutiny.

Justice Gorsuch’s dissents in Does v. Mills and Dr. A. v. Hochul, as well as his concurring opinion in Masterpiece Cakeshop v. Colorado Civil Rights Commission (2018), perpetuate the Court’s confusion in Thomas. In Masterpiece Cakeshop, Gorsuch praises Thomas for having “recognized that Mr. Thomas alone was entitled to define the nature of his religious commitments.” In Does v. Mills, Gorsuch writes that “receiving the COVID-19 vaccines violates [the healthcare workers’] faith because of what they view as an impermissible connection between the vaccines and the cell lines of aborted fetuses.” Similarly, in Dr. A. v. Hochul, the issue, as Gorsuch frames it, is that the health-care workers “cannot receive a COVID-19 vaccine because their religion teaches them to oppose abortion in any form, and because each of the currently available vaccines has depended upon abortion-derived fetal cell lines in its production or testing.” Gorsuch goes on to allow that “many other religious believers feel differently about these matters than they do,” but to him that is irrelevant. Instead, all that matters is that “no one questions the sincerity of [the healthcare workers’] religious beliefs.”

All these cases do involve religious beliefs: in Thomas, beliefs about war; in Masterpiece, about gay marriage; in Dr. A. and Does, about abortion. But it is not a religious belief per se that producing tank turrets makes one complicit in war-making, just as it is not a religious belief per se that baking a cake makes one complicit in same-sex marriage, or that being inoculated with a vaccine tested or developed using a cell line derived from an abortion makes one complicit in abortion. The question to consider in such cases is not whether the judgment is sincerely held, but whether it is reasonable—that is, whether there are good reasons to conclude that the actions in question (making tank turrets, or baking cakes, or getting vaccinated with a particular vaccine) would make a person complicit with the action to which he or she objects on religious grounds.

Consider one more example: Braidwood Management v. Becerra, which was recently decided by a U.S. district court in Texas. The owner of Braidwood, a for-profit corporation, objects to having to provide his employees with health insurance that covers pre-exposure prophylaxis (“PrEP”) drugs that prevent HIV infection. As the district-court judge writes, the owner of Braidwood believes, first, that “providing coverage of PrEP drugs ‘facilitates and encourages homosexual behavior, intravenous drug use, and sexual activity outside of marriage between one man and one woman’” and, second, that “providing coverage of PrEP drugs in Braidwood’s self-insured plan would make him complicit in those behaviors” in violation of his religious liberty under the terms of RFRA. The government’s lawyers argued that the “claim that PrEP drugs facilitate various kinds of behavior is an empirical one that requires factual support,” but the judge, relying on recent Supreme Court precedent, rejected the government’s argument on the grounds that it “inappropriately contest[s] the correctness of [the owner’s] beliefs, when courts may test only the sincerity of those beliefs.” In other words, the owner of Braidwood didn’t need to show that it’s reasonable to believe either that providing insurance that covers PrEP facilitates behaviors to which he objects, or that his company’s provision of that insurance would make him complicit in those behaviors. Instead, he just needed to state that he sincerely believes these claims to be true. This standard strikes me as absurd and unsustainable. It may be a religious belief that certain sexual behaviors are immoral, and that facilitating those behaviors is immoral. But the belief that providing insurance that covers PrEP facilitates those behaviors cannot be understood as itself a religious belief without opening the door to almost any exemption from law one can imagine. Nor does one need expertise in any particular religious tradition to assess the reasonableness of the Braidwood owner’s claim about what constitutes complicity. Logic, prudence, and common sense suffice, and courts generally excel in applying these when they aren’t blinded by bad precedent and ideology.

Some may object that what counts as “reasonable” in the application of general religious teachings to particular circumstances is not entirely separable from the tradition to which those teachings belong. Responding to that objection adequately would likely require a book rather than an article. In Opting Out: Conscience and Cooperation in a Pluralistic Society (2018), David Oderberg argues that the robust teaching on the ethics of cooperation developed within the tradition of Roman Catholic moral theology is “reasonable and plausible in its own right, that is to say, independently of any specifically Catholic doctrine.” Oderberg doesn’t, however, defend that claim at length. For now, the best response to that imagined objection might just be an expression of incredulity: Is there really no way for courts to adjudicate complicity-based conscience claims without appealing to comprehensive doctrines? If so-called public reason is really so out of its depth, then it may soon become possible for Americans to disregard many more laws by appealing to their private and essentially uncontestable sense of complicity.

Zubik v. Burwell, from 2016, takes this possibility all the way to its most absurd conclusion. Like Burwell v. Hobby Lobby two years prior, Zubik was focused on the Affordable Care Act’s contraceptive mandate. At stake was whether the accommodations to the Obama administration’s contraceptive mandate violated the petitioners’ religious liberty by requiring them to inform the federal government, their insurer, or their third-party administrator (TPA) that they were opting out of providing coverage for contraceptives. This requirement, the petitioners argued, would make them complicit in the provision of contraceptive coverage, since the government, their insurer, or their TPA would, upon being so informed, provide the coverage that the petitioners were declining to provide. As Oderberg writes, “If explicitly opting out of some activity ipso facto made a person a cooperator in that activity, then the very idea of opting out would lose all meaning.” It is important to note that Oderberg is a defender of religious liberty against what he calls secular authoritarianism. If friends of religious liberty such as Oderberg find the petitioners’ argument in Zubik absurd, imagine how it will play with people who are already put off by religion on account of its association with reactionary politics. Proponents of religious liberty need to avoid bad arguments, even if those arguments find favor with the current Supreme Court. If this is what religious liberty is supposed to require, then all too many Americans would sooner be rid of it.