Making Sense of 303 Creative: A Free Speech Solution in Search of a Problem
David S. Schwartz[1]
Abstract
In 303 Creative v. Elenis, the Supreme Court held that a website designer had a First Amendment right to refuse to create wedding websites for same-sex couples, even though she would create such websites for opposite-sex couples and despite the fact that her refusal violated a Colorado antidiscrimination law. 303 Creative purports to resolve a tension between freedom of speech and public accommodations laws as applied to “creative professionals” whose products or services are expressive. But this problem is largely theoretical. It did not really exist outside a small handful of ginned-up controversies between purportedly creative wedding-related businesses run by religious conservatives and their largely hypothetical same-sex couple-clients. The Court’s doctrinal “solution” to this supposed problem distorts free speech doctrine and needlessly threatens the foundations of antidiscrimination law by characterizing public accommodations laws as not content-neutral. The case may be better understood as a political gesture, operationalizing the “promise” made in Obergefell v. Hodges, that the Court would treat continued, private resistance to same-sex marriage as legitimate and worthy of protection. In delivering on this “promise,” the Court has made the tension between free speech and public accommodations laws worse.
In 303 Creative v. Elenis,[2] the Court held that a website designer had a First Amendment right to refuse to create wedding websites for same-sex couples, even though she would create such websites for opposite-sex couples and despite the fact that her refusal violated a Colorado antidiscrimination law. The Colorado law was unconstitutional as applied to the website designer, a “creative professional,” because it would compel her to engage in speech against her religious beliefs. This ruling sounds simple, but 303 Creative creates, fosters, or stumbles into a doctrinal muddle. Its precedential impact on future cases remains obscure, suggesting a range of potential doctrinal pathways.
Viewed one way, 303 Creative represents an effort to resolve a difficult free speech problem. Can public accommodations laws command compliance that is expressive of ideas that the regulated party disagrees with? It seems distasteful to compel a creative professional to speak against her conscience, and one can imagine (as the 303 Creative majority does) a parade of horribles in which LGBT-friendly bakers must make cakes emblazoned with homophobic messages, media consultants must issue press releases for clients whose views they detest, and lawyers must take on cases promoting causes they oppose. The Court apparently saved us from such outcomes by holding that a public accommodations law cannot compel a creative professional “to speak in ways that align with [the law’s] views but defy her conscience. . . .”[3]
This solution is superficially attractive and may be right in theory. But the problem the Court solved is largely, perhaps almost entirely, theoretical: the product of legal performance art. The 303 Creative case might have met the technical requirement of a “case or controversy”—though that is not entirely clear. But it was not a real controversy in the pragmatic sense: a dispute arising organically between people conducting the activities of their lives, as opposed to one choreographed by legal entrepreneurs to create an issue for judicial resolution. Nor was the movement-lawyer-created dispute in 303 Creative representative of a widespread real-world category of disputes. Real controversies claiming that expressive businesses—media consultants, law firms, and the like—are compelled by public accommodations laws to produce speech in violation of their beliefs, are virtually non-existent. Cases reviewing the application of public accommodations laws to expressive associations are exceedingly rare. And public accommodations/free speech conflicts involving market participants whose services are at most partly expressive—like bakeries—have arisen only in the context of wedding-services providers claiming religious objections to same-sex marriage. This tiny handful of cases includes at least some that are ginned-up, fake cases intended to make a political statement against LGBT rights.[4] For example, in 303 Creative itself, the web designer, Lorie Smith, dropped her factual assertion—which turned out to be bogus—that she was contacted by a potential same-sex-couple client before the case reached the Supreme Court; and she had apparently still not launched her supposedly planned wedding website business, seven months after winning her Supreme Court case.[5] These are the sorts of cases that the Supreme Court has often avoided, exercising “the passive virtues” to prudently steer clear of making broad rulings to solve small or non-existent social problems.[6] Yet the Court has failed to exercise such prudence, now being drawn into religious objections to same-sex marriage twice: first in Masterpiece Cakeshop,[7] a suit under the same Colorado statute by a same-sex couple against a baker who refused to make a wedding cake for them; and now in 303 Creative.
This is not to say that test cases as such are bad and should be resisted by the Court, nor that the Court should always avoid theoretically challenging questions whose importance is primarily symbolic. But when venturing into “delicate” areas—the word Justice Kennedy has used to describe the free-speech/public-accommodations tension[8]—the Court should carefully weigh the costs and benefits of doing so. And the Court’s theoretically attractive response to the legal performance art giving rise to 303 Creative comes at a very high price.
The price is that the Court blunders into a problem at the heart of antidiscrimination laws. The tension between such laws and the First Amendment is built into the antidiscrimination laws, because invidious discrimination is inherently expressive—a fact that the Supreme Court has never adequately confronted. Instead, the Court has dealt with the problem by sweeping it under the rug and classifying discriminatory acts as non-speech “conduct”—a solution that, while theoretically unsatisfying, has been pragmatically workable. 303 Creative raises new doubts about this pragmatic workaround by issuing broad language suggesting that public accommodations laws express the “government’s preferred message” and are therefore content-based, and indeed, that they may discriminate on the basis of viewpoint.
But the 303 Creative decision can be viewed another way. Given the largely theoretical existence of the “problem” of creative professionals being forced to violate their consciences, the case might be better understood as a do-over of Masterpiece Cakeshop and a sequel to Obergefell v. Hodges.[9] The 303 Creative decision may have more to do with resolving a specific battle in the “culture wars” than with resolving the tension between free speech and antidiscrimination law or creating a new line of free speech doctrine. Judicial norms push the Court to speak in terms of generalizable principles, and the consumers of Court opinions normally take the Court at its word by seeking to doctrinalize those principles or harmonize the new case with existing doctrine.[10] But sometimes a case’s precedential import has more to do with its unique background context in constitutional politics than with the logical thrust of its stated rationale.
This Article argues that 303 Creative creates more problems than it resolves, because it set out to address a tiny or non-existent real-world problem with an overly broad doctrinal solution, in order to deliver a win to a religious conservative constituency. In Part I of the Article, I argue that the 303 Creative majority’s motivation to decide this case—a ginned-up case that barely qualified as a live controversy—was to deliver a win to religious conservative groups by operationalizing a promise made in Obergefell to respect religiously based opposition to same-sex marriage. This section explores the various doctrinal pathways on the table in Masterpiece Cakeshop—the precursor to 303 Creative—and explains how and why that 2018 decision failed to deliver a satisfying win to religious conservatives.
In Part II, I set out the doctrinal background to the free-speech/antidiscrimination law tension. The conventional resolution of this tension is to treat discrimination as conduct, an analytically shaky but pragmatically workable solution that allows antidiscrimination laws, such as public accommodations laws, to be reviewed under the relatively permissive test established in United States v. O’Brien.[11]
Part III shows how 303 Creative has created significant confusion about how to resolve the tension between free speech claims and public accommodations laws. The case departs from applicable doctrine—specifically, the O’Brien test—due to the majority’s analytical errors and fails to clearly articulate its own underlying logic.
Finally, in Part IV, I discuss the implications of 303 Creative if it is taken seriously as a broad doctrinal pronouncement, rather than a one-off political decision. Although there are ways to read the case to cabin its precedential impact, the decision invites further test cases by plaintiffs seeking to use their businesses as platforms for discriminatory speech.
Though the case law on First Amendment protection for discriminatory acts is strikingly thin, the takeaway is that the Court has never previously taken seriously the notion that antidiscrimination laws raise significant free speech issues. That “problem” has only arisen since the rise of same-sex-marriage recognition, and only because the Court cannot resist involving itself in the largely symbolic culture-war issue of continued private resistance to same-sex marriage.
The culture-war battle over LGBT rights did not end with Obergefell, which recognized a constitutional right to same-sex marriage. More pointedly, Obergefell did not even end the battle over same-sex marriage, at least in the courts. Perhaps it’s more accurate to say that Obergefell worked a truce rather than resolving the controversy over same-sex marriage. Under Obergefell, states are constitutionally prohibited from banning same sex marriage, because gays and lesbians are entitled to equal dignity and the full rights of citizenship in our constitutional order. But as the case was part truce and only part settlement, Justice Kennedy’s majority opinion makes a significant concession:
[I]t must be emphasized that religions, and those who adhere to religious doctrines, may continue to advocate with utmost, sincere conviction that, by divine precepts, same-sex marriage should not be condoned. The First Amendment ensures that religious organizations and persons are given proper protection as they seek to teach the principles that are so fulfilling and so central to their lives and faiths, and to their own deep aspirations to continue the family structure they have long revered. The same is true of those who oppose same-sex marriage for other reasons.[12]
This is a partial take-back of the extension of equal dignity to gays and lesbians. Marriage, to those who seek it, is so central to full participation in our social and civic life, that the denial of the right to marry is a significant denial of any meaningful conception of equal dignity. Loving v. Virginia, in contrast, did not honor those whose religious beliefs held that divine precepts condemned interracial marriage; the Court gave no quarter to the “white supremacy” it correctly said underlay the law.[13] Indeed, in a short per curiam opinion the following year, when other litigants tried to argue that racial integration “contravenes the will of God,” the Court dismissed that argument as “patently frivolous,” deeming it worthy of only a footnote.[14] One might read the concession to same-sex marriage opposition in Obergefell, probably a condition of Justice Kennedy’s decisive fifth vote, as merely a magnanimous gesture to the losing side. But magnanimous gestures to the losing side are not a regular feature in the Roberts Court’s “culture war jurisprudence”—consider the decision in Students for Fair Admission v. Harvard, in which the majority and concurring opinions treated the losing advocates of affirmative action as violators of equal protection comparable to segregationists.[15] In any case, this concession in Obergefell was more than magnanimity. It was a promise.
The nature of the promise was defined negatively in Justice Alito’s Obergefell dissent:
[the Obergefell decision] will be used to vilify Americans who are unwilling to assent to the new orthodoxy. . . I assume that those who cling to old beliefs will be able to whisper their thoughts in the recesses of their homes, but if they repeat those views in public, they will risk being labeled as bigots and treated as such by governments, employers, and schools.[16]
Kennedy’s promise may well have been added in response to this passage. In any event, five justices—Kennedy and the dissenters—seem to have committed to the idea that anti-LGBT animus is not bigotry, at least to the extent that it is confined to opposition to same-sex marriage.
Despite a long-standing conservative majority on the Court, gay and lesbian claimants have won a string of victories. Anti-LGBT positions took two-out-of-three at the end of the twentieth century: Hurley v. Irish–American Gay, Lesbian and Bisexual Group of Boston[17] and Boy Scouts of America v. Dale[18] as against Romer v. Evans.[19] But since then, LGBT rights have had the winning record, scoring major wins in Lawrence v. Texas,[20] United States v. Windsor,[21] Obergefell and Bostock v. Clayton County.[22] It is plausible to speculate that the Court’s conservative majority felt that some sort of compensatory win for religious conservatives on the LGBT-rights front was overdue. The messy facts and lack of internal agreement in Masterpiece Cakeshop delivered a perhaps unsatisfying win for the anti-LGBT side, in the form of a highly fact-specific one-off. It did not (as we will see) hold that there was a right to speak against same-sex marriage where it really counted: in the face of public accommodations laws. Against this backdrop, 303 Creative can be understood as an attempt to operationalize the implicit promise made in Obergefell and only imperfectly realized in Masterpiece Cakeshop: to honor the “utmost, sincere conviction that, by divine precepts, same-sex marriage should not be condoned.”[23]
For religious conservatives, Obergefell’s promise would be thin gruel if it merely recognized the right to condemn same-sex marriage in churches and traditional free-speech fora, where they are already fully protected from regulation. This would make religious opposition to same-sex marriage no more special than any other message posted on a placard on the steps of the Supreme Court building. In general, First Amendment rights only have bite when they resist government regulation. And for religious conservatives, the court victories that matter are those that recognize religious exceptions to generally applicable laws. To operationalize Obergefell’s promise, it would therefore be imperative to find some way in which government was restricting that sort of speech, and the only candidate for such a governmental restriction is in the application of generally applicable laws barring discrimination on the basis of sexual orientation.
Masterpiece Cakeshop v. Colorado Civil Rights Commission[24] appeared to be the case that would operationalize a First Amendment right to oppose same-sex marriage. Whether that right would be based on the Free Speech or Free Exercise clauses, or a hybrid of the two, was the basis of the certiorari petition.
- Background
As discussed further below, antidiscrimination laws aim at conduct motivated by beliefs, and therefore unavoidably have an impact on expressive conduct.[25] Normally, the Court has analyzed antidiscrimination laws as “generally applicable,” meaning that they do not target speech or religion, and are said to regulate “conduct” rather than speech. Under the applicable precedents, challenges to antidiscrimination laws under the First Amendment should trigger something less than strict scrutiny. For speech claims, the O’Brien test would apply intermediate scrutiny[26]; for free exercise claims, Employment Division v. Smith[27] would apply rational basis.[28]
The two main types of antidiscrimination statutes are employment and public accommodations laws.[29] For same-sex marriage opponents, employment laws are not a promising avenue to obtain special constitutional recognition for their views. For one thing, religious organizations have significant exemptions built into most employment discrimination statutes[30]; for another, most of the employment actions regulated by these laws—discharge, failure to hire, demotion, etc.—are readily (if superficially) characterized as “conduct, not speech.”[31] Public accommodations laws, in contrast, prohibit discrimination by “places of public accommodation”—businesses and entities that participate in the marketplace or are “open to the public.”[32] These laws regulate a wider variety of transactions and activities than the employment laws, and thereby offer more potential to find activities with strong or predominant “expressive” elements.
The Masterpiece Cakeshop case arose when a baker refused to make a wedding cake for a same-sex couple. The baker, Jack Phillips, claimed that he would sell generic cakes, or “cookies or brownies” to LGBT customers, but that his religious beliefs forbade him from agreeing to celebrate a same-sex marriage through the expressive medium of his wedding cakes. The couple, Charlie Craig and David Mullins, filed charges of discrimination with the Colorado Civil Rights Commission for violation of the Colorado Anti-Discrimination Act (CADA). The administrative law judge found the baker Phillips to be in violation of CADA and imposed sanctions; the order was affirmed by the full Commission and the state court of appeal.[33]
This fact pattern seemed very neatly to tee-up the question—which would ultimately be addressed in 303 Creative—of whether a public accommodations law could compel a business owner to engage in expressive activity that violated his religious beliefs. But on a closer look, the case was messy. Primarily, there was a dispute about exactly what services the baker was unwilling to provide: a custom-designed cake “with words or images celebrating the marriage” or “a refusal to sell any cake at all.”[34] These and other factual uncertainties bore on whether the baker’s activities were expressive. “In defining whether a baker’s creation can be protected,” Justice Kennedy’s majority opinion acknowledged, “these details might make a difference.”[35] Moreover, the incident occurred in 2012, before Windsor began and Obergefell completed the Court’s recognition of same-sex marriage. This would no doubt complicate the inquiry into the strength of Colorado’s interest in eradicating sexual orientation discrimination manifested as opposition to same-sex marriage.[36]
Perhaps these difficulties did not rise to a level warranting dismissal of certiorari as improvidently granted, but they were enough to motivate the Court to avoid “the delicate question of when the free exercise of . . . religion must yield to an otherwise valid exercise of state power”[37]—an avoidance which would keep the issue in dispute for 303 Creative. In March 2014, while Masterpiece Cakeshop was still pending before the Colorado Civil Rights Commission, a conservative Christian activist named William Jack visited three bakeries that were listed as gay-friendly on LGBT websites.[38] He asked each bakery to make him two cakes in the shape of an open bible, one decorated with the statement “Homosexuality is a detestable sin—Leviticus 18:22” and the other with a bible quotation and an image of two men holding hands covered by a “no” or “ghostbusters” symbol—the red circle with diagonal line through it.[39] All three bakeries refused, and Jack filed complaints for religious discrimination in violation of CADA with the Commission.[40] The Commission rejected Jack’s claims.[41] This was presumably what Jack expected and likely part of a coordinated strategy.These cases were included in the Masterpiece Cakeshop cert petition over a year before the National Center for Law and Policy (NCLP), a conservative religious legal center, filed an amicus brief in Masterpiece Cakeshop laying out Jack’s cases.[42]
The Jack cases were probably designed to create a free-exercise-based argument that CADA should be reviewed under strict scrutiny rather than rational basis under Smith. The general rule of Smith applies rational basis review to generally applicable laws that only incidentally affect religion; but a doctrine was developing in the lower courts that the failure to accommodate religious practices or beliefs would be reviewed under strict scrutiny if any secular exceptions were made in the law. The Supreme Court would eventually embrace that position in Fulton v. City of Philadelphia.[43] The Jack argument was set up to allow the Court in Masterpiece Cakeshop to hold that CADA was not generally applicable within the meaning of Smith, because exceptions were being made for secular bakers to refuse decorating cakes with messages that offended them.
Led by Justice Kennedy, the Masterpiece Cakeshop majority took a parallel, but different, approach. For them, the Jack cases supplied an exit ramp for the Court to avoid the difficult free speech issues. The majority found it significant that the gay couple’s claim was successful while the religious complainant’s was unsuccessful. Taking that “disparate treatment” together with the supposed severity of the remedial order against Phillips the baker,[44] and three remarks by one or two commissioners supposedly showing disrespect for Phillips’s religious beliefs, the Court concluded that the Commission demonstrated “clear and impermissible hostility toward [Phillips’s] sincere religious beliefs” in violation of the Free Exercise Clause.[45] The petitioners had dropped this argument by the time of their reply brief, and at oral argument, and it did not come up until it was briefly raised by Justice Kennedy, though the thread was not picked up by the other justices.[46] Clearly, it was on Kennedy’s mind, however, and he evidently persuaded his majority colleagues to make it the dispositive issue.
Perhaps finding the case-fact-specific “hostility” to be an unsatisfying basis of resolution, Justice Gorsuch seemed to want to announce a broader doctrinal rule, as he would eventually do in 303 Creative. He therefore wrote a lengthy concurrence trying to show that CADA targeted religion, at least in practice, by emphasizing the Commission’s supposed double standard evidenced in the three Jack cases. But as Justice Kagan pointed out in her concurrence, the Jack cases were different: the bakers refused to make a cake for Jack that they would not have made for any customer. And the objection was to the explicit messages put on the cake. She might have added that the public accommodations laws do not convert bakeries into free speech zones requiring non-discrimination among all customers who wish to speak through their cake purchases. Justice Gorsuch responded that Justices Kagan and Breyer were gerrymandering the level of generality: just plain wedding cake in Phillips’s case, but message cake in Jack’s. Justice Gorsuch had a point if, but only if, one assumes or argues that the wedding cake in Phillips’s case was “same-sex wedding cake,” and was expressive of a message approving marriage. This is indeed what Justices Gorsuch and Thomas argued. But consider this head-scratcher from Justice Gorsuch:
For just as cakes celebrating same-sex weddings are (usually) requested by persons of a particular sexual orientation, so too are cakes expressing religious opposition to same-sex weddings (usually) requested by persons of particular religious faiths. In both cases the bakers’ objection would (usually) result in turning down customers who bear a protected characteristic.
Let’s pause briefly over “cakes expressing religious opposition to same-sex weddings (usually) requested by persons of particular religious faiths.” Usually? Are anti-same-sex-wedding cakes a thing? The oddity of that assertion about the bakery business reveals the flaw in Justice Gorsuch’s reasoning. The violation of the public accommodations law is not “turning down customers who bear a protected characteristic” but turning down customers because of a protected characteristic. Even in Justice Gorsuch’s imaginary world where bakeries do a brisk business in anti-same-sex-wedding cakes, most of which are ordered by religious people, his argument that this constitutes discrimination on the basis of religion rests on a disparate impact theory—one relying on correlation rather than the causation required to show discrimination under public accommodations laws. But the Court has not recognized disparate impact theories of discrimination under public accommodations laws, the lower courts are largely against such a theory, and that can of worms should not be opened lightly.[47] Worse, Justice Gorsuch’s argument implies that a public accommodations law requires regulated entities to accommodate every tenet of every religion. But a shop with a “no shoes, no shirt, no service” policy does not discriminate on the basis of religion if it turns away a shirtless, shoeless customer who happens to be a member of a sect that devoutly believes in going forth dressed as Jesus was on the cross. And if a restaurant allows one family to say grace before being served a meal, must it also allow another family to conduct a ritual animal sacrifice at the table?[48]
Moreover, same-sex-wedding cakes, as opposed to plain wedding cakes, may not be a thing either. Wedding cakes may differ from “ordinary” cakes in terms of traditional styles—e.g., white icing, floral decorations, and multiple tiers. But same-sex wedding cakes exist only in the minds of those who object to same-sex marriage and those who wish to privilege that objection in constitutional law. No one speaks of “interracial wedding cakes.” Here, Justice Gorsuch’s verb-switch is telling. The cakes ordered by the Christian activist Jack were indeed intended to “express” a message, because of the particular words and images he asked to have put on them. In contrast, Craig and Mullins intended their cake to “celebrate” their wedding. But wedding cakes do not “celebrate” same sex-weddings absent explicit messages. We can consider the cakes themselves agnostic as to the sex of the couples. Putting the point another way, if Craig or Mullins had gone alone to Masterpiece Cakeshop and asked for a pretty wedding cake decorated with flowers, and had let Phillips assume it was for a heterosexual couple, Phillips would have made and sold the cake; the bakers refusing Jack’s (bogus) cake order would not have.[49]
Justice Thomas also seemed disappointed by the fact-bound, narrow resolution of Masterpiece Cakeshop, and sought his own doctrinal rule for resolving conflicts between free speech and public accommodations laws. Though Thomas took a different tack, Gorsuch joined his opinion. Justice Thomas rhapsodized about the symbolism and hallowed history of wedding cakes to support his contention that making them is expressive conduct. With that premise “established,” he argued that the accommodation demanded by CADA was itself speech. According to Justice Thomas, “When a public-accommodations law ‘ha[s] the effect of declaring . . . speech itself to be the public accommodation,’ the First Amendment applies with full force.”[50] To Thomas, this would mean that strict scrutiny applies. But the proposition, far from an established black-letter principle, relies entirely on a broad and contentious interpretation of a single case, Hurley v. Irish–American Gay, Lesbian and Bisexual Group of Boston.[51] That case, which the Court itself acknowledged was “peculiar,” was not well reasoned and makes a poor foundation for a line of doctrine, as I will argue further below.[52] For now, suffice it to say that the “full force” of the First Amendment meant applying strict scrutiny to CADA.[53]
Notably, Justice Thomas copped out at this point and did not undertake the supposedly applicable strict scrutiny analysis.[54] Given that, as we will see, the 303 Creative majority did not apply strict scrutiny either, no current Supreme Court justice has yet walked us readers through an analysis telling us whether or not a public accommodations law meets strict scrutiny.[55] In any case, Justice Thomas’s Hurley theory laid out a template that the Court could, and perhaps did, follow in 303 Creative. Rather than pursuing a free-exercise approach, or a hybrid free-speech/free-exercise approach, the religious conservative advocates behind these cases were beginning to come to the conclusion that perhaps a free speech argument offered the most promising path for religious-belief exceptions to public accommodations laws.[56]
* * *
Nevertheless, “antireligious hostility” became the basis of decision in Masterpiece Cakeshop, with seven justices (including the liberals Breyer and Kagan) signing on. Clearly, Justice Kennedy was contemplating that resolution by the time of oral argument.[57] It may also be worth noting that the 303 Creative case was put on the Justices’ radar screen in the Masterpiece Cakeshop litigation itself, by NCLP in the Petitioner’s merits brief, which included “excerpts of filings” in the 303 Creative case then pending in the Colorado district court. These excerpts included the all-important stipulated case facts which would form the basis of the 303 Creative decision.[58] Thus, the Court may well have known before issuing its opinion in Masterpiece Cakeshop that it could take a mulligan and decide the core issues on a sanitized record in a few years in 303 Creative.
There is a long-standing tension between antidiscrimination laws and the First Amendment freedoms of speech and association that has been largely swept under the rug. Public accommodations laws, like antidiscrimination laws more generally, have not been previously deemed problematic on free speech grounds.[59] This is because discriminatory acts have conventionally been viewed as “conduct,” not speech. As Justice Thomas summarized in his Masterpiece Cakeshop concurrence, “as a general matter, public-accommodations laws do not target speech but instead prohibit the act of discriminating against individuals in the provision of publicly available goods, privileges, and services.”[60] The 303 Creative dissenters and supporting amici stated the same principle.[61]
But that is an unfortunate oversimplification. The core of the problem is that intentional, invidious discrimination is, in the nature of things, expressive conduct. In whatever form it takes—whether a “whites only” sign, or the physical exclusion of a person from a business premises—the “act” of invidious discrimination expresses, and is intended to express, a viewpoint of distaste, contempt, or a host of related sentiments against the discrimination victim’s group. This is true for a simple reason. The act itself does not violate antidiscrimination law; it is the act accompanied by the reason for the act that may violate antidiscrimination law. Kicking a couple out of a restaurant because they are barefoot would not violate public accommodations laws; kicking them out because they are holding hands and presumed gay would. This is the simple effect of the intent requirement. Intentions manifested in words or actions are expressive of those intentions. And when the discriminatory acts consist of verbal harassment creating a hostile environment, the expressive element of discrimination is even more pronounced.[62]
Discrimination thus, uncomfortably, defies crisp or easy resolution by the dogmatic categories that populate free speech doctrine. This section reviews how the law conventionally resolves the tension between freedom of speech and antidiscrimination law.
The critical analytical question at the start of a free speech analysis is whether the regulation at issue is content-based or content-neutral. A law is content-based if it targets speech—that is, the regulatory objective is some form of speech or expressive activity in itself. Content-neutral regulations are those that are not intended to target speech, and raise First Amendment issues only because they affect speech incidentally. Consider a law prohibiting overnight camping in a public park. The law is designed to promote safety and sanitation in the park rather than to regulate speech. When protesters camp overnight in the park to make a political statement, the law restricts their speech, but only incidentally; the law is content-neutral.[63] Laws that are content-based—that target speech—are reviewed under strict scrutiny. Content-neutral laws are reviewed under the more permissive intermediate scrutiny test of United States v. O’Brien.[64] Under the O’Brien test, a law will be upheld against a free-speech challenge if (1) it is “justified without reference to the content of the regulated speech,”[65] (2) it“promotes a substantial government interest that would be achieved less effectively absent the regulation”[66] and (3) it “leave[s] open ample alternative channels for communication of the information.”[67]
It is clear from this basic doctrine that the initial focus in a free speech challenge is on the nature of the law, rather than the nature of the speech. To be sure, it is exceedingly common for courts and commentators to focus initially on whether the First Amendment challenger’s expressive activity can be labelled “conduct,” rather than “speech,” in which case it would get no First Amendment protection.[68] But the work purportedly done by the speech-conduct distinction can be, and usually is, handled by the content-based/content-neutral distinction. This analysis entails three analytical moves that make irrelevant a First Amendment challenger’s purported subjective intentions to express ideas through conduct. First, the analysis will assume arguendo that a person’s allegedly expressive acts are in fact expressive, and therefore presumptively entitled to some degree of speech protection.[69] Second, the analysis will shift away from an effort to characterize the speech, to an effort to characterize the regulation.[70] And, third, the analysis asks whether the law targets speech, or instead attempts to regulate behavior that does not logically entail expression (even if that conduct might be used expressively).[71]
United States v. O’Brien is a paradigmatic example of this. The Supreme Court upheld O’Brien’s conviction for violating a criminal law against mutilating draft (military induction) registration cards. Chief Justice Warren’s opinion fulminated that O’Brien’s burning of the draft card was conduct, not speech—an absurd characterization of an act that was intentionally and patently symbolic and therefore expressive. “We cannot accept the view,” Warren opined, “that an apparently limitless variety of conduct can be labeled ‘speech’ whenever the person engaging in the conduct intends thereby to express an idea.”[72] But that intention is precisely what makes conduct expressive. Warren’s inaccurate statement, though frequently quoted in support of the supposed speech-conduct distinction, is contradicted by the rest of the opinion, beginning with the very next sentence.
However, even on the assumption that the alleged communicative element in O’Brien’s conduct is sufficient to bring into play the First Amendment, it does not necessarily follow that the destruction of a registration certificate is constitutionally protected activity. This Court has held that when ‘speech’ and ‘nonspeech’ elements are combined in the same course of conduct, a sufficiently important governmental interest in regulating the nonspeech element can justify incidental limitations on First Amendment freedoms.[73]
The Court proceeded to articulate the test described above, and ultimately to uphold O’Brien’s conviction on the ground that the law protecting draft cards from defacement was content-neutral. Although that was a gross mischaracterization of the particular law in that case,[74] the principle is sound, and the “O’Brien test” is now the long-established doctrine for judging content-neutral laws against free speech challenges—that is, laws that do not target speech, but that may impinge on speech incidentally. “Content-neutral” laws are sometimes called “generally applicable,” in that they are deemed not to target speech: they apply generally to the speech and nonspeech elements alike.
One can try to resolve the problem of discriminatory speech by defining discriminatory acts as “conduct, not speech” by doctrinal fiat. This is the course that the Supreme Court has taken. In Runyon v. McCrary,[75] perhaps the leading case on this subject, the Court held that the First Amendment speech-and-associational right of a private school to preach racial segregation in its curriculum did not entitle the school to practice it by excluding black children.
[I]t may be assumed that parents have a First Amendment right to send their children to educational institutions that promote the belief that racial segregation is desirable, and that the children have an equal right to attend such institutions. But it does not follow that the practice of excluding racial minorities from such institutions is also protected by the same principle.[76]
The gist of this passage is that discrimination, at least insofar as it has a practical effect, is conduct: a “practice,” in the words of the Runyon Court.
But this assertion has never been closely analyzed in the small number of cases challenging antidiscrimination laws on free speech grounds. In Roberts v. U.S. Jaycees,[77] a case heavily relied on by the 303 Creative dissent, the Court simply asserted, without analysis, that Minnesota’s public accommodations law prohibiting sex discrimination “[o]n its face, . . . does not aim at the suppression of speech, does not distinguish between prohibited and permitted activity on the basis of viewpoint, and does not license enforcement authorities to administer the statute on the basis of such constitutionally impermissible criteria.”[78] This characterization would have made it appropriate to apply the O’Brien test.[79] Characterizing an antidiscrimination law as content-neutral seems to be sound public policy, but it is difficult to sustain as a matter of abstract theory. For decades, legal doctrine has probably coasted on the idea that the question should not be looked at too closely. A handful of cases, mostly from 40 years ago or more, have rejected both speech and religion defenses to race and sex discrimination claims. But these cases have existed largely on the margins of both free speech and antidiscrimination law, probably because few lawyers have had the toxic audacity to press such arguments in court. In the era initiated by the Trump presidency, those sorts of norms may be dissolving, and the issue may therefore have to be confronted again, and more directly.
The precedents are quite thin on the ground. In one of the supposed leading cases, Hishon v. King & Spaulding, LLP,[80] a law firm unsuccessfully asserted a freedom-of-association defense against a sex discrimination claim for denying partnership to a female lawyer. The Court’s analysis of the issue consisted entirely of two sentences:
Although we have recognized that the activities of lawyers may make a “distinctive contribution . . . to the ideas and beliefs of our society,” respondent has not shown how its ability to fulfill such a function would be inhibited by a requirement that it consider petitioner for partnership on her merits. Moreover, as we have held in another context, “[i]nvidious private discrimination may be characterized as a form of exercising freedom of association protected by the First Amendment, but it has never been accorded affirmative constitutional protections.” [81]
Lawyering is expressive activity, the Court plausibly suggests, and law firms are therefore expressive associations. But why are their discriminatory acts not “accorded affirmative constitutional protections” under the First Amendment? This brief assertion is hardly a robust explanation, and its ending with a quotation from Norwood v. Harrison gets us no further. Norwood preceded the above-quoted sentence with this one: “although the Constitution does not proscribe private bias, it places no value on discrimination as it does on the values inherent in the Free Exercise Clause.”[82] That assertion was conclusory, and also dicta in a case holding that Mississippi could not give state-funded textbooks to discriminatory private schools.[83] No less conclusory is the footnote in the 930-word per curiam in Newman v. Piggie Park Enterprises.[84] Newman held that a plaintiff who successfully obtained an injunction enforcing Title II, the public accommodations provision of the 1964 Civil Rights Act, was presumptively entitled to attorneys’ fees. Hinting that this presumption might be rebutted in a close case, the Court noted that the defenses actually raised were “so patently frivolous that a denial of counsel fees to the petitioners would be manifestly inequitable.”[85] Those frivolous defenses included the contention “that the [1964 Civil Rights] Act was invalid because it ‘contravenes the will of God’ and constitutes an interference with the ‘free exercise of the Defendant’s religion.’”[86]
These assertions brushing aside First Amendment defenses to antidiscrimination laws are eminently sound as a matter of legal policy. The problem is that their rationales are largely unexplained, and the little explanation given seems to rest on a distinction between discriminatory acts and discriminatory speech. This opens the door for the majority’s contention in 303 Creative that a discriminatory act, if classified as “pure speech,” is immune from antidiscrimination law—and hence, no level of scrutiny is required.
The tension between free speech and antidiscrimination laws is not easily resolved, though it could be finessed in various ways. The path of inertia would be to simply deny the expressive element of discriminatory acts and continue to categorize them as mere conduct. This might mean that there is now, due to 303 Creative, a “pure speech” exception to public accommodations laws, with harassing speech perhaps constituting an exception to the exception.
Another solution is to hold our noses, acknowledge the expressive aspect of invidious discrimination, and accept that antidiscrimination laws are not content neutral—they penalize or prohibit expressive acts that cause harm on the basis of protected characteristics or against protected groups—but acknowledge, too, that they meet strict scrutiny. This is the approach taken in 303 Creative by the lower court. Other solutions would require recognition of new doctrinal categories. Robert Post, for instance, has suggested that the market participants covered by antidiscrimination laws have no robust First Amendment protection for speech that is not intended to be part of the public discourse.[87]
Alternatively, the Court could simply decide that discriminatory speech by employers and places of public accommodation is a new category of unprotected speech, like defamation, fighting words, or obscenity. Creating exceptions is the familiar way that the law resolves conflicts between doctrinal logic and public policy. Discrimination could be, and arguably has been, treated as “low value” speech. This resolution was arguably implicit in the statements, quoted above, to the effect that discriminatory speech “has never been accorded affirmative constitutional protections.”
For present purposes, the point is that the 303 Creative majority blunders into this “delicate” subject (to borrow Justice Kennedy’s description from Masterpiece Cakeshop) like a bull in a china shop. The majority opinion reads as if Lorie Smith’s speech and conscience are the only interests at stake. And the opinion repeatedly makes the regrettable assertion that enforcement of antidiscrimination law against a “creative professional” in effect “force[s]” Smith “to abandon her conscience and speak [the government’s] preferred message”[88]; and that “Colorado’s very purpose in seeking to apply its law to Ms. Smith” is “the coercive elimination of dissenting ideas about marriage” and is intended to “to excise certain ideas or viewpoints from the public dialogue.”[89]
Undoubtedly, the majority thinks it can say these things without damaging the fabric of public accommodations laws because it claims to be applying the sentiment to Smith’s supposed “pure speech” rather than to discriminatory “conduct.” But it is far from clear that the Court’s blunt and overbroad characterization of the enforcement of antidiscrimination laws is so easily cabined. Antidiscrimination laws may not aim in the first instance at “excising” racist, sexist, or homophobic “ideas or viewpoints from the public dialogue,” but they do aim at stopping people from acting on those viewpoints in the hope that those viewpoints will eventually pass away. In its affirmative action cases, the same justices who formed the 303 Creative majority certainly extol the idea that racism will vanish, or has vanished, due in part to the force of law. Why not homophobia too?
III. Novelty and Confusion in 303 Creative
Justice Thomas’s Hurley analysis holds that free speech protection is at its maximum—strict scrutiny applies—when “the public accommodation itself” is speech. In this section, I argue that Thomas’s Hurley theory is the clearest way to make doctrinal sense out of 303 Creative, even though Justice Gorsuch’s majority opinion does not squarely articulate that theory. Instead, the majority opinion offers a muddy mashup of Hurley and other not-quite-apposite precedents that obscure the facts that the O’Brien test should have been applied, that Hurley is a doctrinal outlier, and that 303 Creative’s handling of the free-speech/public-accommodations tension is a novelty. Finally, this section argues that the 303 Creative majority begs the crucial and difficult questions at the heart of the case by relying on the unfortunate stipulated facts offered by the parties:[90] the fake case creating a largely hypothetical free speech problem.
Hurley v. GLIB is one of the odder cases in First Amendment doctrine. It struck down what the Court described as a “peculiar” application of a public accommodations law against a parade—a purely expressive activity—and stated a principle to resolve disputes that should hardly ever arise and that should not have arisen in that case. It makes for a lousy basis to build a doctrine, and, as 303 Creative demonstrates, offers a somewhat dubious precedent for resolving the tension between free speech and public accommodations laws.
Hurley is an archetypal “bad facts” case. The dispute arose when the Irish–American Gay, Lesbian and Bisexual Group of Boston, known familiarly as “GLIB,” challenged its exclusion from Boston’s St. Patrick’s Day/Evacuation Day parade. The parade is a major civic event, celebrating both the place of Irish-American culture in the United States and the evacuation of British troops from Boston in 1776. As a civic event, the parade should probably have been deemed a government-sponsored public forum, and GLIB’s claims resolved on the basis of the group’s First Amendment right against the exclusion of its message from that civic event. The problem was that, after having organized the parade for over a century, the city of Boston in 1947 turned its organization over to a private association, the South Boston Allied War Veterans Council, which annually applied for and received a permit for the parade.[91] Although privately organized, however, it was still the St. Patrick’s Day parade in Boston. GLIB sued both the city and the Veterans Council to gain entry into the parade, “as its own parade unit carrying its own banner.”[92] Regrettably, the Massachusetts courts found that there was no state action, and thereby rejected GLIB’s First Amendment theory. However, they ruled in favor of GLIB’s alternative theory, that the Veterans Council violated the state public accommodations law, which prohibited discrimination on the basis of sexual orientation by any “place of public accommodation.”[93]
This clever lawyering and clever judging was too clever by half. The Massachusetts law covered “places of public accommodation,” and a parade is neither a place nor the sort of activity at which public accommodations laws are aimed. Perhaps the Massachusetts courts mistakenly believed that their decision would avoid Supreme Court review if based on an interpretation of state law. But, of course, they triggered the Veterans Council’s free speech rights, because the parade as such is a purely expressive activity, no different from a public address. No one would seriously claim a public accommodations law requires a person giving a public speech to include other messages, particularly incompatible ones. Nor would there be any plausible argument that, for example, a public accommodations law could have forced the Nazis infamously marching in Skokie, Illinois to include an anti-Holocaust banner in their ranks.[94] GLIB’s real dispute was not with the parade as such, but with the city of Boston, or with the Veterans Council as the city’s proxy, in excluding speakers and messages from a broad public event. The attraction of conceptualizing the parade as a sort of walking, temporary public accommodation only arose because of the Massachusetts courts’ unwillingness to characterize the case properly as one involving state action. It should either have identified the Veterans Council’s role as a quasi-public permit-issuing body that unconstitutionally denied GLIB a permit to march in the civic event; or have held Boston liable for a First Amendment violation in ceding authority over a public event to an expressive association entitled to exclude particular messages.
Thus, the case came to the Supreme Court on the erroneous premise that GLIB’s dispute was with the Veterans Council’s parade, a dispute which could only be resolved in GLIB’s favor as a violation of the public accommodations law. Deeming itself bound to decide the case only on the public accommodations issue, the Court ignored the quasi-public nature of the civic event and characterized the Veterans Council as a private expressive association with a right to exclude members or speech that would dilute its message. Indeed, the Court stated its holding thus: a “requirement to admit a parade[-]contingent expressing a message not of the private organizers’ own choosing violates the First Amendment.”[95] This is a reasonable holding, once one takes the flawed premises as given. Again, the Nazi parade contingent can’t be forced to include an anti-Holocaust banner. To the extent that one remains dissatisfied because the Court mischaracterized the nature of the St. Patrick’s Day parade and the Veterans Council’s role, the mistake was not in calling that organization a private expressive association, but in failing to question the city’s delegation of a broad civic event to such an association.
The Court might have rested there, with a straightforward holding about parades. This would usefully have confined Hurley to its facts; as a precedent it might have extended no further than from parades to expressive associations or activities with unified or holistic messages. But, in his opinion for the unanimous Court, Justice Souter compounded the case’s erroneous factual premises by indulging in some intriguing musings that produced broad dicta:
In the case before us, however, the Massachusetts law has been applied in a peculiar way. . . . Although the state courts spoke of the parade as a place of public accommodation, once the expressive character of both the parade and the marching GLIB contingent is understood, it becomes apparent that the state courts’ application of the statute had the effect of declaring the sponsors’ speech itself to be the public accommodation.[96]
Rather than a holding about a forced admission of one parade contingent into another, the Court here seemed to be talking about the unconstitutionality of applying public accommodations laws to expressive activities generally. From this dicta, it was a short step to Justice Thomas’s doctrinal principle: “When a public-accommodations law has the effect of declaring speech itself to be the public accommodation, the First Amendment applies with full force.” [97]
This principle, seemingly sound in pristine theoretical isolation, is in practice very problematic. What Thomas plainly means by “applies with full force” is that the O’Brien test is inapplicable to situations where “speech itself” is the public accommodation. But that is plainly wrong, and Hurley’s failure to apply the O’Brien test was itself a serious doctrinal error. When deciding whether to apply strict scrutiny or O’Brien’s intermediate scrutiny, as we have seen, the proper focus is on the nature of the law, not on the nature of the challenger’s speech. In Hurley, the Massachusetts public accommodations statute was certainly content neutral. But the application of that law to the parade failed at least one of the three prongs of the O’Brien test. Given the Court’s characterization of the facts—GLIB was attempting to insert an incompatible or diluting message into the message of an expressive association (the parade)—the application of the public accommodations law was arguably not “narrowly tailored” to the state’s interest in prohibiting exclusion of LGBT people from places of public accommodation. The government could instead have given GLIB its own parade permit without infringing the purported right of the Veterans Council to its message. More importantly, since the parade was, at least according to the Court, the Veterans Council’s own message, diluting that message did not “leave open ample alternative channels for communication.” That is the significance of the statement that “speech itself is the public accommodation.” If the message itself is the public accommodation, the law will always apply to it, in whatever “channel” it appears, leaving no alternative channels for the message. The Hurley Court made no effort to explain its failure to apply the O’Brien test.
In sum, Justice Thomas is wrong to declare that the O’Brien test does not apply in a case like Hurley. Rather, “[w]hen a public-accommodations law has the effect of declaring speech itself to be the public accommodation,” the O’Brien test should apply, but the law fails that test, because it does not “leave open ample alternative channels for communication.” Because the O’Brien test can accommodate activities that are speech in themselves, like a public address or a parade, or that generate speech, like newspapers, media consultants, and law firms, there is no need to create new doctrines to deal with those entities.
We come now to the error at the heart of 303 Creative. The majority held, in essence, that CADA was unconstitutional as applied to Lorie Smith’s “pure speech.” Therefore, it assumed, the O’Brien test could not apply. In rejecting O’Brien, the majority purported to rely on, or at least to be consistent with, current applicable doctrine, rather than creating a new exception of some sort.[98] But it is a stark logical error to say that O’Brien is inapplicable when a content-neutral law is challenged “as applied to ‘pure speech.’”
As noted above, the first question in any black-letter analysis of a free speech problem is whether the regulation at issue is content-based, or content-neutral. The 303 Creative majority nowhere mentions the content-based/content neutral distinction.[99] Had they done so, they would have been forced by precedent to apply the relatively permissive O’Brien test to CADA, a generally applicable, content-neutral law. Instead, they tried to finesse the question by characterizing Smith’s proposed wedding websites as “pure speech.” Justice Gorsuch’s majority opinion referred to Smith’s regulated activity as “pure speech” no fewer than six times.[100] How did the Court come to characterize Smith’s proposed web-designing as “pure speech”? And what are the legal implications of the “pure speech” category?
Plaintiff Lorie Smith is an internet-based website designer who operates her business under the name 303 Creative. Smith claimed that she wanted to expand her business to include wedding websites. As a socially conservative Christian opposed to same-sex marriage, Smith wanted to be free to refuse to make websites for same-sex marriages, even though she would perform all her other web design services for gay couples or individuals.
The facts concerning the expressive and conduct elements of Lorie Smith’s web designing were likely all along to have determined the outcome of the case; they certainly would determine the outcome of applying the O’Brien test. It is therefore mind-boggling why the state of Colorado stipulated to a set of facts that wound up giving the game away. The 303 Creative majority relied heavily on the stipulated facts, quoting the following in detail:
- Smith is “willing to work with all people regardless of classifications such as race, creed, sexual orientation, and gender,” and she “will gladly create custom graphics and websites” for clients of any sexual orientation.
- She will not produce content that “contradicts biblical truth” regardless of who orders it.
- Her belief that marriage is a union between one man and one woman is a sincerely held religious conviction.
- All of the graphic and website design services Ms. Smith provides are “expressive.”
- The websites and graphics Ms. Smith designs are “original, customized” creations that “contribut[e] to the overall messages” her business conveys “through the websites” it creates.
- Just like the other services she provides, the wedding websites Ms. Smith plans to create “will be expressive in nature.”
- Those wedding websites will be “customized and tailored” through close collaboration with individual couples, and they will “express Ms. Smith’s and 303 Creative’s message celebrating and promoting” her view of marriage.
- Viewers of Ms. Smith’s websites “will know that the websites are [Ms. Smith’s and 303 Creative’s] original artwork.”
- To the extent Ms. Smith may not be able to provide certain services to a potential customer, “[t]here are numerous companies in the State of Colorado and across the nation that offer custom website design services.”[101]
These stipulations enabled the Court to finesse the difficult question of whether Smith’s future wedding-website designing would indeed be expressive activity. By definition, Colorado’s lawyers agreed to the stipulations. Presumably, they too wanted a ruling on this factual basis, but their tactical decision was regrettable: it should have been foreseeable that the entire case would turn on the disputable fact that web design is expressive activity. And at least some of the confusions and doubts arise out of the case because the Court was not forced to consider the contours of speech in the realm of purportedly creative commercial activity.
In addition to these stipulations, the Court made much of the Tenth Circuit’s characterization of Smith’s activity as “pure speech.” According to Justice Gorsuch, the Tenth Circuit “acknowledged,”[102] “held,”[103] “said,”[104] and “recognized”[105] that Smith’s proposed website designs would be “pure speech.” Well, so what? Appellate courts do not make factual findings subject to deferential review by the Supreme Court. And in any case, the “pure speech” characterization is plainly a mixed question of fact and law that could be revisited de novo by the Supreme Court. But for all practical purposes, the Court treated the “pure speech” label as another stipulated fact:
The dissent claims that Colorado wishes to regulate Ms. Smith’s “conduct,” not her speech. Forget Colorado’s stipulation that Ms. Smith’s activities are “expressive,” and the Tenth Circuit’s conclusion that the State seeks to compel “pure speech.”[106]
By treating the Tenth Circuit’s legal conclusion as a de facto stipulation, the majority felt itself freed from any obligation to explain or justify that conclusion. These fact stipulations greatly eased the majority’s path to its result in 303 Creative; indeed, they supplied virtually every rebuttal to the dissent’s arguments. But they did not simplify the analysis or the precedential impact of the case; indeed, as we will see below, quite the contrary.
Coupled with its failure to engage in a strict scrutiny analysis, the 303 Creative majority opinion creates an atmospheric impression that “pure speech” is a particularly sacred form of speech—sacred for its “purity”—and therefore given categorical protection. This suggestion is akin to the occasional suggestion that “political speech” is more protected than other forms. (It isn’t.) To be sure, the phrase “pure speech” comes up frequently in First Amendment case law.[107] But if “pure speech” is a thing in free-speech doctrine, it is not a special doctrinal category of speech, and it does not even function as a category of speech at all—i.e., a category adopting the speaker’s vantage point. It is a category of regulation.
In 303 Creative, “pure speech” seems to function as a placeholder for “content-based”: laws that regulate “pure speech” are content-based. Although “pure speech” sounds a tad more elegant or evocative, it is actually less accurate than its functional synonyms, “content-based” or “content-regulatory,” because it is underinclusive of that category. Courts have not described “flag burning” as “pure speech” because, in a literal or common sense, it need not involve speaking at all. But it is unquestionably expressive activity, and as such is treated identically to “pure speech.” That’s because, again, the critical inquiry focuses on the intent of the regulation, and flag-burning prohibitions have been regarded as content-based.[108] In other words, all regulations of “pure speech” are content-based, but not all content-based laws are regulations of “pure speech.”
The 303 Creative majority deployed the “pure speech” characterization of Smith’s custom wedding websites to tell us that the O’Brien content-neutrality test should not apply. And indeed, the Tenth Circuit took this approach, applying strict scrutiny instead. The Gorsuch opinion muddied the waters by omitting any strict scrutiny analysis, a point to which we will return. But it is hard to see how the nature of Smith’s web-designing dispenses with the O’Brien test. In this regard, both the majority and dissent became confused by adopting the alleged speaker’s vantage point. Justice Sotomayor, dissenting, argued that O’Brien applied because Smith’s refusal to do wedding website design for same-sex couples was conduct. The majority’s answer, as we just saw above, was to assert the “pure speech” pseudo-stipulation. Both miss the point, at least under the applicable precedents. If CADA regulates “pure speech,” then it is content-based, and strict scrutiny applies. If CADA primarily regulates conduct, and only incidentally restricts speech—whether “pure” or “impure”—then O’Brien applies.
CADA defines a covered “place of public accommodation” as “any place of business engaged in any sales to the public and any place offering services, facilities, privileges, advantages, or accommodations to the public, including but not limited to any business offering wholesale or retail sales to the public.”[109] The law thus regulates what has been conventionally regarded as “conduct,” even though that conduct may have “speech elements”: sales, and the offering of products or services. The small number of controlling precedents have applied the O’Brien test to this type of law.[110]
Under the O’Brien test, Smith’s activity is selling web designs “to the public,” and her speech incident to this “conduct” is protected only if the law as applied in her case fails the O’Brien test. But it is difficult to make out that case for Smith—or would have been, absent the stipulations. Public accommodations laws have been held to meet the first two prongs of the O’Brien test: content-neutrality and narrow tailoring/substantial government interest. Under a conventional analysis, Smith’s only argument would be that CADA deprives her of “ample alternative channels for communication” for her opposition to same-sex marriage.[111] That argument is a clear loser: the First Amendment protects her right to convey anti-same-sex-marriage messages to her heart’s content outside the context of her sales to the public.
Enter “pure speech.” The argument, at least implicitly, is that public accommodations laws are unconstitutional as applied to “pure speech,” and that therefore O’Brien does not apply to purely expressive goods and services. But this argument makes no sense. All challenges to content-neutral laws involve the laws’ regulatory impact “as applied” to speech: if the regulated activity had no expressive element, there would be no basis for a First Amendment challenge. Putting it another way, the O’Brien test is designed for the exact purpose of handling as-applied free-speech challenges to content-neutral laws. In Clark v. Community for Creative Non-Violence, for example, the protesters sought to sleep overnight in a public park “to call attention to the plight of the homeless.”[112] They were denied a camping permit pursuant to the park’s general no-camping rule. The protesters did not claim—and could not plausibly claim—that the law was unconstitutional on its face, that is, in all its applications. Rather, they claimed the law violated the First Amendment as applied to their would-be camping, which was clearly expressive. The Court upheld the permit denial by applying the O’Brien test. The O’Brien test accounts for a content-neutral law’s restriction on speech, recognizing the special place of that speech in constitutional law with intermediate, albeit not strict, scrutiny.
The purported significance of labelling 303 Creative’s wedding website-making as “pure speech” is to suggest (implicitly) a distinction between it and the expressive camping in Clark or the draft-card-burning in O’Brien. One could argue that the latter have “elements of conduct,” whereas “pure speech” by definition does not. But this argument, too, is flawed. It suggests that “purely” expressive conduct, like flag-burning, is less protected than “pure speech”—an unsustainable proposition, in light of the application of strict scrutiny to flag-burning statutes, which are properly deemed content-based. It also entails an erroneous analytical shift away from the law by focusing entirely on the speaker. The purpose of a law like CADA does not change because of the regulated party’s activities: it is still a law aimed at regulating the sales of goods and services to the public, with an impact on speech in a particular case because the offered goods or services happen to be expressive. There is no reason to raise the level of scrutiny of a law that is content-neutral because of the expressive quality of the activity incidentally regulated. To do so, would be to eliminate the O’Brien test entirely, since O’Brien-type challenges always isolate the expressive element of the regulated activity.
Now enter Hurley. Here the argument is that the “pure speech” characterization means that “the public accommodation is itself speech.” But as we have seen, this argument is flawed as well. The “peculiar” application of a public accommodations law to a parade or other entirely expressive activity does not, again, change the content-neutral nature or purpose of the law. The Hurley theory is not a good argument for applying strict scrutiny (or, worse, a categorical prohibition on speech regulation), but rather an argument that the law, as applied, fails the O’Brien test: when applied to an entirely expressive activity, the public accommodations law does not leave open adequate alternative channels for conveying the message.
Perhaps that’s what Justice Gorsuch meant to convey by his repeated insistence that Smith’s someday wedding websites will be “pure speech”: that CADA would fail the O’Brien test as applied to Smith’s case. But that argument is hard to swallow. It requires us to accept the idea that Smith’s web-design business is like a parade or a public address: something that exists only to convey her own message. But that’s obviously not the case: her business is substantially, indeed predominantly, one that conveys the messages of her clients.
Now enter the stipulations. When pressed on this point by the dissent, Justice Gorsuch pointed to stipulations that “all of [Smith’s] services . . . are ‘expressive’”; that her business “conveys” her own “overall messages”; and that her “wedding websites . . . will ‘express Ms. Smith’s and 303 Creative’s message celebrating and promoting’ her view of marriage.”[113] These facts seem implausible as descriptions of a web design business., and had they been litigated, the courts would have been hard-pressed to find them. But the stipulations conveniently allowed the Court to sidestep the O’Brien test and, instead, to create what appears to be a new exception with very indefinite contours.
To be clear: although 303 Creative follows the logic of Justice Thomas’s Hurley theory, and that theory offers perhaps the most sensible way to doctrinalize a messy opinion, the 303 Creative opinion does not expressly articulate or embrace that theory. Instead, in its determination to make its decision look like a straightforward application of longstanding doctrine, the 303 Creative majority relies heavily on Hurley as a general compelled speech case, buttressed by West Virginia St. Bd. of Education v. Barnette[114] and Boy Scouts of America v. Dale.[115] The majority presents these three cases as though, in combination, they self-evidently establish CADA’s unconstitutionality as applied to Lorie Smith. And the Court takes them as standing for the proposition that the government cannot compel speech—period. (Again, the majority here undertook no strict scrutiny analysis.) Yet these three cases are applicable only atmospherically, or at very high levels of generality—Hurley and Dale both allowed expressive associations to discriminate against LGBT people—rather than as guides to resolving a tension between public accommodations laws and free-speech claims. The Court is saying something unprecedented in 303 Creative.
Like Hurley, Dale involved the application of public accommodations laws to an entity claiming the right to discriminate against gays. In Dale, a gay man fired from his position as Boy Scout troop leader sued under a New Jersey public accommodations law. The Court held that the Boy Scouts had the right to exclude Dale because his homosexuality was inconsistent with the Scouts’ associational message and purpose. Dale is thus even further afield than Hurley, since expressive activity was not directly involved. Together, the two cases stand for the proposition that associations formed for the purpose of expressing certain ideas could not be compelled to accept speakers or members whose participation would conflict with or “dilute” their message. But the two cases fall short of stating a principle that controls 303 Creative, where the Lorie Smith’s web design business was not created to express a particular message.
In Barnette, a WWII-era state law required children to salute the American flag and recite a pledge of allegiance at school on pain of expulsion from school and criminal prosecution of the children’s parents. The plaintiffs objected that the salute and oath violated their religious beliefs as Jehovah’s Witnesses, and the Court struck down the compulsory salute as compelled speech in violation of the Free Speech Clause.
Barnette is obviously distinguishable from 303 Creative in dispositive ways. First, the flag-salute law was not generally applicable or content-neutral. It had no other purpose than to force individuals to engage in government-mandated speech. In significant contrast, the public accommodations law at issue in 303 Creative, like any public accommodations law, was aimed at establishing equal dignity and participation in the public marketplace for members of formerly discriminated-against groups. Speech only becomes an issue incidentally, when the particular business-offering is expressive. The two laws are of very different types.
Second, the pledge of allegiance in Barnette was, in essence, the requirement of a personal oath. The speakers were compelled by law to express the beliefs in question as their own. In contrast, as the 303 Creative dissenters pointed out, the plaintiff Smith’s future creation of wedding websites would not require Smith to say or do anything that could reasonably be construed as expressing her personal views on same-sex marriage. (Here, again, a stipulation took this commonsense fact off the table by insisting that viewers “will know that the websites are” Smith’s.)[116]
The offense to Smith would not be personal speech, but personal association. Her personal efforts would be associated with, by assisting, an activity she disapproved of. Of course, the same could be said about the efforts put out by a wedding venue rented out to a same-sex couple, the resort where they would honeymoon, or even the store that might sell them streamers and balloons to decorate their party. This type of offense or injury is the very thing that public accommodations laws are designed to neutralize. Places of public accommodation are not permitted to refuse to associate with protected classes of people of whom their owners, managers, or employees happen to disapprove. As noted above, the only recognized exception to this principle is for expressive associations. Implicitly, the 303 Creative majority thinks it avoided this principle, rather than undermining it, by relying on the “stipulation” that Lorie Smith’s wedding websites would be “pure speech,” making her claim one of speech rather than association.
In the end, the precedents that purportedly dictate the decision in 303 Creative only loosely apply, at very high levels of generality. They provide the sort of background principles that courts typically lay out before getting to the actually dispositive issues. Barnette is emphasized for ambience. As the iconic compelled-speech case, Barnette offers powerful rhetorical advantages to the side that can claim it. That the objection to the speech was based on religious belief offers a further attraction to the 303 Creative majority. But the law in question in Barnette was aimed at speech and only speech; it was nothing like a public accommodations law. Hurley and Dale make the point that expressive associations need not associate themselves with LGBT messages or people. But Smith’s website business is not an expressive association, nor is it a purely expressive activity dedicated to a particular message, notwithstanding the stipulations. The mashup of these three precedents does not establish any particular free speech principle that governs the 303 Creative case. The 303 Creative decision is a novelty.
The majority states, “Generally, . . . the government may not compel a person to speak its own preferred messages.” That statement is correct, but only insofar as the qualifier “generally” is there, and is no more or less true than the statement, “Generally, the government may not abridge the freedom of speech.” The compelled speech cases do not impose a blanket prohibition on compelled speech or treat it as a special case warranting extra protection. For example, in Rumsfeld v. Forum for Academic and Institutional Rights (“FAIR”),[117] the Court held that compelled speech that was incidental to a law’s regulation of conduct is permissible if it meets the O’Brien test.[118] Like any other law abridging freedom of speech, the general prohibition on compelled speech is subject to exceptions.[119] We know this because laws abridging freedom of speech are upheld if they meet
the appropriate level of scrutiny: intermediate scrutiny if content neutral, strict scrutiny if not.
Black-letter First Amendment doctrine holds that free speech claims are subject to some level of scrutiny, which demands that governmental interests in regulating speech are taken into account. As noted above, content-neutral laws are reviewed under the O’Brien intermediate-scrutiny test. Content-based regulations can be upheld if they meet strict scrutiny. I have argued, as did the 303 Creative dissenters, that intermediate scrutiny should have applied to Smith’s claim. Justice Thomas’s Masterpiece Cakeshop concurrence argued that a public accommodations law must be reviewed under strict scrutiny when “speech itself is the public accommodation.” The petitioner Lorie Smith in 303 Creative argued that strict scrutiny applied. The majority Justices in 303 Creative said that “we align ourselves with much of the Tenth Circuit’s analysis”[120] and “part ways with the Tenth Circuit only when it comes to the legal conclusions that follow.”[121] The Tenth Circuit applied strict scrutiny. And by asserting that the “pure speech” quality of Smith’s activity made the O’Brien test inapplicable, the 303 Creative majority seemed poised to apply strict scrutiny.
Yet, the majority opinion does not apply any level of scrutiny. The 303 Creative majority offered a general acknowledgment “that governments in this country have a ‘compelling interest’ in eliminating discrimination in places of public accommodation.”[122] Yet the Court did not apply strict or intermediate scrutiny in any recognizable form. Which is to say that the 303 Creative majority failed to consider the strength of the state’s legitimate interest in its antidiscrimination law. The only state interest purportedly underlying CADA that was even mentioned by Justice Gorsuch was the patently illegitimate “goal of eliminating views [the government] does not share”—i.e., “the coercive elimination of dissenting ideas about marriage.”[123] The majority refused to consider that Colorado had a substantial interest in preventing discrimination against same-sex married couples. Nor did it consider that public accommodations laws are not generally held to be aimed at the suppression of speech, even though they prevent shopkeepers from posting “whites only” or “no gays” signs in their storefront windows. The majority simply said, citing Hurley and Dale, that “this Court has held, public accommodations statutes can sweep too broadly when deployed to compel speech. . . . When a state public accommodations law and the Constitution collide, there can be no question which must prevail.”[124]
What are we to make of that bald and breathtakingly broad assertion in the second sentence? Do plausible free speech claims now always defeat public accommodations laws? Even the slightly more specific first sentence is alarming: Is a public accommodations law categorically prohibited from compelling compliance if compliance can be characterized as expressive? Depending on how that assertion is limited, it has the potential to undermine public accommodations laws.
Conceivably, the omission of a strict scrutiny analysis was intended, not as a doctrinal statement that the prohibition was categorical, but as a workaround to avoid a difficult question involved in the application of strict scrutiny. Does a state have a compelling interest in prohibiting sexual orientation discrimination, when that category has to date received only rational basis review (albeit, rational basis “with bite”)?[125] In other words, must there be some parallelism between the degree of protection against discrimination afforded by the Constitution and the strength of the governmental interest when weighed against free speech interests? CADA also prohibits discrimination on the basis of “disability,” “gender identity,” “gender expression,” and “marital status.”[126] Disability, like sexual orientation, has received only rational basis (“with bite”) review. Gender identity and expression are a new culture-war front, and marital status is constitutionally protected, at most, as a derivative of sex discrimination. Perhaps the Court was not prepared to decide whether combating discrimination on these grounds is a compelling interest. It is not even clear whether prohibition of sex discrimination is a “compelling interest” in the strict-scrutiny sense. Roberts v. U.S. Jaycees, the case cited for this proposition, may have used the phrase “compelling interest” in the special (and confusing) way the term is used in the O’Brien test, to refer to an interest that meets intermediate, not strict, scrutiny.[127]
Arguably, it was incumbent on the 303 Creative Court to face these issues, since clearly some level of scrutiny—intermediate or strict—should have applied, and the Court’s failure to do this admittedly difficult work was irresponsible. On the other hand, doing so would have complicated an already complex and messy decision. And if the majority was determined to resolve the case in Lorie Smith’s favor, a thorough analysis of the governmental interests for a strict scrutiny inquiry might only have made things worse. The strength of the state’s interest in combatting sexual orientation discrimination remains unresolved.
Why was this case even decided? What was the pressing First Amendment problem that required the court to resolve it at so high a price? The high price comes in the form of doctrinal confusion in raising, but not answering, difficult and troubling questions about the constitutionality of public accommodations laws, about the rights of businesses open to the public to engage in discriminatory “pure speech,” and about the role of religious beliefs in free speech claims.
There is, to be sure, a knotty theoretical problem at the heart of the case, pitting free speech against public accommodations laws. Does someone who speaks or creates for others as a living have the right to turn down a commission to which they object? Do public accommodations laws infringe that right? This issue was pointedly raised in the majority’s more compelling hypotheticals. By privileging public accommodations laws in this situation,
[t]he government could require “an unwilling Muslim movie director to make a film with a Zionist message,” or “an atheist muralist to accept a commission celebrating Evangelical zeal,” so long as they would make films or murals for other members of the public with different messages. Equally, the government could force a male website designer married to another man to design websites for an organization that advocates against same-sex marriage.[128]
One might add to this the expressive services of lawyers: must they take clients who advocate positions that violate their beliefs?[129]
Ironically, Justice Gorsuch ridiculed the dissent for going “adrift on a sea of hypotheticals” when it was the majority that raised these hypotheticals—and they were the most persuasive aspects of the majority’s argument! Yet these persuasive examples were just that: hypothetical. The only “live” controversy was barely live: a cooked-up case based on stipulated facts about a purportedly expressive activity that Smith, as of this writing, had yet to undertake, six years after filing the case and seven months after winning her case in the Supreme Court.[130]
Figure 1. Smith’s 303 Creative website (March 28, 2024).
303 Creative was legal performance art: not a genuine dispute, but a creation to give the courts an opportunity for a do-over of Masterpiece Cakeshop. Some commentators objected that Smith’s future plan to undertake a wedding web-design business meant that the case was unripe, or that she lacked standing. But the more relevant objection is worse than that. It may be that Smith never intended to make wedding websites, but that religious activist lawyers recruited her as a conservative Christian web designer who could say that she wanted to make wedding websites in order to raise the same-sex marriage objection. In the initial suit, in order to show a live controversy, Smith alleged that she had been contacted by a same-sex couple to make a wedding website; but it turns out that that contact was bogus, and that the alleged prospective client was a heterosexual married man.[131] Smith dropped this allegation before the Supreme Court.
To be clear, I am not objecting either to pre-enforcement challenges or to cause lawyers or public interest entrepreneurs finding their plaintiffs rather than the other way around. Both practices have their place in obtaining judicial review of pressing policy questions. Rather, my point is that the thinness of Smith’s claim to a live controversy demonstrates the difficulty of finding people in Smith’s situation to serve as plaintiffs. And that, in turn, is strong circumstantial evidence that the problem is unreal—or at least, far from a pressing one.
The Supreme Court often allows purported legal problems to percolate to see if they become real problems—and they might well have done so here. To be sure, the handful of cases of this type had produced a circuit split between the Eighth and Tenth Circuits.[132] But this split was just one year old when the Court granted cert, and the Court has sometimes let circuit splits linger for years.[133] And the pressure to resolve a circuit split arises, not from the split itself, but from the pressure of numerous inconsistent adjudications in the lower courts. Here, there does not appear to have been such a press of cases, or even an actual, as opposed to hypothetical problem of creative professionals, muralists, and filmmakers being forced to accept commissions violating their beliefs, or being sued under public accommodations laws for following their consciences.[134] The problem is entirely one involving objections to same-sex marriage by a tiny handful of people who provide wedding-related services. And there are not many of these cases; other than Masterpiece Cakeshop and 303 Creative, I have found only eight federal and state cases since 2010 in which an arguable “creative professional” challenged a public accommodations law to assert their right to oppose same-sex marriage.[135]
The hypothetical problem of the public accommodation-conscience conflict might well never have materialized as a significant issue due to private self-sorting. The Zionist group is unlikely to seek out the Muslin director to make its film; the evangelical client is likely to seek out a muralist known to be friendly to religion; and same-sex couples are likely to avoid bakeries with Bible cakes on display in the front window. (As for lawyers turning down clients with offensive cases, ‘the lawyer’s ethical obligation to zealously represent her client gives the lawyer a ready out, on the ground of a conflict of interest.) Disappointed customers are unlikely to turn to law, and in the exceptional instances when they do—like Craig’s and Mullins’s complaints—they need not be transformed into broad constitutional rules. This state of affairs is imperfect, to be sure, but legal ordering is filled with imperfect enforcement of existing laws and unrealized potential rights.
Instead, the Court chose to make a federal case out of this—to weigh in on a grand scale. And in doing so, they have made the problem worse. For most of the messages that would be refused by would-be suppliers of expressive services, even hypothetically, do not reflect categories protected by public accommodations laws. Hate groups, for example, are not protected by such laws. Refusals are thus legally unproblematic. But by raising hypotheticals that equate the refusal of offensive, religiously based messages with discrimination against religion, the Court creates a potentially widespread problem that didn’t previously exist. Now everyone in this “creative professional” category needs the speech protection from public accommodations laws afforded by 303 Creative. The “problem” solved so messily by the 303 Creative Court is one that may exist entirely in the rarified world of test-cases.[136]
We are now left to wonder where the Court is going with this. What does 303 Creative mean for LGBTQ rights, for the First Amendment’s Free Speech and Free Exercise clauses, and for antidiscrimination law? There is a broad range of possibilities, depending on how one interprets the decision as a precedent, and on what inferences one draws about the internal dynamics of the Court from the choices it made about how to structure the decision.
Fully understanding 303 Creative requires attention to a significant issue the Court did not address. The Court decided the case on the basis of the Free Speech Clause of the First Amendment, without reliance on the Free Exercise Clause. This in itself is not new: plaintiffs in other “compelled speech” cases, such as Barnette and Wooly v. Maynard,[137] alleged that the compelled speech in question violated their religious beliefs.
What happened to Smith’s free-exercise claim? The second cause of action in Smith’s complaint alleged a “Violation of Plaintiffs’ First Amendment Right to Free Exercise of Religion.”[138] The lower courts decided both questions. And the two questions raised by the petition for certiorari both involved the Free Exercise Clause. The first question presented was “Whether applying a public-accommodation law to compel an artist to speak or stay silent, contrary to the artist’s sincerely held religious beliefs, violates the Free Speech or Free Exercise Clauses of the First Amendment.”[139] The second question was “Whether a public-accommodation law that authorizes secular but not religious exemptions is generally applicable under [Employment Division v.] Smith, and if so, whether this Court should overrule Smith.”[140]
But the Court granted certiorari “limited to the following question: Whether applying a public-accommodation law to compel an artist to speak or stay silent violates the Free Speech Clause of the First Amendment.”[141] The limited cert grant is noteworthy in two respects. First, it appears that Lorie Smith was pre-destined to win the case, and to win on something like Justice Thomas’s Hurley theory from Masterpiece Cakeshop. Had that been in doubt, there would have been some incentive to include the Free Exercise issue. Second, we can infer that Employment Division v. Smith is safe—for now.
Employment Division v. Smith,[142] gets remarkably little love from conservatives, considering that it was authored by Justice Scalia in his adamant style. Put another way, Smith’s shaky status on the current Court is a fascinating marker of how conservative politics on the Court have shifted. Smith was handed down before the Court’s conservatives made common cause with religious conservatives on issues other than abortion. In Smith, two native American men, Alfred Smith and Galen Black, were fired from their jobs when their employer, a drug rehabilitation organization, learned that they had ingested peyote in a religious ceremony. The peyote use violated the state’s drug laws, and when Smith and Galen applied for unemployment benefits, they were denied them based on the purported good cause for their firing. The Court held that an incidental infringement on religious exercise resulting from a generally applicable law that did not target religion would be reviewed under the rational basis test, rather than strict scrutiny. Smith effectively overruled Sherbert v. Verner[143] and other precedents that had applied strict scrutiny in such cases.
Justice Scalia’s opinion was grounded on the eminently sensible proposition that, “To make an individual’s obligation to obey [a generally applicable] law contingent upon the law’s coincidence with his religious beliefs, except where the State’s interest is compelling” would in effect “permit[] him, by virtue of his beliefs, to become a law unto himself.”[144] This would be the inevitable result of the Court’s wholly appropriate unwillingness to scrutinize the bona fides of purported religious practices,[145] coupled with the strictness of strict scrutiny.
If the “compelling interest” test is to be applied at all, then, it must be applied across the board, to all actions thought to be religiously commanded. Moreover, if “compelling interest” really means what it says (and watering it down here would subvert its rigor in the other fields where it is applied), many laws will not meet the test. Any society adopting such a system would be courting anarchy, but that danger increases in direct proportion to the society’s diversity of religious beliefs, and its determination to coerce or suppress none of them. Precisely because . . . we value and protect that religious divergence, we cannot afford the luxury of deeming presumptively invalid, as applied to the religious objector, every regulation of conduct that does not protect an interest of the highest order. The rule respondents favor would open the prospect of constitutionally required religious exemptions from civic obligations of almost every conceivable kind—ranging from compulsory military service, to the payment of taxes, to health and safety regulation such as manslaughter and child neglect laws, compulsory vaccination laws, drug laws, and traffic laws; to social welfare legislation such as minimum wage laws, child labor laws, animal cruelty laws, environmental protection laws, and laws providing for equality of opportunity for the races. The First Amendment’s protection of religious liberty does not require this.[146]
Smith thus stands for the general proposition that religious exemptions need not be carved out of reasonable laws of general applicability—including antidiscrimination laws. Smith is thus anathema to the religious element of movement conservatism.
In Masterpiece Cakeshop, where the petitioner did not ask for Smith to be overruled, Justice Gorsuch dropped a hint in his concurrence that the Court might be willing to do so, noting that “Smith remains controversial in many quarters.”[147] That hint was soon taken up. Fulton v. City of Philadelphia[148] involved a challenge to the city’s cessation of foster-care referrals to Catholic Social Services because that organization refused to certify same-sex couples as foster parents. The cert petition filed in July 2019, one year after Masterpiece Cakeshop, asked “Whether Employment Division v. Smith should be revisited.”[149] The Fulton Court decided that it didn’t need to reconsider Smith because Philadelphia’s policy was not neutral toward religion and therefore not a generally applicable law within the meaning of Smith.[150] Justices Alito, Gorsuch, and Thomas asserted that “The Court granted certiorari to decide whether to overrule Smith” and argued vigorously that it should have.[151] But the decisive view was that of Justice Barrett, joined by Justice Kavanagh. Despite believing Smith to have been wrong “as a matter of text and structure,” Justice Barrett was “skeptical about swapping Smith’s categorical antidiscrimination approach for an equally categorical strict scrutiny regime, particularly when this Court’s resolution of conflicts between generally applicable laws and other First Amendment rights—like speech and assembly—has been much more nuanced.”[152] There it is: Smith would survive because the other First Amendment rights offered more workable pathways to constitutionalize religious exemptions to generally applicable laws.
The lawyers for Lorie Smith in 303 Creative either missed this signal, or believed that they could shift Barrett and Kavanagh. Their cert petition, which asked, as we have seen, “whether Smith should be overruled,” was filed in September 2021, three months after Fulton was handed down. But the Court’s cert grant in 303 Creative made clear that reconsidering the Smith case was no longer on the table. Nor would the Court evaluate religious exemptions from generally applicable laws on a “free-speech/free-exercise” “hybrid rights” theory.[153] Religious exemptions to generally applicable laws would be channeled through speech and association, which would allow for more “nuanced” treatment. 303 Creative is presumably the “nuanced” treatment.
What did Justice Barrett have in mind as a “nuanced” holding when she stated her preference for the free speech route to religious exemptions from public accommodations laws? Such a holding may be exemplified by two recent efforts to harmonize 303 Creative with existing doctrine.
Professor David D. Cole suggests that 303 Creative “should have minimal impact on the enforcement of public accommodations and antidiscrimination laws” if properly interpreted.[154] According to Cole, the free speech right prevails against a public accommodations law only if “(1) a business objects only to expressing a particular message for anyone, not where it objects to serving certain customers because of their identity; and (2) the state’s interest in requiring the business to provide the service is the suppression of disfavored ideas.”[155] Cole is probably right that this interpretation of 303 Creative would limit its reach, and thereby its damage to public accommodations laws.[156] And his two-prong test is a plausible, lawyerly effort to harmonize the decision with existing doctrine. But Cole’s interpretation of the case is debatable. Because the state’s interest underlying CADA was not “the suppression of disfavored ideas,” Smith should have lost her case under Cole’s test. It is problematic to construe the holding of a case as requiring the direct opposite of the result it reached. To find for the plaintiff, under Cole’s test, one would have to interpret a generic public accommodations law—that is, all public accommodations laws—as intended to suppress disfavored ideas. And that is exactly what the majority opinion in 303 Creative seems to do: it treats the application of a public accommodations law to an expressive business as though it were based on a per se interest in “the suppression of disfavored ideas.” The majority offers no basis to distinguish between that purportedly illegitimate interest and a legitimate interest in suppressing discrimination against protected groups. Indeed, Justice Gorsuch’s conclusion that CADA’s purpose as applied to wedding websites was to “eliminate disfavored ideas”[157] suggests that any application of a public accommodations law to an expressive business would meet this prong of Cole’s test for unconstitutionality.
Professor Dale Carpenter offered a similar interpretation of 303 Creative in a blog post on The Volokh Conspiracy shortly after the decision came down.[158] Carpenter argues that the case should be understood to create a rule that “a vendor cannot be compelled by the government (1) to create customized and expressive products (whether goods or services) that constitute the vendor’s own expression; (2) where the vendor’s objection is to the message contained in the product itself, not to the identity or status of the customer.”[159] This, in fact, was the doctrinal rule proposed in an amicus brief filed by Carpenter along with Professor Eugene Volokh, who together have been arguing for this test for a decade.[160]
The rule, if it is the rule, would place reassuring limits on the potentially sprawling scope of the case. First, like Cole’s test, it would be limited to compelled speech. Second, as Carpenter explains, the vast majority of commercial products and services would not qualify under the first prong of the test. “Ordinary commercial products,” which constitute “almost all of the products we buy,” writes Carpenter, “are neither customized nor expressive.” Moreover, Carpenter asserts, “most customized products are not expressive” (e.g., made-to-order hamburgers) even if the maker considers himself to be an artist. As Carpenter and Volokh have argued, “not all efforts that produce aesthetically pleasing products” are protected speech.[161] And finally, “most expressive products are not customized,” such as books sold in a bookstore—the speech is complete before the sale.
The second part of the test, which must also be satisfied, emphasizes that “it will not suffice to say that the very fact of the sale alone sends a message the vendor does not want to send.” Instead, the message must be “contained in the product itself.” For example, Carpenter says, a business can’t refuse to sell party decorations to a transgender person based on the fact that the decorations will be used to celebrate that person’s gender transition.[162] This, though Carpenter doesn’t say so, is simply a corollary of the customization requirement. Illustrative of the contours of their test, Carpenter and Volokh supported Smith’s claim in 303 Creative, and the earlier claim of a wedding photographer in Elane Photography v. Willock, but filed an amicus brief in support of Colorado and against the baker in Masterpiece Cakeshop.[163]
Of course, there will be disagreements over whether a product is “customized and expressive” and “the vendor’s own expression.” In 303 Creative itself, it is highly debatable whether a custom wedding website is the vendor’s own expression, particularly if the original material (e.g., text and photographs) are supplied by the clients and the web designer merely arranges them nicely and perhaps adds prefabricated artwork—which seems to be a very likely process for such things. Website coding and layout are dubious candidates for “the vendor’s own expression.” This problem was finessed in 303 Creative by the unfortunate stipulated facts, which conceded that the web design in this case was “pure speech.” Had the facts been actualized, rather than hypothetical, and litigated rather than stipulated, there’s a good chance that plaintiff Smith would have failed the Carpenter-Volokh test. In any case, fact-based definitional disputes such as this are nothing the law can’t handle.
But 303 Creative did not actually adopt the Carpenter-Volokh test, making it uncertain whether this is how the case will be read in the future. For starters, the Court did not apply strict scrutiny, even though Carpenter and Volokh argued that it applied.[164] In contrast, a different First Amendment Scholar’s brief tried to argue that compelled speech was categorically barred,[165] and the Court may well have adopted that view.[166] Moreover, the Court’s rationale in 303 Creative includes broad language suggesting that conforming one’s “expressive” business activities—not “expressive and customized” activities—to the public accommodations laws unconstitutionally compels that person to “speak [the government’s] preferred message.”[167] And there are hints that at least some Justices would reject the Carpenter-Volokh test. As Andrew Koppelman observes, the Thomas concurrence in Masterpiece Cakeshop, which Gorsuch joined, took a very broad view of expressive business activity, one that included made-to-order baking.[168] Yet Carpenter and Volokh argued that cake-baking did not meet their test. Justice Thomas, in Masterpiece Cakeshop, began his concurrence by emphasizing that Phillips “had refused to create a custom wedding cake” but then dropped that element and implied that the cake’s expressiveness alone made it protected speech.
The Carpenter-Volokh test does offer a plausible way to make sense of, and place reasonable limits on, an otherwise very messy and potentially far-reaching opinion. The Carpenter-Volokh test treats the key stipulated facts, that the web designs were “customized” and “expressive,” as essential to the holding. But the stipulated facts themselves present problems. Lower courts and litigants, and perhaps even the Supreme Court, are likely to forget, or ignore, that fact stipulations in one case are not binding in others, and that the 303 Creative Court took the stipulations arguendo. It is thus likely that a web designer will be treated, going forward, as an archetypal “creative professional,” even though that conclusion is dubious. Carpenter doesn’t come to terms with the problem that the fact stipulations allowed the Court to gloss over the difficult questions about what makes a business activity expressive. The category of expressive products and services falling within the Carpenter-Volokh exception to public accommodations laws may thus be much broader than Carpenter suggests.
There is a much bigger problem presented by 303 Creative that Professor Carpenter does not consider. There is no particular reason for religious conservatives to be satisfied with a “victory” limited to the contours outlined by Carpenter and Volokh. Why should only “creative professionals” enjoy this particular exemption from the public accommodations laws? That is a somewhat elitist place to draw the line, but the constituency that the justices arguably sought to reward in 303 Creative is not a predominantly professional, college-educated class. It is inevitable that the religious conservative community will come forward with claims of conscience by its less elite members. When this happens, will the Court’s 303 Creative majority have the nerve to turn them away, arguing that their “coerced” compliance with antidiscrimination laws is not “expressive”?
Here, it is worth noting that this bloc of Justices has a history of playing the populist demagogue from time to time, particularly on the issue of LGBT rights and same-sex marriage. In Obergefell, Chief Justice Roberts’s dissent excoriated the “five lawyers” in the majority “who happen to hold commissions authorizing them to resolve legal disputes according to law.”[169] Justice Scalia was (characteristically) far more blunt about the subtext of these “five lawyers”:
[T]he Federal Judiciary is hardly a cross-section of America. Take, for example, this Court, which consists of only nine men and women, all of them successful lawyers who studied at Harvard or Yale Law School. Four of the nine are natives of New York City. Eight of them grew up in east- and west-coast States. Only one hails from the vast expanse in-between. Not a single Southwesterner or even, to tell the truth, a genuine Westerner (California does not count). Not a single evangelical Christian (a group that comprises about one quarter of Americans), or even a Protestant of any denomination.[170]
Like the old references to “New York lawyers” to refer to Jewish lawyers, Scalia’s rant here is a barely disguised euphemism for “coastal, liberal elites.” One could go on.[171]
And why should the religious right be satisfied with opposition to same-sex marriage? Recall that the message William Jack wanted iced onto his cake said “Homosexuality is a detestable sin.”[172] The 303 Creative majority implied that this viewpoint, supported as it was by a Bible quotation, was deserving of respect. And why stop there? White supremacy can be presented as a “sincere religious belief,” and the Court has correctly taken the position that it is not able to assess the bona fides of claimed religious beliefs.[173] Perhaps business people will come forward demanding exemptions from antidiscrimination laws to act on their view that mixing of the races “contravenes the will of God.” On what basis will the 303 Creative majority dismiss such claims, as the Court once did, as “patently frivolous”?
303 Creative’s expansive potential invites tests of its limits along two parameters: speaker and subject matter. The Carpenter-Volokh interpretation limits the damage of 303 Creative only to the extent that reviewing courts apply rigorous views of what sort of business activity counts as expressive. That element tells us what speakers are entitled to this new exemption. But future cases will also test the subject matter parameter. The Carpenter-Volokh test and the 303 Creative opinion itself both purport to rest on content- and viewpoint-neutral free-speech principles. Nothing in them, therefore, limits the “creative professionals” exception to antidiscrimination laws to the “acceptable” (to those Justices) realm of opposition to same-sex marriage. Any sort of homophobic, sexist, or racist belief is fair game. The lower courts will be faced with test cases putting pressure on the limits of 303 Creative along both parameters.
The broadest and most alarming interpretation of 303 Creative is that it is an incremental first step toward a First Amendment assault on LGBTQ rights and perhaps even on antidiscrimination laws more broadly. Here, the Court’s reliance on the Free Speech Clause rather than the Free Exercise Clause is concerning rather than reassuring. To be sure, the Court has previously rejected arguments to recognize First Amendment exemptions for racially exclusive business practices, whether on speech or religious grounds. But the Gorsuch opinion includes disturbing language to the effect that CADA “seeks to compel [pro-same-sex marriage] speech in order to excise certain ideas or viewpoints from the public dialogue,” and that “the coercive elimination of dissenting ideas about marriage constitutes Colorado’s very purpose in seeking to apply its law to Ms. Smith.”[174] Here, Justice Gorsuch was quoting the Tenth Circuit opinion—another quasi-stipulation or finding that the majority felt no need to unpack or explain. But in quoting it, Justice Gorsuch implicitly endorsed that view. What are we to make of that?
And then there are all the references to “pure speech.” Expressive services or products are not the only “pure speech” that might be exercised by businesses. A message in a shop window or a screed on a business website saying that same-sex marriage, or homosexuality—or for that matter, interracial marriage or racial equality—are (to quote the Piggie Park defendant) “against the will of God” would seem to be pure speech, and to convey the same message as “Gays not welcome” or “Whites only.” The Court’s bland acceptance of Smith’s stipulated assertion that her web designs “contribute to the overall message her business conveys”[175] is deeply troubling. It invites businesses to characterize themselves in expressive associational terms, such as “Christian bakery” or “anti-LGBT web-design” or “white supremacy copy shop,” whether or not their services are expressive, in order to claim that their business as a whole conveys a discriminatory message. True, the Court tried to reassure us with a very general bromide about “the vital role public accommodations laws play in realizing the civil rights of all Americans.”[176] But such ugly test cases would draw the Court into making difficult refining characterizations to the effect that some businesses are inherently of a type that must “serve all comers” in order to prevent the resegregation of public accommodations through purported free speech self-description. Even the concept of distinguishing the work of “creative professionals” and their “customized and expressive” work product from that of all other business owners—even if the Court chooses to enforce those limitations—may be too spongy and delicate to do the analytical work necessary to stop the slide down this slippery slope.
The narrowest interpretation of 303 Creative as a precedent, and the one I hope for as a supporter of LGBTQ rights and non-theocratic government, is that the Court’s work is done in this area. Specifically, the Court is (perhaps) now satisfied that it has made good on the “promise” in its Obergefell disclaimer to respect religious conservatives’ persistent refusal to accept same sex marriage. The small subset of business owners who can be categorized as “creative professionals” are entitled to a single, limited exemption from public accommodations laws, as long as their anti-gay bigotry is (1) expressed courteously, (2) framed as religious opposition to same-sex marriage, and (3) accompanied by an assurance that they will otherwise extend all services to LGBTQ customers that do not entail assisting same-sex weddings or endorsing same-sex marriage. No further inroads into antidiscrimination laws under the rubrics of Free Speech or Free Exercise are contemplated—or so one hopes.
Ultimately, the future impact of 303 Creative boils down to whether the case is to be a precedent about free speech writ large, or merely a safe-harbor for opponents of same sex marriage to express themselves. As I have suggested, the Court’s motivation for creating a solution to an almost entirely hypothetical problem, was based on constitutional politics: the felt need to operationalize the promise in Obergefell to reward religious conservatives with a compensatory win. I think it unlikely, therefore, that 303 Creative is intended as the opening salvo of a free speech assault on antidiscrimination laws.
Nevertheless, the muddy doctrinal takeaways of 303 Creative invite numerous ugly test cases to explore the limits of free speech exceptions to public accommodations laws. This creates much potential for damage in lower court decisions before the Supreme Court speaks again. When it does, the Court will have to walk back some of the broader assertions and implications of 303 Creative in some way. The most obvious means of damage control is a judicial fiat limiting the 303 Creative precedent to its facts to some extent. The Court might hold that the exemption applies only to “creative professionals” and thereby confine the realm of arguable claims to arguments about whether a person is a creative professional. The Court might also construe the case as limited to “compelled speech”—that is, a legal requirement that purported speakers engage in conduct that implies support for same-sex marriage—but does not extend to the protection of affirmative expressions of disapproval or religiously inspired anti-gay or other bigotry beyond the minimum message necessary to convey that creative services will not be provided.
Doctrine may have been the last thing on the majority’s mind here. The only thing that is clear about this decision is that the Court was intent on protecting the last refuge of acceptable (to some) anti-gay bias: the religiously based disapproval of same-sex marriage. The doctrinal implications of this move—for free speech, free exercise, and antidiscrimination law—are left for another day.
Appendix A
The following brief empirical research, admittedly far from definitive, tends to support the inference asserted in the text. A LEXIS search for cases since 2010 using the search terms “public accommodation and free speech” yielded 263 hits, the vast majority of which were false positives. There were only 19 discrete cases in which a public accommodations law was challenged on free speech grounds. All 19 (which include Masterpiece Cakeshop and 303 Creative) involved religious-based objections to sexual orientation. While court cases in general represent only the tip of the iceberg of legal disputes, the fact that all such cases were of this type is consistent with my suggestion that the other hypotheticals—those involving secular claims or not involving sexual orientation—may have little or no basis in reality. Conservative religious legal organizations represented the religious claimant in at least 14 of the cases, thirteen of which were the Alliance Defending Freedom. (Counsel were not identified on LEXIS in two of the cases, and a private attorney represented the remaining religious objector.)
Of the 19 cases, five involved objections to extending non-speech services to transgender people, or to organizational policies requiring the use of traditional pronouns. Thus, aside from Masterpiece Cakeshop and 303 Creative, there have been only 12 federal and state court cases involving the kinds of claims at issue in those two cases. All 12 involved wedding services.
Of those 12, seven fell into the Masterpiece Cakeshop “enforcement” pattern in which a public accommodations complaint had been filed against a business that refused to provide services for a same-sex wedding. One of these was the Scardina test case against Masterpiece Cakeshop itself (see note 136 above). Five involved pre-enforcement challenges, in the 303 Creative pattern. Three cases involved refusal to provide a wedding venue, which should not qualify as expressive conduct under 303 Creative.
Appendix A.
| Case | Cite | Type | Wedding Service |
Counsel |
| Elane Photography, LLC v. Willock | 309 P.3d 53 (N.M. 2013) | Enforcement | Photography | ADF |
| Knapp v. City of Coeur D’Alene | 172 F. Supp. 3d 1118, 2016 WL 1180168 (D. Ida. 2016) | Pre-enforcement | Wedding venue | ADF |
| Matter of Gifford v. McCarthy | 137 A.D.3d 30, 42, 23 N.Y.S.3d 422 (N.Y.App. Div. 2016) | Enforcement | Wedding venue | ADF |
| Klein v. Or. Bureau of Lab. & Indus. | 289 Ore. App. 507, 410 P.3d 1051 (2017) | Enforcement | Custom cake | Liberty Institute |
| Country Mill Farms, LLC v. City of E. Lansing | 2017 WL 11444048 (W.D. Mich. 2017) | Enforcement | Wedding venue | ADF |
| Brush & Nib Studios, LC v. City of Phoenix | 247 Ariz.269, 448 P.3d 890, (Ariz. 2019) | Pre-enforcement | Custom invitations |
ADF |
| State v. Arlene’s Flowers, Inc | 193 Wn.2d 469 *; 441 P.3d 1203 (2019) | Enforcement | Floral arrangement |
ADF |
| Telescope Media Grp. v. Lucero | 963 F.3d 740 (2019) | Pre-enforcement | Videography | ADF |
| Dep’t of Fair Emp. & Hous. v. Superior Ct. | 54 Cal. App. 5th 356 *; 269 Cal. Rptr. 3d 9 **; 2020 | Enforcement | Custom cake | Freedom of Conscience Defense Fund |
| Chelsey Nelson Photography LLC v. Louisville/Jefferson Cty. Metro Gov’t | 479 F. Supp. 3d 543 (W. D. Ky. 2020) | Pre-enforcement | Photography | ADF |
| Emilee Carpenter, LLC v. James | 575 F. Supp. 3d 353 (W.D. N.Y. 2021) | Pre-enforcement | Photography | ADF |
| Scardina v. Masterpiece Cakeshop | 528 P.3d 926 (Colo. App. 2023) | Enforcement | Custom cake | Private attorneys |
[1]. Frederick W. & Vi Miller Professor of Law, and Vilas Distinguished Achievement Professor, University of Wisconsin Law School. Copyright © 2024 by David S. Schwartz. I am grateful to Andrew Coan, Anuj Desai, Toni Massaro, and Jill Hasday for their comments, and to Nicholas Surprise for research assistance.
[4]. Or for LGBT rights. See infra note 136.
[5]. See infra notes 130–131 and accompanying text.
[6]. See Alexander Bickel, The Supreme Court, 1960 Term—Foreword: The Passive Virtues, 75 Harv. L. Rev. 40 (1961).
[7]. Masterpiece Cakeshop, Ltd. v. Colo. Civ. Rts. Comm’n, 584 U.S. 617 (2018).
[10]. For an excellent attempt at this, see David D. Cole, “We Do No Such Thing”: 303 Creative v. Elenis and the Future of First Amendment Challenges to Public Accommodations Laws, 133 Yale L.J.F. 499 (2024). According to Cole, “to make sense of the Court’s doctrine, one must seek to harmonize this decision, recognizing a First Amendment exemption, with a long line of cases rejecting seemingly similar claims.” Id. at 501. Must one? There are other ways to make sense of a case than to harmonize it with existing doctrine. Some decisions are just not harmonious.
[13]. 388 U.S. 1, 11–12 (1967).
[14]. Newman v. Piggie Park Enters., 390 U.S. 400, 402 n.5 (1968).
[15]. 600 U.S. 181 (2023). See, e.g., id. at 208 (finding affirmative action is a race distinction that is “odious to a free people whose institutions are founded upon the doctrine of equality”); id. at 232 (Thomas, J., concurring) (identifying affirmative action as a “discriminatory wrong[]” on par with Jim Crow).
[16]. Obergefell, 576 U.S. at 741 (Alito, J., dissenting). Justice Alito has recently reprised this complaint. Mo. Dep’t of Corr. v. Finney, 218 L. Ed. 2d 69 (2024) (Alito, J., dissenting from denial of certiorari).
[23]. Obergefell v. Hodges, 576 U.S. 644, 679–80 (2015).
[25]. See infra text accompanying notes 64–74.
[26]. See infra note 64 and accompanying text.
[28]. See infra text accompanying note 143.
[30]. For example, Title VII of the 1964 Civil Rights Act, the employment discrimination title, “shall not apply . . . to a religious corporation, association, educational institution, or society with respect to the employment of individuals of a particular religion to perform work connected with the carrying on by such corporation, association, educational institution, or society of its activities.” 42 U.S.C. § 2000e-1(a). CADA has an identical religious exemption, and goes even further by excluding religious organizations from the definition of “employer.” See Colo. Rev. Stat. § 24-34-402(6), (7) (2022).
[31]. Employment discrimination consists of “discrete discriminatory acts [or] hostile work environment claims.” Nat’l R.R. Passenger Corp. v. Morgan, 536 U.S. 101, 109 (2002), superseded by statute on other grounds by 42 U.S.C. § 2000e-5(e) (2009); cf. Masterpiece Cakeshop, Ltd. v. Colo. Civ. Rts. Comm’n, 584 U.S. 617, 655 (2018) (Thomas, J., concurring in part) (“[p]ublic accommodations laws do not target speech but instead prohibit the act of discriminating. . . .”) (internal quotations omitted). Although sexual and racial harassment typically involve speech, the Court has been careful to refer to workplace harassment as “conduct.” See Harris v. Forklift Sys., 510 U.S. 1, 22 (1993) (“the harassing conduct”).
[32]. See, e.g., 303 Creative LLC v. Elenis, 600 U.S. 570, 590 (2023); id. at 606 (Sotomayor, J., dissenting).
[33]. Masterpiece Cakeshop, 584 U.S. at 628–30.
[38]. Brief of Amici Curiae William Jack and the National Center for Law and Policy in Support of Petitioners at 4–8, Masterpiece Cakeshop, Ltd., 584 U.S. 617 (No. 16-111) [hereinafter Jack Amicus]; Stephanie Mencimer, Did the Supreme Court Fall for a Stunt, Mother Jones (June 7, 2018), https://www.motherjones.com/politics/2018/06/did-the-supreme-court-fall-for-a-stunt.
[39]. Supposedly, it was Jack who described the symbol as a “ghostbusters” symbol to the bakers. Mencimer, supra note 38.
[41]. Jack Amicus, supra note 38, at 4–8; Masterpiece Cakeshop, 584 U.S. at 639–40.
[42]. See Petition for Writ of Certiorari at vii, 28, Masterpiece Cakeshop, Ltd., 584 U.S. 617 (No. 16-111) (citing and arguing Jack cases); Jack Amicus, supra note 38. The National Center for Law & Policy describes itself as a “legal defense organization which focuses on the protection and promotion of religious freedom, the sanctity of life, traditional marriage, parental rights, and other civil liberties.” About Us, Nat’l Ctr. for L. & Pol’y, https://nclplaw.org/about-us.Its president has close ties with the Alliance Defense Fund, which represented Phillips before the Supreme Court. Id.
[43]. 593 U.S. 522 (2021). See infra text accompanying note 150.
[44]. Rightly or wrongly, the Court found the particular remedial measures imposed by the Colorado Human Rights Commission to be draconian and impermissibly disrespectful to the baker’s religious exercise. Had Justice Scalia lived to sit on that case, he would undoubtedly have called the remedy—which included a mandated course of sensitivity training—“unconstitutionally ‘woke.’”
[45]. Masterpiece Cakeshop, 584 U.S. at 634. Two of the three purportedly disparaging, anti-religious statements were clearly innocuous, stating only, as Kennedy essentially conceded, “that a business cannot refuse to provide services based on sexual orientation, regardless of the proprietor’s personal views.” Id. at 635; see Joint Appendix at 202, 205–06, Masterpiece Cakeshop, Ltd., 584 U.S. 617 (No. 16-111). A more inflammatory remark, by an unidentified commissioner, said that “it is one of the most despicable pieces of rhetoric that people can use to—to use their religion to hurt others.” Masterpiece Cakeshop, 584 U.S. at 635. Curiously, this statement, quoted in Phillips’s briefs, does not appear in the Joint Appendix; the Court cites it as “Tr.,” so presumably it is in the full hearing transcript, though the context for it is not provided in the Supreme Court filings. The Court decided this statement impugned Phillips’s sincerity, though it might well have been intended to make the general point that public accommodations laws should not make exceptions for religious beliefs, whether sincere or not. Recall that the Supreme Court in Newman v. Piggie Park dismissed the claim that racial integration was “against the will of God” as “patently frivolous.” 390 U.S. 400, 403 n.5 (1968). The commissioner’s statement here is in line with this sentiment in Newman, unless one views anti-LGBT bias as not despicable.
[46]. [Justice Kennedy:] “Well, suppose we—suppose we thought there was a significant aspect of hostility to a religion in this case. Could your judgment stand?” Oral Argument at 42:24, Masterpiece Cakeshop, Ltd., 584 U.S. 617 (16-111), https://www.oyez.org/cases/2017/16-111.
[47]. The Supreme Court has not decided whether disparate impact claims are viable under Title II of the Civil Rights Act of 1964. See Hardie v. NCAA, 876 F.3d 312, 319 (9th Cir. 2017). Lower courts are divided over the issue. See id.; Monson v. Rochester Athletic Club, 759 N.W.2d 60 (Minn. Ct. App. 2009) (rejecting a disparate impact theory of liability under the Minnesota Human Rights Act by comparing the statute to Title II); Currier v. Nat’l Bd. of Med. Exam’rs, 462 Mass. 1, 20 (2012) (recognizing disparate impact claim under the state’s public accommodation statute); Nancy Leong & Aaron Belzer, The New Public Accommodations: Race Discrimination in the Platform Economy, 105 Geo. L.J. 1271, 1309–10 (2017). A disparate impact theory of liability is said to be available under Title III of the Americans with Disabilities Act, see Bell v. Fam. Dollar Store Jane, No. 23-CV-5307, 2023 U.S. Dist. LEXIS 138560, at *4 (S.D.N.Y. Aug. 7, 2023), but disability-discrimination statutes’ accommodation requirements are not based on intent to discriminate in the first place.
[48]. See Church of the Lukumi Babalu Aye v. City of Hialeah, 508 U.S. 520 (1993) (upholding the right of religious sect to conduct ritual animal sacrifices as against a city ordinance targeting that practice).
[49]. To be sure, messages can be added to wedding cakes. In theory, a client could ask for plastic figurines, a photo reproduction, or a verbal message, indicating that the cake is for a same-sex wedding. That would have made the case closer to 303 Creative. But the record did not show Craig and Mullins asking for such messages. And Justices Gorsuch and Thomas did make such explicit messages a requirement of their conclusions.
[50]. Masterpiece Cakeshop, 584 U.S. at 655–56 (Thomas, J., concurring in part) (quoting Hurley v. Irish-Am. Gay, Lesbian & Bisexual Grp. of Bos., 515 U.S. 557, 573 (1995)).
[52]. See infra section III.A.
[53]. Masterpiece Cakeshop, 584 U.S. at 656 (Thomas, J., concurring in part).
[54]. Id. at 664 (“The Court of Appeals did not address whether Colorado’s law survives strict scrutiny, and I will not do so in the first instance.”).
[55]. The Court has at most indicated in conclusory fashion that public accommodations laws serve a “compelling” state interest, but they did so in a case that could arguably be understood as applying intermediate scrutiny. See Roberts v. U.S. Jaycees, 468 U.S. 609 (1984); supra notes 79, 122, and accompanying text.
[56]. See David L. Hudson, Jr., & Emily H. Harvey, Dissecting the Hybrid Rights Exception: Should it Be Expanded or Rejected?, 38 U. Ark. Little Rock L. Rev. 449, 471–75 (2016) (recommending the free-speech approach for “advocates of religious freedom”).
[57]. See supra note 46 and accompanying text.
[58]. Brief for Petitioner at v–vi, Masterpiece Cakeshop, Ltd., 584 U.S. 617 No. 16-111).
[59]. There are two main types of antidiscrimination laws. Public accommodations laws prohibit discrimination against identified protected groups in “places of public accommodation”—businesses and public services open to the public. Employment discrimination laws prohibit discrimination against protected groups by employers. Compare, e.g., Title II (public accommodations), with Title VII (employment) of the 1964 Civil Rights Act.
[60]. 584 U.S. at 655 (internal quotations omitted) (quoting Hurley v. Irish-Am. Gay, Lesbian & Bisexual Grp. of Bos., 515 U.S. 557, 572 (1995); accord, e.g., Rumsfeld v. F. for Acad. and Institutional Rts., Inc., 547 U.S. 47, 62 (2006) (“Congress, for example, can prohibit employers from discriminating in hiring on the basis of race. The fact that this will require an employer to take down a sign reading ‘White Applicants Only’ hardly means that the law should be analyzed as one regulating the employer’s speech rather than conduct.”); R.A.V. v. City of St. Paul, 505 U.S. 377, 390 (1992) (“Where the government does not target conduct on the basis of its expressive content, acts are not shielded from regulation merely because they express a discriminatory idea or philosophy.”); Norwegian Cruise Line Holdings Ltd. v. State Surgeon Gen., 50 F.4th 1126, 1135–36 (11th Cir. 2022) (“Anti-discrimination statutes ordinarily regulate non-expressive conduct.”).
[61]. 303 Creative LLC v. Elenis, 600 U.S. 570, 604 (2023) (Sotomayor, J., dissenting) (“[T]he law in question targets conduct, not speech, for regulation, and the act of discrimination has never constituted protected expression under the First Amendment”); see, e.g., Brief of Professor Tobias B. Wolff as Amicus Curiae in Support of Respondents at 2, 303 Creative LLC, 600 U.S. 570 (No. 21-476) (“Anti-discrimination laws regulate conduct in the marketplace, not speech”).
[62]. See Eugene Volokh, How Harassment Law Restricts Free Speech, 47 Rutgers L. Rev. 563 (1995).
[63]. See Clark v. Comty. for Creative Non-Violence, 468 U.S. 288, 294, 299 (1984).
[64]. The O’Brien test asks whether the government’s interest is “important or substantial.” 391 U.S. 367, 377 (1968); see, e.g., Clark, 468 U.S. at 293–94; 303 Creative, 600 U.S. at 626 (Sotomayor, J., dissenting). So-called “time, place, and manner” regulations are another species of the genus of content-neutral regulations of speech, and can be thought of as regulating the conduct elements of traditional speech. These concepts are virtually interchangeable and use the same O’Brien test. See Clark, 468 U.S. at 294 (agreeing that the challenged regulation “is defensible either as a time, place, or manner restriction or as a regulation of symbolic conduct”).
[66]. Rumsfeld v. F. for Acad. & Institutional Rts., 547 U.S. 47, 67 (2006).
[68]. See, e.g., Virginia v. Black, 538 U.S. 343, 388 (2003) (Thomas, J., dissenting) (arguing that cross burning is “conduct” with no “expressive component”).
[69]. See, e.g., Clark, 468 U.S. at 296 (“[W]e have assumed for present purposes that the sleeping [overnight in the park] banned in this case would have an expressive element.”); O’Brien, 391 U.S. at 376 (“[E]ven on the assumption that the alleged communicative element in O’Brien’s conduct is sufficient to bring into play the First Amendment.”).
[70]. “It is, in short, not simply the verbal or nonverbal nature of the expression, but the governmental interest at stake, that helps to determine whether a restriction on that expression is valid.” Texas v. Johnson, 491 U.S. 397, 406–07 (1989); see supra text accompanying note 67 (stating the O’Brien test).
[71]. Content-neutral laws regulate “non-speech” elements of conduct for reasons “unrelated to the suppression of free expression.” O’Brien, 391 U.S. at 377.
[74]. The Committee Reports in support of the draft-card law made crystal clear that Congress intended to suppress draft-card mutilation when used as a form of anti-war protest. Id. at 387 (reprinting excerpt of committee reports) (“The committee has taken notice of the defiant destruction and mutilation of draft cards by dissident persons who disapprove of national policy. If allowed to continue unchecked this contumacious conduct represents a potential threat to the exercise of the power to raise and support armies.”).
[79]. See F. for Acad. & Institutional Rts., 547 U.S. 47, 62, 67 (2006) (using antidiscrimination law as paradignmatic example of content-neutral law subject to O’Brien test); 303 Creative LLC v. Elenis, 600 U.S. 570, 630 (2023) (Sotomayor, J., dissenting) (“Because any burden on petitioners’ speech is incidental to CADA’s neutral regulation of commercial conduct, the regulation is subject to the standard set forth in O’Brien.”). In Roberts itself, it was unclear what test was applied. Although recognizing the law in question as content-neutral, the Court seemed to suggest that the application of the law in that case would have met something sounding like strict scrutiny. See Roberts, 468 U.S. at 623–24. That could be read to say that public accommodations laws would satisfy strict scrutiny, even though they need only satisfy intermediate scrutiny. But that ambiguity might also reflect the Court’s long-standing tendency in First Amendment cases to reformulate its intermediate scrutiny tests, and to use strict-scrutiny terminology when applying intermediate scrutiny. Compare O’Brien, 391 U.S. at 377 (stating original test) with Clark v. Comty. for Creative Non-Violence, 468 U.S. 288, 293 (reformulating O’Brien test and using strict-scrutiny “compelling interest” terminology) with F. for Acad. & Institutional Rts. at 67 (further reformulating O’Brien test and omitting its third prong). Sometimes the Court is unsure or mistaken about whether it has applied the O’Brien test. Compare id. at 66 (asserting that Court applied O’Brien test to strike down flag-burning statute in Texas v. Johnson) with Texas v. Johnson, 491 U.S. 397, 410 (1989) (concluding that “[w]e are thus outside O’Brien’s test altogether”).
[81]. Id. at 78 (quoting Norwood v. Harrison, 413 U.S. 455, 470 (1973)).
[83]. The Court was suggesting that the values of the Free Exercise Clause might allow some state support for private religious schools notwithstanding the Establishment Clause.
[87]. Robert Post, Public Accommodations and the First Amendment: 303 Creative and “Pure Speech,” 2023 Sup. Ct. Rev. 251 (2023).
[88]. 303 Creative LLC v. Elenis, 600 U.S. 570, 597 (2023).
[89]. Id. at 586–88 (internal quotations and brackets omitted). The majority purports to attribute the “coercive elimination” statement to the Tenth Circuit, but the Gorsuch opinion embraces it, so my attribution in the text is fair game.
[90]. The stipulations are discussed infra section III.B.1.
[91]. Hurley v. Irish-Am. Gay, Lesbian & Bisexual Grp. of Bos., 515 U.S. 557, 560 (1995).
[93]. Mass. Gen. Laws ch. 272, § 98.
[94]. See Nat’l Socialist Party of Am. v. Skokie, 432 U.S. 43 (1977).
[95]. Hurley, 515 U.S. at 566.
[97]. Masterpiece Cakeshop, Ltd. v. Colo. Civ. Rts. Comm’n, 584 U.S. 617, 655–56 (2018) (Thomas, J., concurring in part) (internal quotations omitted) (quoting Hurley, 515 U.S. at 572).
[98]. My colleague Anuj Desai suggests that the 303 Creative Court viewed the purportedly compelled speech as violating “freedom of conscience,” which could be viewed as distinct from government regulations restricting speech. That suggestion makes sense, but the distinction would represent a new doctrinal direction. And such an explanation fails to account for the problem that antidiscrimination laws are consistently viewed as content-neutral and subject to the O’Brien test.
[99]. The seven appearances of the word “content” in the majority opinion all refer to the content of speech, not the content-basis or content-neutrality of the government regulation. See, e.g., 303 Creative LLC v. Elenis, 600 U.S. 570, 592 (2023) (“[T]he State could not use its public accommodations statute to deny speakers the right to choose the content of their own messages.” (internal quotations and brackets omitted)); id. at 596 (“[N]o government may alter the expressive content of her message.” (internal quotations omitted)). The dissent contends that the Colorado law is content-neutral. Id. at 629, 631, 635 (Sotomayor, J., dissenting) (“[A]ny burden on the company’s expression is incidental to the State’s content-neutral regulation of commercial conduct.”).
[100]. Id. at 583, 586, 587, 593, 597, 599.
[101]. Id. at 582–83 (citations omitted) (bullets and brackets in original).
[106]. Id. at 597. See Andrew Koppelman, Why Gorsuch’s Opinion in ‘303 Creative’ Is So Dangerous, Am. Prospect (Jul. 12, 2023), https://prospect.org/justice/ 2023-07-12-gorsuch-opinion-303-creative-dangerous (arguing that Gorsuch treated the Tenth Circuit’s “pure speech” characterization as a factual finding).
[107]. See, e.g., Tinker v. Des Moines Ind. Cmty. Sch. Dist., 393 U.S. 503, 505 (1969); Bigelow v. Virginia, 421 U.S. 809, 817 (1975); Virginia v. Hicks, 539 U.S. 113, 124 (2003).
[108]. See Texas v. Johnson, 491 U.S. 397, 407, 420 (1989) (applying strict scrutiny rather than the O’Brien test to flag-burning statute).
[109]. Colo. Rev. Stat. § 24-34-601(1) (2023).
[111]. See Clark v. Cmty. for Creative Non-Violence, 468 U.S. 288, 293 (1984); cf. Roberts v. U.S. Jaycees, 468 U.S. 609, 623 (1984) (focusing analysis on whether the law interfered with the Jaycees’ message).
[113]. See supra text accompanying note 101.
[116]. 303 Creative LLC v. Elenis, 600 U.S. 570, 582 (2023).
[119]. Consider oath requirements imposed by statutes or state constitutions on jurors, peace officers, and various government officials not covered by the Constitution’s oath requirements.
[122]. Id. at 590 (quoting Roberts v. U.S. Jaycees, 468 U. S. 609, 628 (1984)).
[123]. Id. at 588, 599 (internal quotations and brackets omitted) (brackets supplied). The majority purports to attribute these statement to the Tenth Circuit, but adopts them as its own.
[125]. See, e.g., Kenji Yoshino, The New Equal Protection, 124 Harv. L. Rev. 747, 760 (2011); Gerald Gunther, Foreword: In Search of Evolving Doctrine on a Changing Court: A Model for a Newer Equal Protection, 86 Harv. L. Rev. 1, 18–20 (1972) (coining the term “bite” to describe less-deferential rational basis cases); David S. Schwartz & Lori Ringhand, Constitutional Law: A Context and Practice Casebook 1097 (3d ed. 2021) (categorizing sexual orientation and disability as “not-quite-suspect classifications” reviewed under “rational basis with bite”).
[126]. Colorado Antidiscrimination Act, Colo. Rev. Stat. § 24-34-601(2)(a) (2022).
[127]. “Infringements on [expressive association] may be justified by regulations adopted to serve compelling state interests, unrelated to the suppression of ideas, that cannot be achieved through means significantly less restrictive of associational freedoms.” Roberts, 468 U.S. at 623. See 303 Creative, 600 U.S. at 626 (Sotomayor, J., dissenting) (describing O’Brien test as “lesser” than strict “constitutional scrutiny”); Masterpiece Cakeshop, Ltd. v. Colo. Civ. Rts. Comm’n, 584 U.S. 617, 663–64 (2018) (Thomas, J., concurring in part) (same).
[128]. 303 Creative, 600 U.S. at 589–90. The quoted hypotheticals came from Judge Tymkovich’s dissenting opinion in the Tenth Circuit, which the majority praises.
[129]. The dissent answered these hypotheticals by pointing out that the muralists or filmmakers in the hypothetical typically do not in fact offer their services to the general public, and are therefore not public accommodations. And, as for the rest, they may refuse to make certain messages as long as their refusal applies to all customers. But that argument is vulnerable to Gorsuch’s critique that the distinctions are based on opportunistically changing the level of generality.
[130]. Presumably, it is understandable that Smith would not begin making wedding websites until clarifying her legal rights. But she won her Supreme Court case on June 30, 2023. As of March 28, 2024, Smith’s 303 Creative website still said that the wedding websites are “coming soon.” 303 Creative LLC, https://303creative.com (last visited, Mar. 28, 2024). In other words, Smith had yet to create a portal page for her wedding website line of business.
[131]. Melissa Gira Grant, The Mysterious Case of the Fake Gay Marriage Website, the Real Straight Man, and the Supreme Court, New Republic (June 29, 2023), https://newrepublic.com/article/173987/mysterious-case-fake-gay-marriage-website-real-straight-man-supreme-court.
[132]. Compare 303 Creative LLC v. Elenis, 6 F.4th 1160 (10th Cir. 2021), rev’d, 600 U.S. 570 (upholding state public accommodations law against free speech challenge), with Telescope Media Grp. v. Lucero, 936 F.3d 740 (8th Cir. 2019) (holding that wedding videographers had First Amendment right to refuse to make videos of same-sex weddings notwithstanding state public accommodations law).
[133]. To take just one recent example, in Van Buren v. United States, 593 U.S. 374 (2021), the Court resolved an eight-year-old circuit split between United States v. Nosal, 676 F.3d 854 (9th Cir. 2012) (en banc) and United States v. Rodriguez, 628 F.3d 1258 (11th Cir. 2010). According to a 2015 study of circuit splits, only 5% of circuit splits that first occurred in 2005 had been resolved by 2013. See Deborah Beim & Kelly Rader, Evolution of Conflict in the Federal Circuit Courts (Mar. 19, 2015) (unpublished manuscript), https://ssrn.com/abstract=2623304.
[134]. I have found no cases since 2010 in which a public accommodations law was challenged on free speech grounds outside the context of religious objections to same-sex marriage or accommodation of trans people. See Appendix A.
[135]. My research, which is admittedly brief and non-dispositive, found only twelve cases since 2010, other than Masterpiece Cakeshop and 303 Creative, involving the application of a public accommodations law to a business opposed to same-sex marriage. Three of the twelve involved refusal to provide a wedding venue, which would not have qualified as expressive activity by a “creative professional” under 303 Creative. Of the remaining nine cases arguably involving a creative professional, one was the Scardina test case that was ginned up against Masterpiece Cakeshop itself, described in infra note 136. Five of the cases were pre-enforcement challenges brought by the Alliance Defending Freedom, which is suggestive of made-up test cases. For a table of these cases and a methodological statement, see Appendix A.
[136]. And the “problem” of test-case creation in the LGBTQ-rights culture war continues. A case is currently pending before the Colorado Supreme Court in Masterpiece Cakeshop, et al. v. Scardina, 2023 Colo. Lexis 960 (2023), in which Autumn Scardina, a transgender woman attorney, sued Masterpiece Cakeshop and Jack Phillips for transgender discrimination. On the same day that the Supreme Court granted cert in Masterpiece Cakeshop, according to Andrew Koppelman, Scardina went to the baker and requested a custom cake decorated with blue-and-pink icing. Phillips initially accepted the order, but when Scardina next disclosed that she wanted the cake to celebrate her
gender transition, Phillips changed his mind and refused to bake the cake. He later explained “that he ‘won’t design a cake that promotes something that conflicts with
[his] Bible’s teachings’ and that ‘he believes that God designed people male and female, that a person’s gender is biologically determined.’” Andrew Koppelman, The Colorado cake wars continue, with a literally colorful twist, Hill (Jan. 18, 2024), https://thehill.com/opinion/judiciary/4414848-the-colorado-cake-wars-continue-with-a-literally-colorful-twist. Scardina initially won her administrative complaint under CADA, but the Commission dismissed her case after the Supreme Court issued its Masterpiece Cakeshop opinion. Scardina then sued Phillips under CADA’s private right of action. See id. Scardina’s test case is presumably intended to test the limits of the free speech argument by focusing on a cake that lacks the purported symbolism of a wedding cake.
[138]. Joint Appendix at 285, 303 Creative LLC v. Elenis, 600 U.S. 570 (2023) (No. 21-476).
[139]. Petition for a Writ of Certiorari at 1, 303 Creative LLC v. Elenis, 600 U.S. 570 (2023) (No. 21-476) (emphasis added).
[141]. 303 Creative LLC v. Elenis, 142 S. Ct. 1106 (2022) (mem.).
[143]. 374 U.S. 398 (1963). The Court did not formally overrule Sherbert, but limited its applicability to situations where the state’s unemployment insurance law made exemptions for secular reasons but not similar religious ones. Smith, 494 U.S. at 884–85.
[144]. 494 U.S. at 885 (internal quotations omitted).
[145]. “Repeatedly and in many different contexts, we have warned that courts must not presume to determine the place of a particular belief in a religion or the plausibility of a religious claim.” Smith, 494 U.S. at 887; see also Masterpiece Cakeshop, Ltd. v. Colo. Civ. Rts. Comm’n, 584 U.S. 617, 666 (2018) (Thomas, J., concurring in part) (“This Court is not an authority on matters of conscience. . . .”).
[146]. Smith, 494 U.S. at 888–89 (emphasis added) (citations omitted). Each of Scalia’s examples was accompanied by a case cite referring to a claim for such a religious exemption.
[147]. Masterpiece Cakeshop, 584 U.S. at 643 (Gorsuch, J., concurring).
[149]. Petition for Writ of Certiorari at 1, Fulton, 593 U.S. 522 (No. 19-123).
[151]. Id. at 618 (Gorsuch, J., concurring). Thomas and Alito joined that opinion. See also id. at 543–44 (Alito, J., concurring).
[152]. Id. at 543 (Barrett, J., concurring).
[153]. Religious exemptions from public accommodations laws have been advocated under a “hybrid rights” theory based on a passage in Justice Scalia’s opinion in Emp. Div. v. Smith, 494 U.S. 892, 881 (1990). Scalia distinguished prior cases upholding religious exemptions from generally applicable laws, such as Wisconsin v. Yoder, 406 U.S. 205 (1972), by observing that those cases “have involved not the Free Exercise Clause alone, but the Free Exercise Clause in conjunction with other constitutional protections, such as freedom of speech. . . .” Smith. But the “hybrid rights” theory never gained traction with the Court, see Hudson & Harvey, supra note 56, at 474, and was not taken up in 303 Creative.
[154]. Cole, supra note 10, at 501.
[156]. Id. at 501–02 (arguing that this test would not be met in “the vast majority of instances in which antidiscrimination laws are applied to expressive businesses”).
[157]. 303 Creative LLC v. Elenis, 600 U.S. 570, 602 (2023). I omit the internal quotations and brackets because the Court’s quotation from the Tenth Circuit opinion was approving.
[158]. Dale Carpenter, How to Read 303 Creative v. Elenis, Reason: The Volokh Conspiracy (Jul. 3, 2023), https://reason.com/volokh/2023/07/03/how-to-read-303-creative-v-elenis.
[159]. Id. (page citations omitted).
[160]. Brief of Amici Curiae Prof. Dale Carpenter et al., 303 Creative, 600 U.S. 570; see Brief of Amici Curiae Cato Institute, Eugene Volokh, & Dale Carpenter in Support of Petitioner at 17–20, Elane Photography LLC v. Willock, 572 U.S. 1046 (2014) (No. 13-585) (cert. denied).
[161]. Brief of American Unity Fund & Profs. Dale Carpenter and Eugene Volokh as Amici Curiae in Support of Respondents at 4, Masterpiece Cakeshop, Ltd. v. Colo. Civ. Rts. Comm’n, 584 U.S. 617 (2018) (No. 16-111).
[162]. Carpenter, supra note 158.
[163]. See supra notes 160–161.
[164]. See Brief of Amici Curiae Prof. Dale Carpenter et al., supra note 160, at 22.
[165]. “The Court’s precedents overwhelmingly favor a categorical approach prohibiting compelled speech.” Brief of Amici Curiae First Amendment Scholars in Support of Petitioners at 11, 303 Creative LLC v. Elenis, 600 U.S. 570 (2023) (No. 21-476).
[166]. See supra text accompanying note 124.
[168]. See Koppelman, supra note 106.
[169]. Obergefell v. Hodges, 576 U.S. 644, 688 (2015) (Roberts, C.J., dissenting).
[171]. Thus, for example, Justice Scalia in Lawrence v. Texas:
Many Americans do not want persons who openly engage in homosexual conduct as partners in their business, as scoutmasters for their children, as teachers in their children’s schools, or as boarders in their home. They view this as protecting themselves and their families from a lifestyle that they believe to be immoral and destructive. The Court views it as “discrimination” which it is the function of our judgments to deter. So imbued is the Court with the law profession’s anti-anti-homosexual culture, that it is seemingly unaware that the attitudes of that culture are not obviously “mainstream[.]”
Lawrence v. Texas, 539 U.S. 558, 602–03 (2003) (Scalia & Thomas, JJ., dissenting); see also Romer v. Evans, 517 U.S. 620, 636 (1996) (Scalia & Thomas, JJ., dissenting) (“This Court has no business imposing upon all Americans the resolution favored by the elite class from which the Members of this institution are selected, pronouncing that ‘animosity’ toward homosexuality is evil.”).
For this sort of demagoguery in another context, see Students for Fair Admissions v. Harvard, 600 U.S. 181, 280 (2023) (Thomas, J., concurring) (finding affirmative action is “a call to empower privileged elites, who will ‘tell us what is required to level the playing field’ among castes and classifications that they alone can divine” (internal brackets omitted)); Grutter v. Bollinger, 539 U.S. 306, 366 (2003) (Thomas, J., concurring in part, dissenting in part) (“Apparently where the status quo being defended is that of the elite establishment—here the Law School—rather than a less fashionable Southern military institution, the Court will defer without serious inquiry and without regard to the applicable legal standard.”).
[172]. See supra text accompanying note 39.
[174]. 303 Creative LLC v. Elenis, 600 U.S. 570, 588 (2023) (internal quotations omitted).