Free Speech and Generally Applicable Laws: A New Doctrinal Synthesis

I.     Introduction

Categories are central to free-expression jurisprudence. The Supreme Court, for example, has sharply distinguished between protected and unprotected speech, content-discriminatory and content-neutral statutes, and public and non-public forums.1 Another key dividing line separates laws that directly burden speech from laws that burden speech only “incidentally.”2 Restrictions on defamation, for example, fall into the direct-burden category because they provide redress for reputation-harming expression itself.3 In contrast, a law that prohibits vandalism burdens speech only incidentally, even when applied to the dissident who spray paints “MAKE LOVE, NOT WAR” on the Washington Monument. Vandalism laws, after all, do not target expression. Nor do they concern expression in many of their applications.4 As a result, courts often describe such laws as “generally applicable,” with the consequence that they receive only limited, if any, judicial scrutiny in their application to expressive activity.5

These concepts lie at the heart of First Amendment law.6 They also are of far-reaching practical importance because most laws are generally applicable, and such laws interact with expressive activities in “virtually limitless” ways.7 Even so, these concepts are deeply undertheorized and widely misunderstood—so much so that commentators have decried the Supreme Court’s rulings in this area as “perplexing,”8 “inconsistent,”9 and “conflicting.”10

Confusion springs from a mix of sources. To begin with, decisions that call for deferential review of “generally applicable laws” based on their “incidental” effects have not supplied useful definitions of these key terms.11 In addition, the Court has held that some generally applicable laws call for meaningful judicial scrutiny, while other generally applicable laws merit no scrutiny at all.12 But the Justices have failed to explain with clarity how lower courts are to distinguish between these separate sets of cases.13 Another complication arises from the Court’s landmark Free Exercise Clause ruling in Employment Division v. Smith.14 In that case, the Court held that a religious practitioner who ingested peyote as a sacramental act could not challenge an across-the-board prohibition on peyote use as applied to him precisely because of its “generally applicable” character.15 Common sense might suggest that a serious speaker should be no more able to challenge a generally applicable law than a serious worshipper; however, Supreme Court doctrine does not track this intuition, and the Justices have never told us why.

This Article grapples with these matters. It posits that the Court’s Free Speech Clause jurisprudence, on close examination, supports an approach that divides up generally-applicable-law cases into three mutually exclusive categories, as illustrated by three seminal cases: Holder v. Humanitarian Law Project,16 United States v. O’Brien,17 and Arcara v. Cloud Books, Inc.18

In Humanitarian Law Project, the Court confronted a federal statute that barred the giving of “material support” to terrorist groups and defined “material support” to include “training.”19 Humanitarian Law Project, a nonprofit organization, argued that the statute was unconstitutional as applied to its activities because it sought to communicate with a government-listed terrorist group only to offer instruction on peaceable matters, such as pursuing change through legislative reform.20 In response, the government argued that the statute should escape strict scrutiny because it was “generally applicable”; after all, the government insisted, the material-support prohibition covered speakers and non-speakers alike, and it did so in an effort not to stifle expression but to stem terrorist violence.21 The Court, however, rejected this position. It reasoned that the law as applied to the challenger did target speech because the expression at issue in this particular case—the giving of instruction—was what constituted the statutorily prohibited material support.22 In other words, to use the category-defining terms developed in this Article, the statute placed a direct-in-effect burden on expression in its application to Humanitarian Law Project because that law operated against that organization precisely and only because it was speaking.23

A contrasting set of facts gave rise to O’Brien. That case involved a Free Speech Clause defense asserted by a draft-card-burning political protestor prosecuted under a law that banned all forms of draft-card mutilation.24 Faced with these facts, the Court assumed that the draft-card burning that triggered the prohibition’s operation constituted “speech” for First Amendment purposes.25 Unlike in Humanitarian Law Project, however, the communicative impact of the speech in O’Brien had nothing to do with the criminality of the defendant’s behavior. The mutilation of the draft card was in and of itself the actus reus of the crime, regardless of any communicative impact that the mutilation might have had on any listener or observer. For this reason, in contrast to the situation in Humanitarian Law Project, the across-the-board ban on draft-card destruction placed only an incidental burden on the war protestor’s speech, as opposed to burdening his speech in a direct-in-effect way.

Arcara v. Cloud Books, Inc. illustrates a third type of generally applicable law.26 That case involved the application to bookstore operators of a New York statute that called for the closure of any place of business involved in acts of unlawful public lewdness.27 Invoking the statute, a state court shut down the defendants’ store for a year because prostitution and other unlawful sex-related activity had been occurring on its premises.28 On appeal, the defendants asserted that a government-mandated closure of an entire bookstore involved such a draconian interference with communicative liberty that the state law, as applied to them, violated the Free Speech Clause.29 The Court, however, rejected this claim, employing an even less exacting form of review than it had put to work in O’Brien. The Court reasoned that the bookstore case differed from the draft-card case because the operation of the legal prohibition at issue in Arcara was not triggered by speech at all; rather, it was triggered by lewd acts that involved “absolutely no element of protected expression.”30

On its face, the closing of a bookstore interfered with communicative activity.31 But such a result, the Court explained, often arises when applications of non-speech-related laws impede opportunities to engage in expression—as when, for example, a large damages award levied against a negligent driver reduces that person’s financial capacity to send out political messages.32 These sorts of cases, according to the synthesis developed here, involve a doubly incidental burden on speech. This label applies because the challenged law in its application neither purposefully targets the expressive effects of conduct (as in Humanitarian Law Project) nor non-purposefully outlaws speech in the sense that the prohibited conduct turns out to be itself expressive (as in O’Brien). Rather, the claimed interference with speech arises only because the operation of the challenged prohibition on non-speech—which in Arcara took the form of prostitution and other lewd, non-communicative acts—has ripple effects that include an inhibition on engaging in communicative activity. In effect, the Court in Arcara held that generally applicable laws that impose doubly incidental burdens on speech are categorically distinguishable from laws that impose direct-in-effect or incidental burdens on expression.

So why, exactly, does it matter whether the law that faces First Amendment challenge imposes a direct-in-effect, incidental, or doubly incidental burden on speech? The answer, not surprisingly, is that—at least as a general matter—different levels of scrutiny apply to cases that fall into these different legal categories. More specifically, (1) when a generally applicable law has the effect of burdening speech in a direct-in-effect way (as in Humanitarian Law Project), the Court has applied strict scrutiny;33 (2) when such a law burdens speech in an incidental way (as in O’Brien), the Court has applied intermediate scrutiny;34 and (3) when such a law burdens speech in a doubly incidental way (as in Arcara), the Court has applied no First Amendment scrutiny at all.35 On the face of things, this organizational structure seems straightforward. In fact, however, complexities lurk around every turn. This Article lays bare and grapples with those complexities.

Part II highlights the many difficulties that mark the Court’s past application of the Free Speech Clause to generally applicable laws. At the heart of this discussion lies Cohen v. Cowles Media Co.36 and the Court’s murky treatment in that case of its earlier ruling in Hustler Magazine, Inc. v. Falwell.37 Part II does not visit every complication raised by the Court’s treatment of First Amendment challenges to generally applicable laws, but it tells enough of the tale to show that analytical quandaries pervade this constitutional field.

Part III begins the process of untangling the doctrinal knot by focusing attention on direct-in-effect-burden cases. Central to this discussion is Professor Eugene Volokh’s claim that some generally applicable laws should be subject to aggressive First Amendment scrutiny because they are “content-based as applied.”38 Part III explains how the Court’s ruling in Humanitarian Law Project, while potentially subject to other interpretations, is best understood as vindicating the speech-protective thesis set forth in Professor Volokh’s earlier work. It also suggests that some generally applicable laws may impose particularly significant burdens on speech even though they do not involve content discrimination—a fact that suggests why the term “direct-in-effect burden” may better capture the principle embraced in Humanitarian Law Project than the “content-based as applied” formulation put forward by Professor Volokh.

Part IV shifts attention to incidental-burden cases. In particular, it considers whether the Court’s free-exercise ruling in Smith has knocked the legs out from under O’Brien’s intermediate-scrutiny approach to Free Speech Clause cases. Part IV advances the position that O’Brien remains good law notwithstanding Smith because different policy concerns are at work in free-speech and free-exercise cases. In addition, Part IV identifies and rejects the normative claim that O’Brien-based review has become so feckless in practical operation that retaining it is not worth the candle.

Part V turns to the set of problems raised by Arcara. It posits that the no-review approach laid down in that case for doubly-incidental-burden cases has proven to have a bark that is worse than its bite. This is the case because both courts and commentators have concluded that the Arcara rule should be laced with exceptions. This development gives rise to both irony and opportunity. Irony exists because Arcara offered the promise of simplifying the law by injecting into it an easy-to-apply, bright-line rule. The well-marked safe harbor that the Justices sought to construct in that case, however, has turned out to be neither well-marked nor safe because of the many doctrinal qualifications that have grown up around it. Opportunity arises for this same reason—that is, the opportunity for thoughtful lawyers to skirt Arcara’s no-review rule by creatively invoking exceptions to it.

Part VI steps away from these matters of doctrinal synthesis to reflect more broadly on the interaction of the First Amendment and generally applicable laws. It emphasizes the need for lawyers to approach these cases with the three-part doctrinal structure developed here firmly in mind, both to avoid missteps of analysis and to ensure that they do not miss chances to advance the interests of their clients. Part VI also notes that recent changes in the Court’s membership have created favorable conditions for lawyers interested in expanding speech-based protections against the operation of generally applicable laws.

At bottom, the aim of this Article is to extract from an existing doctrinal morass an organizing, if still-evolving, synthesis of the law. This synthesis places all of the Court’s key precedents within a coherent organizational framework, and—at least as a general matter—shows that this framework sensibly takes account of the different levels of threat to expressive liberty posed by different types of generally applicable laws. It is concerning that the Court’s own rhetoric often has done more to obfuscate than to illuminate this doctrinal architecture.39 As this Article will demonstrate, however, the full body of the Court’s rhetoric is clear enough, and the Court’s actions speak louder than its words. The end result is that the Court’s treatment of generally applicable laws under the Free Speech Clause reflects the three-part analytical structure—based on direct-in-effect, incidental, and doubly incidental burdens on speech—that is both laid bare and significantly refined in the pages that follow.

II.     The Challenges Presented by Generally Applicable Laws

Any discussion of “generally applicable laws” must begin at the beginning—that is, by reflecting on what types of laws qualify as “generally applicable.” Courts often act as if the term is self-defining, but it is not. Consider an ordinance that bans “leafletting” and goes on to define the term as “distributing in a public place any papers to passersby.” At first blush, this law seems not to be generally applicable because it targets only leafletting, a quintessential First Amendment activity.40 On close inspection, however, the law targets more than expressive leafletting because even a person who hands out blank sheets of paper (perhaps because she is in a gift-giving mood) falls within the legal prohibition. What is more, the ordinance does not discriminate based on the content of the distributed material, and its underlying purpose is to cut down on littering, not to cut down on speech. Indeed, for all these reasons, the ordinance bears a close resemblance to the draft-card-mutilation statute that the Court deemed “generally applicable” in O’Brien.41 Using the verbiage of that case, one might well say that the anti-leafletting law—because it covers both communicative and non-communicative activity in addressing the non-speech-related goal of discouraging littering—“is unrelated to the suppression of free expression.”42

Complications arise, however, because existing law suggests that the Supreme Court would not apply the “generally applicable” label to this ordinance—and with good reason.43 The key point is that the ordinance lays its sting on speech-related activity as a practical matter, because precious few individuals in the real world display their munificence by handing out blank sheets of paper. As this Article later explains, the Court has overlaid its treatment of generally applicable laws in free-speech cases with a significant qualification based on real-world effects.44 This approach seems sensible, in part because it tracks the Court’s handling of similar problems in analogous settings,45 and because it responds to the sensible notion that “[t]he proper focus of constitutional inquiry is the group for whom the law is a restriction, not the group for whom the law is irrelevant.”46 The point is that sometimes a law that appears to be “generally applicable” on its face will not qualify as “generally applicable” for legal purposes because of its predominant speech-inhibiting practical effects. And because the term “predominant speech-inhibiting practical effects” is not (to say the least) self-defining, neither is the term “generally applicable laws.”

A related set of questions regarding the general applicability of laws concerns legislative purpose. In Texas v. Johnson, for example, the Court invalidated a state statute that criminalized desecration of the American flag.47 A major problem with the law sprang from its express terms, which limited the law’s reach to flag burnings “that the actor knows will seriously offend one or more persons.”48 Because this turn of the statutory phrase tied criminality to the particular flag-burning audience’s mental reaction, the Court in effect deemed the law not to be generally applicable because it was content-discriminatory on its face.49 In the wake of Johnson, however, an all-star team of free-speech lawyers set about drafting a flag-burning statute that would withstand constitutional challenge. The product of their efforts was the federal Flag Protection Act of 1989, which—in pointed contrast to the statute at issue in Johnson—made no mention of the taking of “offense” or any other form of viewer reaction.50 The result, so the argument went, was that this new statute—because it omitted any reference to audience reaction—was “generally applicable” and therefore outside the reach of the Court’s earlier flag-burning ruling.51

In United States v. Eichman, however, the Court refused to distinguish Johnson and struck down the new federal flag-burning law. Decisive to the Court was not the new statute’s text, but its underlying purpose. According to Justice Brennan, because the purpose of the new law remained centered on stifling messages inconsistent with the unity-promoting meaning of the flag, the law was not “unrelated to the suppression of free expression” for purposes of O’Brien and thus did not bring into play the ramped-down style of scrutiny made applicable by that case to generally applicable laws.52 Put another way, a law that is “related to the suppression of free expression” is in its nature not generally applicable, and a law can be related to the suppression of expression on its face (as in Texas v. Johnson), in its overwhelming effect (as in the leafletting case), or in its underlying purpose (as in Eichman). On the other hand, so long as a law does not fall into one of these categories, the law will qualify as generally applicable.53

Once a court concludes that a law qualifies as generally applicable, its work is hardly done. Among other things, the court must evaluate whether the challenged law burdens speech in a direct-in-effect, incidental, or doubly incidental way and then determine the impact of placing the law into the applicable type-of-burden pigeonhole. Grappling with challenged laws in this way is not for the faint of heart, and that point is well-illustrated by the Court’s 1991 ruling in Cohen v. Cowles Media Co.54 That case (which is not to be confused with Cohen v. California, another soon-to-be-discussed precedent that also has importance in this context) warrants close attention for two reasons. First, Cohen v. Cowles Media Co. itself constitutes a major Supreme Court precedent on generally applicable laws and thus must find a place within any synthesis of the law in this field. Second, the case illuminates the rich mix of analytical problems presented by Free Speech Clause challenges to generally applicable laws.

Cohen v. Cowles Media Co. arose out of a $200,000 compensatory-damages award assessed against a reporter and his employer under state promissory-estoppel law.55 The action was filed after the reporter breached a promise not to disclose the name of a would-be confidential source, Dan Cohen, thus causing him to suffer significant harms that included the loss of his job.56 The reporter’s promise had come in return for Cohen’s provision of information about past criminal charges brought against Marlene Johnson, who was then running as the Democratic-Farmer-Labor candidate for Lieutenant Governor of Minnesota.57 Notwithstanding the promise, the reporter and his editors ultimately decided to publish Cohen’s name, apparently because (1) he turned out to be a Republican Party associate of a leading gubernatorial candidate; (2) most of the charges against Johnson had been dropped; and (3) the one conviction that resulted from those charges had previously been vacated and involved only a minor matter.58 Put another way, the defendants deemed it newsworthy that prominent opponents of a candidate for a major statewide office might have been trying to sabotage her campaign by circulating overblown charges of wrongdoing against her. They also concluded that sharing the name of their source was necessary to provide a proper reporting of this story.59

Faced with these facts, the Minnesota Supreme Court concluded that “in this case enforcement of the promise of confidentiality under a promissory estoppel theory would violate the defendants’ First Amendment rights.”60 A five-Justice majority of the U.S. Supreme Court disagreed. In an opinion written by Justice White, the Court acknowledged that its past decisions broadly precluded the imposition of penalties for reporting information that had lawfully come into a reporter’s hands—for example, the name of a rape victim made available in non-confidential police reports.61 The majority observed, however, that

this case . . . is not controlled by this line of cases but, rather, by the equally well-established line of decisions holding that generally applicable laws do not offend the First Amendment simply because their enforcement against the press has incidental effects on its ability to gather and report the news.62

According to Justice White:

The press, like others interested in publishing, may not publish copyrighted material without obeying the copyright laws. Similarly, the media must obey the National Labor Relations Act and the Fair Labor Standards Act, may not restrain trade in violation of the antitrust laws, and must pay nondiscriminatory taxes . . . . Accordingly, enforcement of such general laws against the press is not subject to stricter scrutiny than would be applied to enforcement against other persons or organizations.63

In addition, Justice White observed that promissory estoppel constituted a “doctrine . . . generally applicable to the daily transactions of all the citizens of Minnesota”64 and that “Minnesota law simply requires those making promises to keep them.”65 In sum, “generally applicable laws do not offend the First Amendment simply because their enforcement against the press has incidental effects on its ability to gather and report the news.”66

In dissent, Justice Souter took a very different approach. For starters, he declared that “‘there is nothing talismanic about neutral laws of general applicability,’ for such laws may restrict First Amendment rights just as effectively as those directed specifically at speech itself.”67 In addition, this case fell outside “the line of authority holding the press to laws of general applicability where commercial activities and relationships, not the content of publication, are at issue.”68 For this reason, Justice Souter could not accept “the majority’s position that we may dispense with balancing”; rather, because he refused to view “the fact of general applicability to be dispositive,” he deemed it “necessary to articulate, measure, and compare the competing interests involved” in the case.69 Justice Souter acknowledged that it would be wrong “to say that the breach of such a promise of confidentiality could never give rise to liability.”70 In this case, however, “the State’s interest in enforcing a newspaper’s promise of confidentiality [was] insufficient to outweigh the interest in unfettered publication of the information . . . .”71

In a separate dissent, Justice Blackmun agreed with Justice Souter that the cases relied on by the majority—such as those that concerned tax, antitrust, and labor laws—were uninformative because they “did not involve the imposition of liability based upon the content of speech.”72 Instead, he found the Court’s earlier ruling in Hustler Magazine v. Falwell73 “to be precisely on point.”74Hustler involved a jury’s award of damages to the well-known religious leader, Jerry Falwell, based on a liquor-advertisement parody that referenced Reverend Falwell’s “first time” with his mother in an outhouse.75 As Justice Blackmun explained:

There, we found that the use of a claim of intentional infliction of emotional distress to impose liability for the publication of a satirical critique violated the First Amendment. There was no doubt that Virginia’s tort of intentional infliction of emotional distress was “a law of general applicability” unrelated to the suppression of speech. Nonetheless, a unanimous Court found that, when used to penalize the expression of opinion, the law was subject to the strictures of the First Amendment.76

According to Justice Blackmun: “As in Hustler, the operation of Minnesota’s doctrine of promissory estoppel in this case cannot be said to have a merely ‘incidental’ burden on speech; the publication of important political speech is the claimed violation.”77 It followed for Justice Blackmun that Minnesota promissory-estoppel law could “not be enforced to punish the expression of truthful information,” at least when—as in this case—“the State’s interest in enforcing its promissory estoppel doctrine . . . was far from compelling.”78

Not surprisingly, Justice White paused to respond to Justice Blackmun’s invocation of Hustler. That case was not controlling, Justice White wrote, because “Cohen is not seeking damages for injury to his reputation or his state of mind”; instead he had recovered compensatory damages “for breach of a promise that caused him to lose his job and lowered his earning capacity.”79 Justice Blackmun was unsatisfied. He reminded the majority that its basic reason for rejecting the defendants’ First Amendment challenge was that the case involved a “law of general applicability.”80 Nothing in Justice White’s analysis of Hustler, Justice Blackmun insisted, changed the fact that the common-law, intentional-infliction-of-emotional-distress cause of action—just like the common-law, promissory-estoppel cause of action—constituted a generally applicable law.81 In addition, according to Justice Blackmun, there was “no meaningful distinction between a statute that penalizes published speech in order to protect the individual’s psychological well-being or reputational interest and one that exacts the same penalty in order to compensate the loss of employment or earning potential.”82 At the least, he emphasized, “our decision in Hustler recognized no such distinction.”83

These dueling contentions about the Hustler precedent give rise to a host of questions that illuminate the sorts of problems posed by First Amendment challenges to generally applicable laws. To begin with, much of Justice White’s analysis seemed to focus on whether the press possesses any measure of expressive freedom above and beyond the expressive freedom enjoyed by other speakers.84 To be sure, it could be that the First Amendment’s Free Press Clause imposes more exacting limits on generally applicable laws than does the First Amendment’s Free Speech Clause; however, one holding of Cohen v. Cowles Media Co. seems to be (unfortunately for the defendants in the case) that it does not.85 The problem is that Justice White’s opinion did not take care to clarify what portions of his opinion concerned the Press Clause and what portions concerned the Speech Clause—as any proper treatment of a First Amendment challenge to a generally applicable law should be sure to do.

Nor did the analytical problems in Justice White’s opinion stop there. For example, he grouped Cohen v. Cowles Media Co. together with pre-Hustler precedents that left undisturbed state-imposed duties to make payments of taxes or wages even though those payments effectively inhibit a publisher’s ability to engage in communication. This analogy, however, is deeply problematic. The difficulty is that the constitutional challenge in Cohen v. Cowles Media Co. arose because the defendants’ duty to pay damages resulted specifically from the defendants’ expressive activity—that is, the printing of a newspaper account of Mr. Cohen’s actions that included his name. In contrast, earlier challenges by publishers to payments that resulted from the enforcement of minimum-wage, tax, and other similar laws did not concern government action that came to bear on speech itself; instead, the operative theory underlying those challenges was that these forms of non-speech regulation would have ripple effects that inhibited speech by shrinking the pool of resources that a publisher could direct toward speech activity. These cases, in other words, involved laws that imposed (to use the nomenclature of this Article) only doubly incidental burdens on speech. But Cohen v. Cowles Media Co.—because it grew out the imposition of liability based on the content of speech itself—involved a speech burden of a more direct kind.

Nor did Justice White do much in the way of countering Justice Blackmun’s invocation of Hustler. Justice White emphasized, for example, that state promissory-estoppel law applied broadly “to the daily transactions of all the citizens of Minnesota,” many of which transactions obviously did not involve promises about speaking at all.86 But no one questioned the proposition that the cause of action for intentional infliction of emotional distress—just like the cause of action for promissory estoppel—reached across “the daily transactions” of all state residents and could be triggered by both speech-based and non-speech-based activities.87 In addition, while Justice White sought to distinguish Hustler by highlighting the difference between monetary and non-monetary injuries, he never explained why this posited distinction made sense for First Amendment purposes.88 One might say that Justice White meant to suggest that monetary injuries are in general more severe, and thus less tolerable, than injuries to one’s psyche and emotions.89 That proposition, however, is hardly self-evident. Moreover, attributing to Justice White this unstated one-injury-is-worse-than-the-other logic seems to clash with his overarching approach to the case. An assessment of the varied weightiness of relevant state interests, after all, would have engaged the Court in the very sort of interest-balancing analysis that its no-review treatment of the case was, from all appearances, designed to avoid.90

No less perplexing was the majority’s course of action once it declared that Minnesota’s promissory-estoppel doctrine constituted a generally applicable law. As previously detailed, the Court in United States v. O’Brien responded to a war protester’s challenge to a generally applicable draft-card-mutilation law by applying intermediate scrutiny.91 The Court thus put to work a mode of analysis that fell short of strict scrutiny but nonetheless involved a structured balancing of the claimed free-speech right and countervailing state interests at play in that case. In Cohen v. Cowles Media Co., however, the majority did not apply either strict or intermediate scrutiny. Rather, Justice White deemed it appropriate to—as Justice Souter put it—“dispense with balancing” altogether.92 In other words, the majority applied no free-speech-based scrutiny to the challenged application of the promissory-estoppel law, and it did so without even mentioning, far less distinguishing, the Court’s earlier ruling in O’Brien.

This omission is not easy to understand.93 In particular, the state promissory-estoppel law, just like the federal draft-card-mutilation law, applied to speakers and non-speakers alike, and both laws advanced underlying interests—having persons keep their promises in one case and keep their draft cards in the other—that were “unrelated to the suppression of speech.”94 Indeed, for two separate reasons, Cohen v. Cowles Media Co. seemed to present an even more appealing set of facts than did O’Brien for triggering some form of meaningful First Amendment review. First, Cohen v. Cowles Media Co. did not involve only “speech . . . brigaded with action” (that is, the mutilation of a physical object);95 rather, it involved “pure speech” (that is, the publishing of actual words in an actual newspaper).96 Second, in Cohen v. Cowles Media Co., unlike in O’Brien, the operation of the challenged law hinged on the content of the relevant communication in a meaningful sense. The defendant’s liability, after all, was triggered only by publishing a story that included particular information—that is, the source’s name. Justice Blackmun emphasized this point in his dissent when he observed that “the publication of important political speech is the claimed violation” of the challenged state law.97 Thoughtful analysts have highlighted this same point, too.98

If Cohen v. Cowles Media Co. presented a stronger case for applying First Amendment interest-balancing analysis than O’Brien itself, why did Justice White opt for a less stringent mode of scrutiny than the Court had applied in the draft-card case? In a related vein, why did Hustler, which Justice Blackmun described as involving “a law of general applicability,”99 produce an even stricter mode of constitutional review than the Court applied to the paradigmatic generally applicable law at issue in O’Brien? A full understanding of Cohen v. Cowles Media Co., O’Brien, and Hustler also requires investigating how they fit together with both Humanitarian Law Project and Arcara. This Article will explain in due course how the Court’s ruling in Cohen v. Cowles Media Co. is reconcilable with these precedents and with O’Brien and Hustler as well. The trail to that destination, however, is not easy to follow, and that is the key point for now. Put simply, Cohen v. Cowles Media Co. highlights how difficult it is to process the many and varied questions presented by free-speech cases that concern generally applicable laws. The remainder of this Article seeks to identify the key questions these cases present, as well as the most useful analytical tools for working through those questions in a sound way.

III.     Direct-in-Effect Burdens on Speech

How could it be that the Court applied a highly aggressive mode of First Amendment scrutiny in the Hustler case, but not in either O’Brien or Cohen v. Cowles Media Co.? One possible answer draws on First Amendment law’s specialized, speech-protective treatment of libel and slander. On this view, because the tort of intentional infliction of emotional distress bears a close resemblance to the tort of defamation, courts faced with constitutional challenges should take much of the same approach in both contexts. In fact, the form of elevated scrutiny that the Court applied in Hustler made use of the same “actual malice” test previously developed in New York Times v. Sullivan to deal with defamation directed at public officials and public figures.100 One way of thinking about Hustler, then, is as a logical build-out of the principle of the New York Times case. It might be that the Hustler rule—which applies aggressive First Amendment scrutiny in intentional-infliction-of-emotional-distress cases that involve public-officials or public-figure plaintiffs—was fashioned to counteract would-be work-arounds of state defamation law.101

This way of thinking about Hustler, however, is hard to defend. To begin with, the defamation tort specifically targets expression, while the intentional-infliction tort does not. This difference is of foundational importance because it means that defamation laws are not “generally applicable.”102 In contrast, the intentional-infliction tort simultaneously targets speakers and non-speakers alike. One might respond to this observation by saying that most cases of infliction of emotional distress involve speech that generates a “listener reaction” in the form of emotional upheaval.103 This view of things, however, is doubtful at best. The Restatement (Second) of Torts, for example, offers the following illustration of when a claim of intentional infliction of emotional distress is available:

A is invited to a swimming party at an exclusive resort. B gives her a bathing suit which he knows will dissolve in water. It does dissolve while she is swimming, leaving her naked in the presence of men and women whom she has just met. A suffers extreme embarrassment, shame, and humiliation. B is subject to liability to A for her emotional distress.104

It is apparent that the distress-inducing activity in the naked-swimmer case does not involve speech.105 Nor do many other actions that trigger the operation of the intentional-infliction tort.106 For this reason, it is hard to say that the cause of action for intentional infliction of emotional distress almost always involves speech.107 And this point is all the more salient because the Court in Hustler never justified its approach to that case on this ground.

So how can it be that the facts of Hustler and O’Brien differ in such a way that strict scrutiny applies in the former case but not in the latter? In an article published in 2005, Professor Eugene Volokh sought to answer this question by setting forth a theory that deals with one major category of problems concerning generally applicable laws.108 His answer built on the bedrock notion that laws almost always violate the First Amendment if they penalize speech based on its content.109 In Police Department v. Mosley, for example, the Court held that the First Amendment bars the government from outlawing all forms of picketing—except for labor picketing—near a school.110 Such a law is content-based on its face because its express terms distinguish between communication on labor-related and non-labor-related subjects. At the heart of Professor Volokh’s theory rests the thought that not all content-based laws work this way. Rather, some laws are content-based as applied.111 In such a case, the law is generally applicable in the sense that it targets both speakers and non-speakers in an effort to address a harm that is remediable by the government. The law gives rise to special free-speech problems, however, to the extent that it addresses that harm because the harm results from the content of speech itself.112

Hustler illustrates the point. There, the publisher committed the tort of intentional infliction of emotional distress as a matter of state law. Moreover, no one questioned the fact that this tort was redressable against non-speakers because of the government’s interest in protecting victims from outrageous inflictions of emotional injury. In Hustler itself, however, the publisher was a speaker. Moreover (and this is the key point), the emotional injury suffered by Reverend Falwell arose out of the particularly hurtful content of what the magazine had published. Put another way, the legally remediable harm came about precisely because Reverend Falwell (unlike, for example, the mortified naked swimmer) was reacting to particular words and images set forth in a communicated message. Thus, it was Reverend Falwell’s “listener reaction”113 to the content of the speech—that is, his mental response of shock and horror to the particular expression embodied in the cartoon—that gave rise to his claimed right of recovery. Because the content of the speech itself caused the harm that triggered the state-imposed sanction, the law of intentional infliction of emotional distress was content-based as applied.114

O’Brien is distinguishable from such a case for a simple reason: The relevant harm caused by the draft-card burner—unlike the relevant harm caused by Hustler Magazine—had nothing to do with the content of the defendant’s speech. Instead, the sole harm targeted by the draft-card-mutilation law had to do with the destruction of physical property that served a useful role in the administration of a government program. To be sure, O’Brien’s destruction of his draft card involved speech.115 But O’Brien tripped the switch of the draft-card-mutilation ban without regard to his engagement in expressive activity. Thus, in contrast to the situation in Hustler, the law in O’Brien was not content-based as applied.116

    Two refinements of Professor Volokh’s analysis merit attention. First, he does not (and cannot) suggest that governments may never put in place content-based laws. He acknowledges, in particular, that unprotected speech—such as incitement or obscenity—is properly subject to prohibition even though it is the content of such speech that renders it unprotected.117 Volokh also recognizes that courts can validate content-discriminatory laws, under strict scrutiny analysis, even if those laws do not target traditional categories of unprotected speech.118 According to Volokh, however, the government’s ability to restrict speech based on its content by way of generally applicable laws should reach no further than that; in other words, courts should permit the content-based-as-applied regulation of speech only to the extent that (1) the challenged legal restraint targets unprotected expression, or (2) the legal restraint, as applied, satisfies the strict-scrutiny test.119

By way of example, a generally applicable breach-of-the-peace statute that covers both expressive and non-expressive activity can operate to prohibit fighting words because fighting words are a form of unprotected speech. But such a statute cannot, under Professor Volokh’s approach, similarly operate on speech that, for example, embodies a problematic rant not “personally abusive” enough to qualify as fighting words under the present-day doctrine.120 Moreover, this limitation operates even though the non-fighting-words speech creates a sufficient risk of peace-breaching violence that the government could prosecute non-speech for triggering exactly the same degree of risk.

Second, Professor Volokh recognized that some generally applicable laws that operate to burden speech remain subject to First Amendment challenge even if they are not content-based as applied.121 This acknowledgement was hardly surprising because O’Brien endorsed that very point by deeming “intermediate scrutiny” applicable in many cases that involve no content discrimination at all.122 This Article will later explore how O’Brien is fairly subject to criticism, and all the more so in light of the Court’s intervening free-exercise ruling in Smith.123 For now, however, the critical point is that a case like Hustler (which involved a content-based-as-applied law) is distinguishable from a case like O’Brien (which did not involve a content-based-as-applied law) in a way that helps to explain why the Court applied a more aggressive form of scrutiny in the former case than in the latter.

This is not to say that the content-based-as-applied principle derived by Professor Volokh from Hustler and other rulings is uncontroversial, nor that the principle now operates in as sweeping a way as he posits it should.124 Notably, Volokh’s own survey of the work of lower courts indicates that they often fail to honor the content-based-as-applied principle—a finding that is hardly surprising in light of the analytical muddiness of the relevant Supreme Court precedents.125 Even so, Volokh’s content-based-as-applied principle sensibly reconciled key rulings of the Court, including Hustler and O’Brien. Even more important, that approach foreshadowed a major later development, which came with the Court’s ruling in Humanitarian Law Project.126

In Humanitarian Law Project, the Court confronted a statute, set forth in 18 U.S.C. § 2339B, that outlawed the giving of material support to terrorist organizations, including in the form of educational services regarding peaceable law-reform activities.127 The government contended (in an argument presented by then-Solicitor General Elena Kagan) that the challenged statute should escape strict scrutiny on the theory that, as in O’Brien, it operated in a generally applicable way.128 This approach made sense, so the government claimed, because § 2339B barred the giving of every form of material support—valuable arms transfers, valuable food transfers, valuable information transfers, etc.—in one fell swoop.129 The Court, however, rejected this position, deeming it necessary to apply a “more demanding” form of scrutiny than the law in O’Brien had triggered.130 In reaching this conclusion, Chief Justice Roberts reasoned:

The Government is wrong that the only thing actually at issue in this litigation is conduct, and therefore wrong to argue that O’Brien provides the correct standard of review. O’Brien does not provide the applicable standard for reviewing a content-based regulation of speech . . . and § 2339B regulates speech on the basis of its content. Plaintiffs want to speak to [terrorist organizations], and whether they may do so under § 2339B depends on what they say. If plaintiffs’ speech to those groups imparts a “specific skill” or communicates advice derived from “specialized knowledge”—for example, training on the use of international law or advice on petitioning the United Nations—then it is barred. See Brief for Government 33–34. On the other hand, plaintiffs’ speech is not barred if it imparts only general or unspecialized knowledge. See id., at 32.

The Government argues that § 2339B should nonetheless receive intermediate scrutiny because it generally functions as a regulation of conduct. That argument runs headlong into a number of our precedents, most prominently Cohen v. California. Cohen also involved a generally applicable regulation of conduct, barring breaches of the peace. But when Cohen was convicted for wearing a jacket bearing an epithet [that read “Fuck the Draft”], we did not apply O’Brien. Instead, we recognized that the generally applicable law was directed at Cohen because of what his speech communicated—he violated the breach of the peace statute because of the offensive content of his particular message. We accordingly applied more rigorous scrutiny and reversed his conviction.

This suit falls into the same category. The law here may be described as directed at conduct, as the law in Cohen was directed at breaches of the peace, but as applied to plaintiffs the conduct triggering coverage under the statute consists of communicating a message. As we explained in Texas v. Johnson: “If the [Government’s] regulation is not related to expression, then the less stringent standard we announced in United States v. O’Brien for regulations of noncommunicative conduct controls. If it is, then we are outside of O’Brien’s test, and we must [apply] a more demanding standard.”131

There are four possible ways to reconcile the Court’s application of strict scrutiny in Humanitarian Law Project with its application of only intermediate scrutiny in O’Brien: (1) as hinging on the challenger’s planned engagement in pure expression by way of teaching, as opposed to symbolic speech caught up with “conduct” as was present in the act of burning a draft card in O’Brien; (2) as hinging on the statute’s express textual definition of “material support” to include “training,” thus rendering the challenged law content-based on its face; (3) as hinging on the government’s concession that the statute applied only to instruction on “specific” subjects as opposed to “general” subjects; and (4) as hinging on the proposition that (à la Volokh) the statute was content-based in its application to the plaintiffs because, as to them, it targeted the communicative effects of speech.

The first reading would accord fewer protections to symbolic speech than to “pure speech,”132 even when a challenged law involves content discrimination.133 This view of the Court’s ruling is off base, however, because there is no sound reason for distinguishing between these two types of speech in applying content-discrimination rules.134 To be sure, some loose-lipped language of the Court in cases that predated Humanitarian Law Project creates a measure of confusion on this score.135 But the Court’s logic in Humanitarian Law Project itself cuts sharply against this would-be distinction for the simple reason that nothing in that logic suggests that protected speech (at least for purposes of the content-discrimination principle) falls into two different categories depending on whether or not it qualifies as “pure.” Indeed, in defending its application of strict-scrutiny to the case, the Court relied squarely on Texas v. Johnson—the flag-burning case—which involved not “pure” speech, but the immolation of a physical object.136

Some commentators have raised the possibility of the second distinction, which would lay critical weight on the express statutory identification of “training” as one form of prohibited “material support.”137 Again, however, in Humanitarian Law Project the Court did not hang its hat on this analytical hook. In particular, the Court made no mention of speech-related statutory language in deciding what level of scrutiny to apply. To the contrary, it centered its analysis on Cohen v. California, in which there was no suggestion that the challenged breach-of-the-peace statute had drawn a distinction on its face between speech-based and non-speech-based peace-disturbing activities.138 In any event, it would be wrong to find a First Amendment violation in Humanitarian Law Project only because the statutory text specifically defined “material support” to include “training.” This is so because, under such an approach, Congress would have been incentivized to leave out specific references to “training” even though courts would remain free—if not required—to view such activity as a proscribed form of “material support.” Put another way, deeming the statute in Humanitarian Law Project to be content-based solely because of its express mention of training would glorify form over substance, while rewarding statutory obliqueness and congressional obfuscation.

The third reading of Humanitarian Law Project—which accords determinative significance to concessions about the presence of content discrimination at pages 32–34 of the government’s brief—also fails to carry the day. In part, this line of analysis falters because the cited government submission, which noted the legal difference between “specific” and “general” training, merely echoed the recognized meaning of the material-support statute itself.139 In any event, this third proposed distinction inevitably collapses into the fourth. The concessions made by the government were important, after all, not because they were concessions; they were important because they confirmed how the statute operated—in particular, that it operated in a content-discriminatory way. In other words, any supposed concessions made by the government about “general” and “specific” training did not themselves trigger strict scrutiny. Rather, they merely confirmed that the statute was content-based as applied. And it was that proposition that set the stage for invoking the content-based-as-applied, strict-scrutiny principle that Professor Volokh had previously put forward.

Any doubt about the appropriateness of this fourth reading of Humanitarian Law Project is laid to rest by the Court’s treatment of Cohen v. California. At bottom, the Court described that case in exactly the way Professor Volokh would have described it. Yes, the breach-of-the-peace statute at issue in Cohen v. California sought in general fashion to protect the tranquility of the community, which the government’s police powers entitled it to do. Yes, the breach-of-the-peace statute swept up speech with non-speech in thus safeguarding community tranquility. Yes, the speech-based threat to community tranquility posed by displays of the F-word could be regarded as no less severe than pure conduct-based invasions of that tranquility (for example, through the “buzzing” of pedestrians with low-flying miniature airplanes). But none of that mattered because Cohen’s alleged breach of the peace was caused by the audience-distressing content of his speech. Thus, as the Court explained in Humanitarian Law Project, the law at issue in Cohen v. California “may be described as directed at conduct . . . [in that it] was directed at breaches of the peace, but as applied to plaintiffs the conduct triggering coverage under the statute consists of communicating a message.”140 For this same reason, the issue in Humanitarian Law Project was “not whether the Government . . . may prohibit material support in the form of conduct.”141 It was “instead whether the Government may prohibit what plaintiffs want to do—provide material support . . . in the form of speech.”142

Did Professor Volokh’s article lead the Court to analyze the Humanitarian Law Project case as it did? The Court never cited the article, although the challenger of the material-support statute had relied on Volokh’s work.143 In any event, who cares? The important point does not concern whether the Court meant to track Volokh’s analysis. The important point is that it did track that analysis, thus vindicating—as a matter of controlling Supreme Court doctrine—his approach to generally applicable laws that impose direct-in-effect burdens on speech.144

As courts move forward in applying this concept, they would do well to consider one build-out of Professor Volokh’s theory that seems sensibly grounded in the logic on which that theory rests. The attractiveness of this refinement arises because sometimes prohibitions that reach across both speech and non-speech activities can be unconstitutional based on their communicative effects even though they are not content-based as applied. Assume, for example, that a municipality has enacted an ordinance that prohibits the placement near a roadway of any object that might cause driver distraction. Assume also that a court has construed the ordinance to prohibit the near-road placement of all signage that contains words based on the rationale that drivers might be distracted by reflecting on the messages those signs convey. This law, to the extent it operates against word-bearing signs, would not seem to be content-based as applied because it bars all such signs without any regard to the content of the messages included on them. Even so, such a law raises profound First Amendment problems because it wipes out an entire set of communicative activities and does so because of the mental reactions that those communicative activities generate. For this reason, the content-based-as-applied verbiage of Professor Volokh may well fall short in its effort to capture the full set of cases to which his listener-reaction-centered principle should logically extend. An alternative articulation of that principle thus is offered here: At least absent special circumstances, a generally applicable law is subject to close judicial scrutiny because (1) it is content-based as applied, or (2) it otherwise directly burdens speech qua speech in its application to communicative activity. Laws that operate in either of these two ways stand in marked contrast to generally applicable laws, such as the statute at issue in O’Brien, that place only incidental burdens on speech. This Article turns now to an examination of the wide-ranging set of problems presented by incidental-burden cases.

IV.     Incidental Burdens on Speech

Professor Volokh’s article focused on generally applicable laws that merit strict scrutiny because they impose direct-in-effect burdens on speech. But what of laws, like the one in O’Brien, that penalize speech only incidentally—incidentally in the sense that the operation of the challenged law does not hinge on the communicative impact of the speech? In O’Brien, the Court considered the constitutionality of a generally applicable law that banned the destruction of military draft cards. No one questioned that this law operated in some cases without raising constitutional problems—for example, in its application to a pyromaniac or to a parent who mutilated a draft card to shield a child from learning he was subject to induction. In O’Brien itself, however, prosecutors targeted a political dissident who burned his card as part of a public protest against the Vietnam War.145 Although the Court upheld the conviction, it did not exempt generally applicable laws of this kind altogether from Free Speech Clause attack. To the contrary, the Court considered whether the law was invalid as applied under a specialized form of “intermediate scrutiny.”146 In particular, the Court stood ready to reject a speaker’s challenge to such a law only

if it is within the constitutional power of the Government; if it furthers an important or substantial governmental interest; if the governmental interest is unrelated to the suppression of free expression; and if the incidental restriction on alleged First Amendment freedoms is no greater than is essential to the furtherance of that interest.147

In short, the Court’s free-speech ruling in O’Brien authorized and invited as-applied challenges to generally applicable laws under the First Amendment.

From O’Brien’s earliest days, its intermediate-scrutiny approach to incidental-burden, free-speech cases has engendered criticism.148 But the chorus of boos has grown especially loud in the wake of the Court’s ruling in Smith.149 That case involved the application of a general criminal prohibition on peyote use to a religious practitioner who ingested peyote as part of a sacramental activity recognized by his faith tradition.150 In a landmark ruling, the Court held—in the face of strong counter-pressures raised by prior rulings—that no Free Exercise Clause challenge was available to the religious practitioner because the peyote ban was “a generally applicable criminal law.”151 Relying on Smith, Justice Scalia later argued that the same no-scrutiny approach endorsed in Smith for free-exercise cases should carry over to free-speech challenges to generally applicable laws in cases like O’Brien.152Smith thus put a bullseye on O’Brien’s back, and later rulings—including Cohen v. Cowles Media Co.153—have raised additional concerns about O’Brien’s continued soundness.154 All of these developments have brought to the fore three key questions about the current status of the O’Brien principle:

(1)             Do Smith and other rulings that post-date O’Brien call for abandonment of its intermediate-scrutiny approach in cases that involve generally applicable laws that impose incidental burdens on speech?

(2)             Even if later cases do not dictate an abandonment of O’Brien’s intermediate-scrutiny rule, do those cases—and particularly Cohen v. Cowles Media Co.—dictate that courts should now recognize significant exceptions to that rule?

(3) Regardless of the impact of rulings such as Smith and Cohen v. Cowles Media Co., has the real-world operation of the O’Brien principle revealed that it has so little speech-protective effect that retaining it is no longer justifiable?

As the following discussion will reveal, these questions are not subject to quick and easy answers. To make a long story short, however, the answers to them turn out to be “no,” “no,” and “no.”

A.     The Precedent-Based Case for Abandoning O’Brien

Do post-O’Brien rulings of the Supreme Court dictate that the Court should abandon that decision’s intermediate-scrutiny Free Speech Clause analysis, thus extending the no-review approach applicable in doubly-incidental-burden cases to incidental-burden cases as well? As it turns out, two separate precedent-based arguments exist for concluding that O’Brien is no longer good law. One line of argument, which focuses on Free Speech Clause precedents (prominently including Cohen v. Cowles Media Co.), is considered and rejected in another work.155 The contention addressed here is the one based on Smith—namely, that the hands-off approach to free-exercise cases endorsed by the Court in that case should logically carry over to the free-speech context.

As we have seen, the Court’s ruling in Smith established that a sincere religious believer engaged in even sacramental activity of central importance to his faith lacked the ability to challenge on free-exercise grounds the operation of a generally applicable law.156 In support of this result, Justice Scalia reasoned that every individual could “become a law unto himself” if courts had to recognize exceptions to the operation of laws that restrict specified forms of problematic behavior when that behavior stems from religious reasons.157 To be sure, Justice Scalia acknowledged the existence of precedent-driven limits on the no-exemption rule set forth in Smith.158 But the basic holding endorsed by five Justices had a sharp-edged quality: “[I]f prohibiting the exercise of religion . . . is not the object . . . but merely the incidental effect of a generally applicable and otherwise valid provision, the First Amendment has not been offended.”159

The potential implications of the Court’s free-exercise ruling in Smith are far-reaching for free-speech law. The essential argument goes like this: If the Free Exercise Clause does not protect participants in religious activity from incidental burdens imposed on them by generally applicable laws, the Free Speech Clause—which is located right next door in the same First Amendment—should likewise not protect participants in expressive activity from the incidental burdens imposed on them by exactly those same laws. There is no apparent textual difference between the Free Exercise Clause and the Free Speech Clause in this regard. So, if generally applicable laws do not abridge free-exercise rights for the sole reason that they are generally applicable (rather than targeted at religion), why should we not say that generally applicable laws likewise do not abridge free-speech rights because they are generally applicable (rather than targeted at speech)? The bottom line, according to this reasoning, is that Smith dictates that O’Brien is no longer good law. Indeed, Justice Scalia took little time after penning the Court’s opinion in Smith to advance precisely this position.160 Not surprisingly, several scholars have agreed that there is no meaningful distinction between free-speech and free-exercise cases in this respect.161

Careful examination suggests, however, that there exist strong reasons for declining to extend the no-review, free–exercise ruling in Smith to O’Brien-type free-speech cases. The overarching point is that the Free Exercise Clause and the Free Speech Clause operate in different contexts to protect different values.162 In particular, the Court’s ruling in Smith found support in a distinctive concern about religious-liberty claims—namely, that singling out religious practitioners for special treatment in applying generally applicable laws creates a tension with the constitutional norm, rooted largely in the Establishment Clause, of ensuring the state’s “complete neutrality toward religion.”163 Stated another way, exempting members of particular religious traditions from laws that apply to everyone else smacks of advantaging both religion in general and some religions over others. But a core goal of the First Amendment is to foreclose exactly this sort of governmental preferentialism.164 These concerns are not present when it comes to applying the Free Speech Clause.165 And so it follows that the ruling of Smith does not easily carry over to the free-speech context.

In addition, the Free Speech Clause plays a distinctly critical role in ensuring the proper operation of our system of republican self-government. The capacity of any society to govern itself depends at its root on affording that society’s members wide-ranging opportunities for criticism (and defense) of the political and social status quo.166 Given this reality, there is reason to look with heightened skepticism on any laws that limit free expression. Notably, the Court has done just that in framing free-speech doctrine—for example, by developing special liberty-protective rules with regard to overbreadth, vagueness, and prior restraints that do not apply outside the free-expression context.167 From these points, it does not follow that the free-speech right is more important in all respects than the free-exercise right. It does follow, however, that the two rights serve different purposes within our constitutional system.168 Indeed, the Framers foresaw that creation of an open society made up of citizens with widely divergent world views—that is, just the sort of open society that an aggressive application of the Free Speech Clause helps to foster—would itself help to safeguard religious liberty.169 It may well be for this reason that the Court has declared that freedom of speech “is the matrix, the indispensable condition, of nearly every other form of freedom.”170 Against this backdrop, it makes sense that the Court would apply the First Amendment with added vigor in assessing state-imposed limits on speech, including limits that stem from generally applicable laws.171

Free-speech rulings that post-date Smith support this understanding. Indeed, in Holder v. Humanitarian Law Project, the Court accepted O’Brien as a relevant precedent and worked hard to show why its principle was not controlling with respect to the constitutionality of the material-support statute.172 Other post-Smith rulings likewise proceeded from the premise that O’Brien remains good law.173 There is, however, a fly in the ointment. As detailed in Part II, the Court in Cohen v. Cowles Media Co. refused to apply any form of Free Speech Clause scrutiny to a law it deemed generally applicable in character.174 A question thus now exists about how far Cohen v. Cowles Media Co. goes in limiting the operation of the O’Brien rule.

B.     Precedent-Based Limits on O’Brien

In Turner Broadcasting System, Inc. v. FCC, the Court—with a nod of the head in the direction of Cohen v. Cowles Media Co.—observed that “a generally applicable law may or may not be subject to heightened scrutiny under the First Amendment.”175 This passage is rife with obscurity. The context in which it appeared, however, suggests that the Court meant to say that O’Brien-style intermediate scrutiny will not apply in every incidental-burden, free-speech case.176 This same conclusion finds support in the Court’s failure to apply any form of review in Cohen v. Cowles Media Co. itself.177 So just how far does the exception to O’Brien recognized in Cohen v. Cowles Media Co. reach?

The answer to this question is shrouded in uncertainty, but two possibilities seem worthy of consideration. Under the first approach, courts would not apply any form of scrutiny in incidental-burden cases (including O’Brien-style intermediate review) so long as the challenged generally applicable law imposed a traditionally recognized restriction on individual liberty.178 Take, for example, the prosecution under a murder statute of an assassin of a political leader. In such a case, the court might conclude that the assassination—much like the burning of a flag—in its nature embodies a publicly communicated expression of opposition to the government. In other words, an assassination might well be seen as speech under the Court’s current definition of that term.179 Even so, many analysts would hesitate to apply O’Brien-based intermediate scrutiny in such a case even if the murder law as applied to the assassin does impose an incidental burden on the free-speech right.180

For some analysts, the reason for this hesitance might lie in the thought that the Founding Fathers would never have imagined a First Amendment challenge to a murder law. For others, the driving force might be that, whatever outlooks held sway in 1791, modern judges should consider “our tradition” and “longstanding history”181 in developing constitutional-decision rules.182 For another group of objectors, the reason for skepticism reflects the thought that some legal prohibitions (beginning with murder laws) are so central to the social order that they should be deemed constitutionally unassailable. And for others (including me), the source of hesitance springs from a more visceral and primal source: It simply makes the skin crawl to think that an assassin might assert a First Amendment defense to a murder prosecution.

Before tempers flare at the thought of letting assassins invoke O’Brien, however, it is worth recalling that exposing a murder law to intermediate scrutiny hardly means that assassins will go free. Indeed, if any First Amendment challenge is destined to fail this form of review, it is sure to be the one put forward by a cold-blooded killer.183 In any event—and whatever our intuitions might be about the assassin-murder-statute case—there is no good reason to view Cohen v. Cowles Media Co. as having installed a traditional-prohibition limit on the intermediate-scrutiny rule. Among other things:
(1) not one word in the Court’s opinion assigned significance to the traditional acceptance of the promissory-estoppel cause of action;184 (2) this omission was not surprising in light of the relatively recent recognition of the promissory-estoppel concept and the common law’s long history of not enforcing promises based solely on detrimental reliance (as opposed to bargained-for consideration);185 (3) significant line-drawing difficulties would arise under a regime that required judges to distinguish between traditionally-recognized and not-traditionally-recognized legal restrictions (as Cohen v. Cowles Media Co. itself indicates); and (4) in any event, there was and is no demonstrable need to embrace a traditional-prohibition exception to O’Brien, as illustrated by the example of the assassin-murderer case itself, since any O’Brien-based challenge to the murder law raised by the assassin would obviously fail.186

So, what is the second way to think about the exception to the O’Brien rule created by Cohen v. Cowles Media Co.? This question circles us back to the work of Professor Volokh. In his treatment of generally applicable laws, Volokh concluded that Cohen v. Cowles Media Co. is best read as rejecting the defendants’ free-speech-based defense on a narrow ground—namely, on the ground that the reporter waived any First Amendment rights by way of his express agreement not to disclose Mr. Cohen’s identity.187 A major problem with this waiver theory is that the Court’s wide-ranging treatment of the First Amendment issue in Cohen v. Cowles Media Co. does not offer much support for it. At the same time, the Court’s meandering analysis does not provide much (if any) support for any other theory either.188 As a result, there is no choice but to extrapolate a limit on O’Brien’s operation from what few signals about the nature of that limit the Court’s analysis in Cohen v. Cowles Media Co. supplies.

Against this backdrop, some form of agreement-based waiver theory probably does the best job of explaining why the Court declined to apply any First Amendment scrutiny in Cohen v. Cowles Media Co. To begin with, the Court did mention—indeed, it emphasized—the fact that the reporter-defendant had promised not to disclose the plaintiff-source’s name.189 In addition, the waiver theory has moorings in long-accepted notions of equitable behavior. Commonly endorsed fairness-based norms—such as not being able to “have it both ways”190 or to “bait and switch”191 or to avoid “tak[ing] the bitter with the sweet”192—all were at play in the promissory-estoppel case precisely because it involved foreseeable reliance on a focused promise that the promisor later willfully breached. In addition, two considerations tied to the reporter’s voluntary agreement reduced the force of First Amendment policy concerns in Cohen v. Cowles Media Co. First, as a systemic matter, upholding the promissory-estoppel cause of action may have advanced, rather than undermined, the animating purposes of the Free Speech Clause by maximizing the net overall flow of socially useful information. On this view, by vindicating legal protections for anonymity-seeking news sources, Cohen v. Cowles Media Co. made it easier for potential sources to step forward to share information otherwise unavailable to the public.193 Second, any worry about government suppression of disfavored speech—that is, the core worry that underlies the free-speech protection—was at a low ebb in Cohen v. Cowles Media Co. Why? Because, as the Court itself noted, the state sanctioned the defendants for disclosing information in violation of only a purely private agreement, as opposed to an edict handed down from on high by the rightly watch-dogged institutions of government.194

Even if one concludes that a waiver theory best explains the Court’s treatment of the free-speech issue in Cohen v. Cowles Media Co., it remains to be asked how far the resulting waiver-based limit on the operation of O’Brien properly extends. On this point, Professor Volokh may have skipped over a significant limiting principle. In particular, it may be that a voluntary agreement not to exercise speech rights does not always preclude the assertion of First Amendment defenses in a follow-up, breach-of-promise case. Rather, the best way to view Cohen v. Cowles Media Co. may be to see it as supporting only a limited waiver exception to O’Brien—that is, an exception hemmed in by important public-policy considerations.195 Along these lines, some commentators have called for judicial recognition of constitutional limits on the enforceability of private employment contracts that broadly bar the disclosure of information of significant public importance.196 Illustrative would be a case in which an employee observes a food-seller employer’s trafficking in germ-infested meat after signing an employment contract that bars any disclosures about the employer’s business practices. Notably, in such a case, some key waiver-based justifications for barring the defendants from challenging the promissory-estoppel law in Cohen v. Cowles Media Co. are not applicable. To begin with, any disclosure of information about germ-infested meat would not bring into play the same “bait-and-switch” concerns raised by breaching a focused promise not to disclose a highly discrete piece of information concerning only an informant’s name.197 In addition, the rationale rooted in fostering the overall flow of newsworthy information, by enforcing reporters’ promises not to disclose a source’s name, has no application whatsoever in cases such as the one involving germ-infested meat. In Cohen v. Cowles Media Co., after all, the whole point of promising not to disclose a source’s identity was to obtain information that otherwise would not be available to the public. But no such non-speech-for-more-speech tradeoff is in the picture in cases that involve only the enforcement of standard employment contracts.

At the least, breach-of-promise cases like the one involving germ-infected meat raise public-policy concerns that some courts are sure to deem significant as they wrestle with waiver-related First Amendment issues. And that fact alone suggests that whatever exception Cohen v. Cowles Media Co. overlays on O’Brien is likely to prove to be a narrow one. After all, if courts balk at applying even a waiver-centered exception to cases involving explicit promises not to speak, there is good reason to believe that they will hesitate to extend that exception to cases that do not involve claims of waiver at all.

C.     Abandoning O’Brien on Policy Grounds

The preceding discussion shows that recent decisions of the Court do not support abandoning O’Brien’s intermediate-scrutiny approach for generally applicable laws that only incidentally burden speech.198 That fact, however, does not mean that applying intermediate scrutiny to such laws is a good idea. In Barnes v. Glen Theatre, Inc., for example, Justice Scalia argued that the Court should extend the Smith no-review rule to incidental-burden, free-speech cases based in part on specialized normative grounds.199 The thrust of his argument was that social-order concerns that favor across-the-board compliance with legal restrictions carry special weight in the free-speech context because “almost anyone can violate almost any law as a means of expression.”200 As earlier analysis suggests, however, there are countervailing considerations that lend support to just the opposite approach. According to this line of analysis, the centrality of the free-speech right to a successful system of self-rule logically supports a judicial stance that directs especially meaningful scrutiny of generally applicable laws that impose burdens on communicative (as opposed to religious) liberty.201

Faced with this clash of normative arguments, lower courts might hesitate to push aside the analysis put forward by Justice Scalia. But in this instance, they should not hesitate at all. The reason why is that Justice Scalia found himself writing in Barnes only for himself. No other Justice expressed even a hint of support for his view that the Court should abandon the long-recognized approach of O’Brien; indeed, each of them readily applied the principles of O’Brien in deciding the issue put forward in Barnes.202 In short, Justice Scalia’s policy argument for abandoning O’Brien was aired before the full Court, and the Court declined to embrace it.

There is, however, another normative argument—one not advanced by Justice Scalia—for doing away with O’Brien’s intermediate-scrutiny methodology. This argument rests on the idea that the O’Brien-style intermediate-scrutiny rule is so feckless in actual operation that there is no good reason to keep it in place.203 The case for viewing the O’Brien rule as an empty vessel begins with O’Brien itself, in which the Court—even while applying intermediate scrutiny—had no difficulty upholding the challenged draft-card-mutilation law. Likewise, in Barnes and other post-O’Brien cases the Court has validated challenged statutes when called on to apply the O’Brien test.204 In Cohen v. Cowles Media Co., the Court also created an indeterminate exception to O’Brien that threatens further confusion by rendering its intermediate-scrutiny principle inapplicable in an as-yet-undefined set of incidental-burden cases.205 Thus, so the argument goes, any benefits that O’Brien might otherwise yield are overwhelmed by the real-world costs it creates in requiring courts to process cases that are almost certain to fail—costs that are significant in terms of undermining certainty in the law, efficiency in its operation, and the consistent treatment of litigants.206 It follows, so the argument concludes, that the Court should abandon the O’Brien approach on normative grounds, regardless of any precedent-based challenge to it under Smith and other authorities.

For at least three reasons, this argument is unpersuasive. First, in a variety of contexts (albeit ones that do not involve typical incidental-burden cases), the Court has found that statutes challenged under the First Amendment do fail the intermediate-scrutiny test, thus demonstrating that the application of this mode of review can be more—indeed, much more—than a make-work exercise.207 Second, the sample size of Supreme Court decisions that directly involve O’Brien is small, and lower courts in fact have repeatedly struck down laws under the banner of O’Brien’s intermediate-scrutiny approach.208 Finally, Arcara v. Cloud Books, Inc.—the case that will take center stage in the remainder of this Article—signals that intermediate scrutiny of generally applicable laws pursuant to the O’Brien principle is far more than a meaningless act.209 In that case, the New York Court of Appeals invoked O’Brien in overturning the state’s closing of a bookstore because episodes of unlawful conduct had occurred on its premises.210 The Supreme Court reversed, but it did so only because it declined to apply the O’Brien standard, while never questioning the lower court’s determination that intermediate scrutiny required it to afford the bookstore owners with judicial relief. Arcara, in short, illustrates that the Court neither does, nor should, view intermediate scrutiny under O’Brien as so impotent that it should be abandoned.

To be sure, one might argue that the case for discarding the O’Brien rule gains (rather than loses) force in light of Arcara because that ruling, like the ruling in Cohen v. Cowles Media Co., creates a range of cases in which the O’Brien rule does not apply at all. This argument, however, suffers from problems of its own. To begin with, Arcara creates a limiting principle that reaches only cases in which a generally applicable rule imposes a doubly incidental burden on expression. Because many cases—including O’Brien itself—involve incidental (rather than doubly incidental) burdens on speech, the principle of that case still­ has much work to do. In any event, as the next Part of this Article shows, the Arcara rule itself has spawned a complex mix of exceptions that in practical effect extend the operation of O’Brien to many doubly-incidental-burden cases. The next part of this Article considers the nature of the rule established by Arcara and the many exceptions to that rule that post-Arcara judicial decisions and scholarly commentary have advanced.

V.     Doubly Incidental Burdens on Speech

The dispute in Arcara centered on a New York statute that authorized the closing of any business that served as a site for prostitution and other forms of lewd behavior.211 Based on this statute, the trial court closed an adult bookstore because persons on the premises, with the knowledge of the store’s operator, had offered to sell sexual services and openly engaged in masturbation and fellatio.212 The New York Court of Appeals overturned the trial court’s order.213 It reasoned that the law, as applied in this case, failed the “less restrictive means” prong of the O’Brien test because an injunction that barred only the lewd activities themselves would adequately address the evils the state sought to remedy.214

The Supreme Court disagreed. It concluded that the limiting principle of O’Brien did not apply at all on the facts presented, so that neither intermediate scrutiny nor any other form of Free Speech Clause review was operative in the case.215 The Court reasoned that O’Brien was distinguishable because it involved the penalizing of conduct—namely, the mutilation of a draft card, a public protest that itself embodied a “significant expressive element.”216 In contrast, the prostitution and open sexual activities subject to sanction in Arcara involved “no element” of expressive behavior, so O’Brien was beside the point.217

The majority rooted its drawing of this distinction in a parade-of-horribles rationale. It noted, for example, that a person subjected to a money damages award in a civil suit might thereby become far less able to publish political messages.218 But it could not be the case that the risk of a civil damages award in any sort of tort or breach-of-contract lawsuit would generate a potential First Amendment defense.219 In a concurring opinion, Justice O’Connor made the same point, describing such a result as “absurd.”220 Surely, she explained, the arrest of a reporter for a traffic violation should not trigger First Amendment scrutiny even if (to build out her thought) that arrest would impede the reporter’s ability to arrive on time for an important newscast.221Arcara, a majority of the Court concluded, was indistinguishable from cases like the one involving the delayed reporter because lewdness (like speeding) did not embody protected expression and the statutory closure remedy applied to any business, not just bookstores.

In short, Arcara launched a new principle that recognized and responded to a third type of generally applicable law. Arcara did not involve a direct-in-effect burden on speech (as in Humanitarian Law Project) or an incidental burden on speech (as in O’Brien) because the sanctioned conduct (namely, prostitution and lewdness) did not involve speech at all. Rather, to use the rhetoric put forward in this Article, the case involved only a doubly incidental burden on expression, and the Court concluded that this sort of burden triggered no First Amendment scrutiny whatsoever.

The story of Arcara, however, did not end there. To begin with, the case produced a vigorous dissent. In an opinion joined by Justices Brennan and Marshall, Justice Blackmun decried the Court’s failure to apply any scrutiny at all to a law that had the practical effect of closing down an entire bookstore engaged overwhelmingly in protected First Amendment activity.222 The majority seemed ready to tolerate this result because it would bring to this sprawling field of law the benefits of a bright-line, state-protective rule. The dissenters, however, emphasized the costs that such a wooden approach would impose:

The Court’s decision creates a loophole through which counties . . . can suppress “undesirable,” protected speech without confronting the protections of the First Amendment. Until today, the Court has required States to confine any book banning to materials that are determined, through constitutionally approved procedures, to be obscene. Until today, States could enjoin the future dissemination of adult fare as a nuisance only by “adher[ing] to more narrowly drawn procedures than is necessary for the abatement of an ordinary nuisance.” A State now can achieve a sweeping result without any special protection for the First Amendment interests so long as the predicate conduct—which could be as innocent as repeated meetings between a man and a woman—occurs on the premises. . . . [W]hen a State’s only intention is to eliminate sexual acts in public, a 1-year closure has a severe and unnecessary impact on the First Amendment rights of booksellers.223

Perhaps in response to this expression of alarm, the majority recognized an exception to Arcara in Arcara itself: Its limiting principle would not apply when the challenged law had the “inevitable effect” of singling out protected speech for adverse treatment.224 The Court in Arcara also in effect recognized another exception, derived from its earlier decision in United States v. Albertini,225 by indicating that the O’Brien rule would remain operative so long as an “intimate relation” exists between non-speech activity targeted by a generally applicable law and separate activity subject to First Amendment protection.226 Post-Arcara rulings also have signaled the Court’s willingness to review speech-restricting laws in another set of doubly-incidental-burden cases—in particular, when the legislature’s purpose in enacting a generally applicable law is to stamp out protected forms of speech.227 In addition, the Court has issued doctrinal pronouncements in other fields of constitutional law—with regard to the so-called “hybrid rights”—that lay the groundwork for recognition of yet another exception to Arcara.228 Lower court judges have joined in this effort too, asserting that Arcara should not control in (1) cases in which the generally applicable law regulates concededly lawful activity, as opposed to activity that is wrongful and therefore prohibited229 and (2) cases in which the generally applicable law targets communicative activity not protected by the First Amendment (such as obscenity) as opposed to non-communicative activity not protected by the First Amendment (such as prostitution).230 Finally, thoughtful commentators have proposed approaches that would significantly limit Arcara’s reach. According to Michael Dorf, courts should abandon Arcara’s no-review approach whenever the challenged law imposes a “substantial” burden on speech.231 And Srikanth Srinivasan has advanced a theory under which courts should refrain from applying Arcara when they detect an intolerably high risk that content discrimination is afoot in the operation of the challenged law.232

This Article cannot provide a full review of each of these actual or potential carve-outs from the Arcara rule. Even so, a quick assessment is offered here for two reasons. First, these exceptions are themselves important features of the overarching architecture of Free Speech Clause limits on generally applicable laws. Second, the number, scope, and complexity of these limits raise questions about whether the Court’s current approach to doubly-incidental-burden cases has become so unstable that it is unlikely to endure. At the least, a brief look at the law in this area leaves no doubt that this question is deserving of serious, continuing reflection.

The exceptions to Arcara set forth above fall into four categories:
(1) exceptions already endorsed by the Supreme Court; (2) exceptions put forward by lower courts; (3) exceptions potentially derived from Supreme Court doctrines not yet explicitly tied to Arcara; and (4) exceptions proposed by legal scholars. The remainder of Part V shows that each of these categories raises questions of interest and importance.

A.     Exceptions Recognized by the Supreme Court

1.     The Inevitable-Effect Exception

The Court in Arcara endorsed in express terms one exception to the rule set forth in that case: Judges would continue to apply elevated First Amendment scrutiny when a generally applicable regulation of “nonexpressive activity has the inevitable effect of singling out those engaged in expressive activity.”233 The Court grounded this exception in its earlier ruling in Minneapolis Star and Tribune Co. v. Minnesota Commissioner of Revenue,234 where it had struck down “a tax imposed on the sale of large quantities of newsprint and ink because the tax had the effect of singling out newspapers.”235 Although “the tax was imposed upon a nonexpressive activity,” its burden “inevitably fell disproportionately—in fact, almost exclusively—upon the shoulders of newspapers exercising the constitutionally protected freedom of the press.”236 In contrast, the anti-lewdness law at issue in Arcara, like most “governmental regulations of general applicability,”237 did not “impose a disproportionate burden upon those engaged in protected First Amendment activities.”238 Therefore, it fell outside the reach of the Minnesota Star principle.239

The inevitable-effect exception, not surprisingly, has given rise to difficult line-drawing problems.240 One lower court case, for example, concerned a so-called anti-paparazzi law, which requires the imposition of heightened punishments for traffic offenses motivated by a driver’s desire to obtain information for personal gain.241 Does such a law fall within the Minnesota Star exception because it most often applies to news gatherers? Or is the exception inapplicable because the law targets not only news gatherers but also private investigators or other persons seeking information for pay? A California appellate court concluded that the statute was constitutional, rejecting the argument that it singled out speakers “as a practical matter.”242 In doing so, the court reasoned that “taking photographs and making recordings for personal gain are not always or necessarily journalistic activities” because such activities might be pursued by a private eye or a blackmailer.243 This argument is hardly airtight, however, because private eyes and blackmailers—no less than paparazzi journalists—may well seek and secure photographs and sound recordings to facilitate follow-up communicative activities.244

Anti-paparazzi statutes highlight the difficulty of applying the Minnesota Star principle across a complex range of generally-applicable-law cases.245 Moreover, this difficulty is heightened by the rhetoric of Arcara itself. The problem arises because the majority in that case described the sort of laws stripped of protection by Minneapolis Star with a hodgepodge of verbal constructs—speaking of “a disproportionate burden” on speakers at one point;246 “the inevitable effect of singling out” speakers at another;247 and consequences that “fall disproportionately—in fact, almost exclusively” on speakers at yet another.248 Uncertainty-creating linguistic formulations are inescapable in law. But uncertainties are compounded for no good reason when the Court states the same principle in shifting terminology in the process of declaring that principle’s existence.

2.     The Intimate-Relation Exception

In Arcara, the Court distinguished O’Brien on the ground that there “the ‘nonspeech’ which drew [the] sanction was intimately related to expressive conduct protected under the First Amendment.”249 As we have seen in O’Brien itself, the intimate relationship that existed between protected speech and unprotected conduct was self-evident because—as the Court put the point in Arcara—“it was conduct with a significant expressive element that drew the legal remedy . . . .”250 In another pre-Arcara ruling, however, the Court signaled that the limits imposed by O’Brien do not apply only when the generally applicable law takes aim at protected speech itself. United States v. Albertini arose out of the proper issuance by military authorities of a so-called “bar letter” that prohibited the defendant from reentering Hickam Air Force Base without a grant of specific permission issued by military authorities.251 A federal statute criminalized the violation of such bar letters regardless of the reason for entry.252 Even so, Albertini came onto the base without permission to participate in a political demonstration.253 In response to the ensuing federal prosecution for this conduct, he argued that his entry onto the base in violation of the bar letter was protected by the First Amendment.254 The Court rejected this argument, but in doing so it did not declare, along the lines of Arcara, that the Free Speech Clause had no role to play; rather, it applied intermediate scrutiny pursuant to O’Brien.255 Of even greater importance for present purposes, the Court in Arcara itself distinguished Albertini without questioning in any way the Court’s application in that case of O’Brien-style review.256 Put another way, the Court in Arcara recognized an exception to the Arcara rule—wholly separate from the Minnesota Star exception—rooted in Albertini.

What is the nature of this exception? Or (to put the same question differently), how did Albertini differ from O’Brien in such a way that it gave rise to a major extension of the O’Brien intermediate-scrutiny principle that the Court in Arcara carried forward? The answers to these questions hinge on one key fact—namely, that the actus reus of the crime in Albertini involved merely entering the grounds of the base.257 This fact is of critical significance because the act of entering the base did not (in contrast to O’Brien’s burning of his draft card) embody communicative behavior.258 Rather, the relevant expressive behavior came only later, when Albertini thrust himself into the on-base demonstration. For this reason, Albertini reflected a significant expansion of O’Brien, and (as a result) a significant limitation on the no-review principle made applicable by Arcara to doubly-incidental-burden cases.

The Court in Arcara posited that this limitation would operate in cases where the statutorily prohibited conduct was “intimately related” to one’s engagement in protected speech. But just how far does this “intimate relation” exception to the Arcara rule extend? One possibility is to say that conduct undertaken to facilitate expressive activity is ipso facto “intimately related” to that activity, so that generally applicable laws that outlaw such facilitative activity must survive intermediate scrutiny notwithstanding Arcara.259 A principle this broad in scope, however, would plainly reach too far.260 Indeed, the hypothesized case that involved unlawful speeding by a late-running news reporter—that is, the very case put forward by Justice O’Connor as a paradigmatic example of when Arcara’s no-scrutiny rule has to apply—itself involved interference with action undertaken to facilitate the engagement in protected speech.261 Perhaps Albertini stands for the proposition that securing physical access to a site of speech should automatically qualify as intimately related to that speech. Such a principle would, however, produce discomforting results as well. Why, for example, should O’Brien, rather than Arcara, apply to the operation of a generally applicable trespass law that bars unauthorized entry onto my property simply because the trespasser wants to stand on my porch to denounce my political views?262

There is also no self-evident reason why the intimate-relation exception should apply only to activities that facilitate speech. Why, for example, should exiting property be distinguished from entering property? It seems odd to say, for example, that an Albertini-like base-invader can mount a First Amendment defense to a prosecution for improperly scaling a fence to join a political demonstration but not to a prosecution for exiting exactly the same base over exactly the same fence after exactly the same demonstration is over. The overarching point is that the “intimate relation” exception hangs over the Arcara principle in an uncertainty-engendering way.263 As with the “inevitable effect” exception of Minnesota Star, the intrinsic amorphousness of the term “intimately related” opens the door for innovative judicial efforts to restrict the operation of Arcara’s no-review rule.264

3.     The Improper-Purpose Exception

In Arcara, the majority emphasized that a person who encountered the application of an otherwise unobjectionable generally applicable law could assert a First Amendment claim if the government’s action stemmed from a wrongful speech-targeting purpose.265 One set of improper-purpose cases involves the wrongful speech-suppressing enforcement of a generally applicable law. In State ex rel. Konstam v. Video Visions, Inc., for example, the Ohio Court of Appeals confronted the closure of an adult bookstore, based on known acts of masturbation in video booths, pursuant to an ordinance that authorized the shutting down of business establishments wherever such acts of “lewdness” occurred.266 Not surprisingly, the State resisted the bookstore operator’s First Amendment defense by invoking Arcara, which at first blush seemed to control the case in a directly-on-point fashion.267 Things became trickier, however, when the “[a]ppellants produced overwhelming evidence that the nuisance investigation against Video Visions was pretextual in nature.”268 Specifically, the town had repeatedly failed in earlier efforts to obtain criminal obscenity convictions against the store’s operators; its governing council had then brought in experts to provide advice on “dealing with pornography problems”; and the nuisance investigation directed at the bookstore “was instituted immediately after” the council instructed local enforcement authorities to take action “to fight ‘hard-core’ pornographic materials.”269 On these facts, the Ohio court concluded that Arcara did not preclude, but instead supported, defendants’ First Amendment challenge.270 Put another way, the court recognized that an improper purpose in bringing an enforcement action—here to root out protected “pornography,” as opposed to unprotected “obscenity”—would render unconstitutional the application of a law otherwise unobjectionable under the principle of Arcara.271

A second type of improper-purpose problem concerns not the enforcement, but the adoption, of generally applicable laws. Assume, for example, that no anti-lewdness law was on the books prior to the events that gave rise to State ex rel. Konstam. Assume further that the city council enacted a new anti-lewdness law because it was unhappy with Video Vision’s operations (even though it purveyed no obscenity) and that the legislative record showed that the council took this action “just so we can keep such places from selling pornographic filth.” Assume finally that an enforcement officer, with no speech-related motives at all, thereafter brought a proceeding to shut down the business upon discovering that an open act of masturbation had occurred on site. Would Video Visions have a First Amendment defense against the city’s action?

Of no little importance with respect to this question, the Court in O’Brien itself declared that “this Court will not strike down an otherwise constitutional statute on the basis of an alleged illicit legislative motive.”272 This rhetoric, however, stands in no small tension with the Court’s more recent free-speech-law pronouncements, which suggest that strict scrutiny applies whenever “the purpose and justification for the law are content based.”273 In addition, a never-consider-legislative-purpose principle is hard to square with post-O’Brien rulings under the Equal Protection,274 Establishment,275 and Free Exercise276 Clauses.277 To be sure, the difficulty of deciphering the shared purposes of a multi-member legislative body278 counsels against judicial adventurism in making findings of wrongful motives.279 Even so, under the full corpus of present-day law, there can be little doubt that courts would hesitate to apply the immunity otherwise bestowed by Arcara when “overwhelming evidence” indicates that lawmakers passed the challenged law with the specific purpose of extinguishing protected speech.280

B.     Exceptions Put Forward by Lower Courts

Lower courts have worked hard to apply the exceptions to the Arcara rule recognized by the Supreme Court, but judges on these courts also have done something more: They have advocated the recognition of additional exceptions to the Arcara rule. Two potential exceptions—one dealing with lawful activities and the other dealing with unprotected speech—illustrate how this process has unfolded.

1.     The Lawful-Activity Exception

The Pennsylvania Supreme Court’s ruling in In re Condemnation by Urban Redevelopment Authority of Pittsburgh arose out of a local government’s eminent-domain-based acquisition of some 47 land parcels in a blighted in-town neighborhood.281 On one parcel stood an adult bookstore, the owners of which for this reason claimed that the taking of this property violated their First Amendment rights.282 A majority of the court rejected this claim by relying on the no-scrutiny rule of Arcara.283 It reasoned in effect that the condemnation law imposed only a doubly incidental burden on speech because it prohibited the continued operation of all businesses in the area, much as the law in Arcara prohibited the continued operation of all businesses that were sites of lewd activity.284 However, two dissenters—following the lead of another dissenter in the lower Commonwealth Court—concluded that the logic of Arcara dictated precisely the opposite result.285 In Arcara, after all, the Court expressed concern that a litigant’s ability to insist on carve-outs from otherwise generally applicable laws would provide a “cloak for obviously unlawful public sexual conduct.”286 In addition, the Court in Arcara analogized the lewd-bookstore operators who had their business shut down to perpetrators of “Fire Code violations,” to someone who is “liable for a civil damages award,” and to “a thief who is sent to prison [only to] complain that his First Amendment right to speak in public places has been infringed.”287 In striking contrast to these situations, according to the dissenters in the Pennsylvania case, the speaker–landowners whose claims were before the Court had not engaged in any “unlawful conduct” at all; rather, they were operating an entirely legal business in what simply happened to be a blighted neighborhood.288

To be sure, a majority of the Pennsylvania Supreme Court saw no good reason to apply the lawful-conduct/unlawful-conduct distinction put forward by the dissenters in the case.289 Indeed, the majority might have gone so far as to condemn the dissenters’ proposed unlawful-conduct exception as internally incoherent. On this view, every generally applicable legal restraint in its nature targets “unlawful activity” precisely because it renders unlawful what was lawful before. For the dissenters, however, it made no sense to view the activity of the unluckily located bookstore in the blighted-neighborhood-condemnation case as wrongful, far less unlawful, in any meaningful sense.290 It remains to be seen whether lower courts—or the Supreme Court—will in time write into law the theory of the dissenters in In re Condemnation by Urban Redevelopment Authority. The case, however, highlights how lower court judges might craft further limits on the Arcara rule.

2.     The Unprotected-Speech Exception

Another proposed limitation on Arcara found a receptive lower-court audience in People v. Sequoia Books.291 That case stemmed from the trial court’s issuance of an injunction that barred a bookstore operator from continuing to do business at its existing site for one year based on its past sale of several obscene publications.292 Pushing aside the government’s reliance on Arcara, the Illinois Supreme Court found that the injunction violated the First Amendment.293 According to that court, Arcara was distinguishable because the prostitution and lewd conduct that triggered governmental intervention in that case “had ‘nothing to do with books or other expressive activity.’”294 In contrast, the case at hand involved a penalty imposed for selling obscene material, which constituted a form of expressive behavior.295 To be sure, obscenity constituted “unprotected” speech under longstanding First Amendment doctrine.296 Even so, the Illinois Supreme Court concluded that it would be intolerable to permit a state to close down any business for engaging in only a few acts of unprotected speech—for example, by shutting down an entire newspaper because a few libelous passages had found their way at some point in the past into its compendious pages.297 By symmetry of logic, according to the court, the state could not close down an entire bookstore because a few obscene books had found their way onto its sprawling shelves.

Sequoia Books held sway until the Supreme Court rejected the case’s principle four years later in Alexander v. United States.298 That case involved an even more severe penalty imposed on a bookstore that sold obscene materials—namely, the permanent forfeiture of the operator’s place of business, together with all of its inventory.299 In a five-to-four ruling, the Court rejected the approach that had carried the day in Sequoia Books, determining instead that Arcara mandated a rejection of the bookstore operator’s First Amendment objection.300 Writing for the majority, Chief Justice Rehnquist did not dispute the proposition that obscene materials “may be ‘expressive’” in some sense.301 For him the key point was that, as with the prostitution and public lewdness involved in Arcara itself, obscenity could be “proscribed consistent with the First Amendment” precisely because it was unprotected speech.302

The Court’s ruling in Alexander, however, was not easily reached. In fact, it generated a stinging dissent, joined by four members of the Court. According to Justice Kennedy, it made a world of difference that the challenged forfeiture in Alexander resulted from a “speech offense,”303 whereas the wrongful conduct in Arcara had “nothing to do with books or other expressive activity.”304 Justice Kennedy bemoaned a principle under which an entire “bookstore or press enterprise could be forfeited as punishment for even a single obscenity conviction,”305 thus producing—as in this very case—the destruction “of hundreds of original titles and thousands of copies, all of which are presumed to be protected speech.”306 In sum, “[i]n a society committed to freedom of thought, inquiry, and discussion,”307 it was “deplorable”308 to make “speakers and the press . . . vulnerable for all of their expression based on some errant expression in the past.”309

The bottom line is that four members of the Court, led by Justice Kennedy, deemed Arcara inapplicable in Alexander based on the distinction between unprotected conduct and unprotected speech. This approach was rejected by five Justices, but only one of them—Justice Thomas—remains on the Court today. Meanwhile, the present-day Justices have moved aggressively to protect free-expression rights in a wide variety of contexts,310 and Justice Kennedy’s opinion in Alexander is marked by an unusually strident tone.311 Against this backdrop, would the current Court adhere to Alexander if the chance to reconsider it arose next week? There is no way to know the answer to that question, but the very fact that the answer is uncertain signals the vulnerability of Arcara to additional qualifications.

C.     Exceptions Based on Developing Constitutional Principles

The Court’s jurisprudence of generally applicable laws does not exist in a vacuum. Rather, it is part of an ever-shifting landscape of free-speech law and constitutional law more generally. Of particular significance, recent constitutional developments raise the question whether further limits on the Arcara no-review rule might emanate from an emerging jurisprudence of so-called “hybrid rights.”

As it turns out, the prospect of recognizing a hybrid-rights exception to Arcara springs from the most famous of all the Court’s generally-applicable-law rulings—namely, its landmark free-exercise ruling in the Smith sacramental-peyote case. There, even as the Court barred the courthouse door to the religiously motivated peyote user, it recognized that some religious practitioners might remain able to invoke the Constitution to skirt the reach of generally applicable laws. This qualification of the principle of Smith emanated from the difficulty the Court encountered in reconciling that decision with earlier rulings—including Murdock v. Pennsylvania312—in which it had exempted religious practitioners from generally applicable laws on what seemed to be free-exercise grounds. In distinguishing those cases, Justice Scalia reasoned that they had not involved “the Free Exercise Clause alone.”313Murdock, for example, concerned “a flat license tax, the payment of which [was] a condition of the exercise of . . . constitutional privileges” asserted by the publisher of a religious magazine.314 That case, he explained, thus presented a hybrid situation in which the right of free exercise operated “in conjunction with . . . freedom of speech” in such a way that the publisher could sidestep the operation of the generally applicable tax law.315

Smith and Murdock signal that courts sometimes will direct elevated scrutiny at generally applicable laws so long as a free-speech claim operates “in conjunction with” a free-exercise claim—and perhaps in conjunction with constitutional claims of other kinds. This principle might best explain the Seventh Circuit’s ruling in Hodgkins ex rel. v. Peterson,316 although the court did not expressly make use of a hybrid-rights rationale. There the court invalidated a juvenile curfew law on First Amendment grounds, finding that the O’Brien standard—rather than the Arcara no-review rule—applied because curfew violations are “intimately related” to later-occurring speech activity.317 Also in the picture, however, were hybrid-rights concerns because the challenged law simultaneously raised issues about the due-process-based liberty of free movement and the right of parents to control the activities of their children.318 Notably, in Smith, the Court indicated that free-exercise claims could operate in tandem not only with free-expression claims, but also with substantive fundamental-rights claims rooted in the Due Process Clauses.319 And if due-process-based claims can thus operate “in conjunction with” free-exercise rights, why not with free-speech rights as well?

To be sure, the subject of hybrid rights remains deeply undertheorized by the Supreme Court and (perhaps for that reason) subject to criticism from many quarters.320 Even so, recent scholarship suggests that hybrid-rights review has taken hold across many areas of constitutional law.321 Of particular importance here, the Court in Smith signaled its openness to using this mode of analysis for the specific purpose of evading a no-review constitutional rule otherwise applicable to generally applicable laws.322 This aspect of that case may thus point the way to judicial recognition of a hybrid-rights-based exception to the kindred no-review rule set forth in Arcara.323

D.     Exceptions Advanced by Academic Commentators

What other exceptions to the no-scrutiny rule of Arcara lurk in the legal materials? Two potential limits have been put forward in scholarly commentary—one focused on countering the risk of purposeful government speech suppression and the other focused on dealing with generally applicable laws that have particularly far-reaching, speech-inhibiting effects.

1.     The Bad-Purpose-Prophylaxis Exception

Srikanth Srinivasan—now a distinguished federal circuit court judge but then still a recent law school graduate—grappled with O’Brien and Arcara in a penetrating paper published in 1995. In that article, Judge Srinivasan endorsed an approach that called for heightened judicial evaluation of generally applicable laws—including in cases otherwise controlled by Arcara—“if there exist special reasons to suspect a speech-restrictive motive” lay behind the government’s action.324 This proposed exception to Arcara differs from the improper-purpose exception325 in an important respect: In the view of Srinivasan, it is the “danger” of the operation of an improper purpose that properly triggers heightened judicial review;326 thus the challenger need not go so far as to prove an actual improper purpose to escape the operation of Arcara.327

The Court has never expressly embraced this approach. According to Srinivasan, however, his view of things provides the best explanation of the Court’s application of intermediate scrutiny in Albertini, rather than Arcara’s no-scrutiny rule, while also accounting for other key rulings, including Cohen v. Cowles Media Co.328 There is much reason to question Judge Srinivasan’s proposed synthesis.329 If he is right, however, yet another exception to the Arcara rule may be waiting in the wings. Indeed, Srinivasan’s analysis suggests that the Court has been assessing—albeit without saying so—whether “special reasons” suggest the presence of bad motives in deciding whether to invoke the rule of O’Brien or Arcara in close cases.330 And if that is true, the Court already is applying, as a practical matter, yet another limit on Arcara’s no-review rule.331

2.     The Substantial-Effects Exception

Another mechanism for tempering the no-review rule of Arcara finds expression in the work of Professor Michael Dorf.332 In a major treatment of all forms of generally applicable rules, Dorf advocated a principle under which heightened scrutiny would attach to all “laws having the incidental effect of substantially burdening fundamental rights to engage in primary conduct,” including primary conduct in the form of speech.333 While Dorf has claimed that this approach is “implicit” in the Court’s cases,334 there is reason to question this assertion, especially in light of the real-world speech burdens tolerated by the Court in Arcara, and even more so in Alexander.335

At the same time, it may be that the nature of the burden imposed on First Amendment freedoms has played a role, and will continue to play a role, in the resolution of borderline Arcara-related cases. Notably, the Court has specifically employed substantial-burden rhetoric in some past free-speech cases,336 and the significance of the burden imposed by the statute at issue in Alexander weighed heavily on the minds of the four dissenters.337 Indeed, Arcara itself offers support for assessing the burden on speech imposed by the challenged law because Chief Justice Burger paused in that case to minimize the impact of the store-closing order. As he put the point: “The severity of this burden is dubious . . . [because] respondents remain free to sell the same materials at another location.”338

The broader point is that Arcara both reflects and has given rise to cross currents in the law. On the one hand, the Court has sought to respond to opening-of-the-floodgates concerns founded on the reality that all forms of generally applicable laws—from tax laws to labor laws to speeding laws—can and do have ripple effects that impede the exercise of free-speech rights.339 On the other hand, both Supreme Court Justices and lower court judges have balked at the idea that they can never consider the speech-suppressive impacts of generally applicable laws, including in doubly-incidental-burden cases, precisely because those impacts can be so severe.340 In sum, the emergence of multiple exceptions to the Arcara no-review rule suggests that neither the floodgates concern nor the severe-effects concern has carried the day in an all-out fashion.341 Against this backdrop, it should come as no surprise if courts, going forward, more openly take account of the speech-impairing consequences of challenged statutes in deciding whether to apply the Arcara rule.342

VI.     The Future of the Court’s Generally Applicable Law Doctrine

What is there to be said about First Amendment limits on generally applicable laws? The foregoing discussion lays the groundwork for six overarching observations.

First and foremost, there is a pressing need in this doctrinal field to exercise care in using legal labels. Courts—including the Supreme Court—tend to deploy terms such as “generally applicable” and “incidental burdens” far too loosely, thus generating confusion in the law. Key steps in addressing this problem involve (1) endorsing the basic tripartite structure presented here and (2) clarifying the nature of the rules (and the nature of the exceptions to those rules) that operate with regard to each key legal category. The Supreme Court must take the lead in this effort. For thoughtful lawyers, however, the inevitably ongoing nature of this process will open up opportunities. This Article shows, for example, that all is not lost simply because a case that involves a generally applicable rule does not trigger strict scrutiny under the principle of Humanitarian Law Project, or even intermediate scrutiny under the principle of O’Brien. Why not? Because it turns out that the Arcara no-scrutiny rule is laced with exceptions that have both a broad scope and a malleable quality.

Understanding the three-part typology outlined here also brings into view the possibility of producing new and surprising results. As we have seen, for example, hybrid-rights analysis may operate in some cases to hold back the operation of the Arcara no-scrutiny doctrine.343 Indeed, a strong judicial embrace of hybrid-rights thinking could even change the results in earlier-decided cases, beginning with Alexander.344 Also in the picture is the question whether some generally applicable laws should fall entirely outside the tripartite doctrinal structure developed here. In another Article, for example, I explain how generally applicable rules of evidence that operate to burden speech may differ in a constitutionally significant way from generally applicable rules of conduct.345 I also have noted that evidence rules bear a kinship to generally applicable laws regarding remedies—all of which might even suggest that the Arcara case itself was wrongly decided, at least if the bookstore closing in that case is viewed as remedial in nature.346 To be sure, there are many complications here. But the overarching point is that such arguments become available only as analysts sharpen their ability to see the nuances that mark the Court’s emerging doctrinal treatment of generally applicable laws.

Second, it bears emphasis that the three-part methodology put forward here provides only a starting point—albeit an important starting point—for First Amendment analysis. Even with that methodology in clear view, courts will have to work hard to sort through a welter of analytical problems.347 Of particular importance, the organizing cases of Humanitarian Law Project, O’Brien, and Arcara all involved what we might call “standard” free-speech issues—that is, issues that arose out of government restrictions on the ability of persons to communicate ideas or information to others. Large questions remain about how the three-part synthesis will map onto other types of free-expression cases, including (for example) cases that involve specialized doctrines with regard to information gathering, expressive association, and access to public forums.348 These complexities highlight a key point: The doctrinal structure put forward here—like any doctrinal structure—is inherently incomplete and inevitably destined to take on shape only as courts process future fact-specific cases. At the same time, the preceding analysis reveals the existence of an overarching scheme that is already in place and that courts must keep in view as they grapple with free-expression disputes. In particular, this scheme should provide the starting point for thinking about how the Free Speech Clause operates in all cases involving generally applicable laws, including those cases that do not involve “standard” forms of expression control.

Third, the most vexing questions that arise in this field center on problems of characterization. As a result, courts must take care not to gloss over subtle challenges raised by characterization choices. As we have seen, for example, Albertini presented a knotty question that the Court never even paused to notice: Was the ban on entering a military base at issue in that case so “intimately related” to the defendant’s subsequent on-base speech activity that it properly triggered intermediate scrutiny under O’Brien?349 Put another way, the case required a decision as to whether the defendant’s entry onto the base was rightly viewed in isolation or as a part of a broader pattern of activity. As it turns out, these sorts of linkage-based characterization problems pervade constitutional law, and in dealing with the First Amendment—just as in any other context—courts would do well not to act as if they did not exist.350 More generally, as we have seen along the way, the resolution of hard characterization questions often will hinge on considerations of underlying constitutional policy—concerns that in this context involve matters such as the risk of improper motivation, the harms created by chilling effects, and the presence of particularly severe or unjustified burdens on speech.351 Because these considerations are sure to carry weight as judges grapple with close-call characterization choices, lawyers would do well to bring them to the surface. No less important, courts that rely on these considerations should openly and frankly explain why they are doing so, thus encouraging other courts to join the effort to avoid the sort of obfuscation that too often has marked this area of law.

Fourth, meta-policies that pervade constitutional decision-making lie at the root of the Court’s treatment of free-speech issues in cases concerning generally applicable laws. Of particular importance is the ubiquitous struggle between rules and standards.352 Most notably, Arcara is a rule-favoring decision, launched largely to advance the cause of legal certainty across a broad swath of real-world conflicts. As Oliver Wendell Holmes famously observed, however, “[t]he life of the law has not been logic: it has been experience.”353 And post-Arcara developments illustrate how doctrines crafted to exempt generally applicable laws from life-complicating judicial scrutiny may turn out to have a fool’s-gold quality. Complexities arise in part because there are many different ways in which courts can and do sidestep the operation of these doctrines. They might, as in Eichman, characterize what looks like a generally applicable law as not generally applicable upon close inspection.354 They might, as in Humanitarian Law Project, determine that a case involves a special kind of generally applicable law, thereby bringing into play strict judicial review.355 They might, as in O’Brien, deal with some types of generally applicable laws by ratcheting the level of judicial scrutiny downward, while nonetheless refusing to abandon judicial scrutiny altogether.356 Or they might—in the pattern that marked even the Court’s free-exercise ruling in Smith—recognize exceptions to a constitutional principle that otherwise forecloses judicial review altogether.357

As we have seen, this last approach has generated a complex mix of actual and potential exceptions to the no-review standard of Arcara.358 It would not be accurate to say that these exceptions now reach so far that they have swallowed the Arcara rule. They do, however, bring to this field a measure of complexity that undermines any claim that the Court’s standoffish approach to doubly-incidental-burden cases has sown only certainty in its wake. In the end, the inevitable challenges of drawing a line between O’Brien-type and Arcara-type cases,359 the odd results that drawing that line can create,360 the emergence of a checkerboard of exceptions to the no-review principle that usually controls the latter set of cases,361 the plasticity that marks those exceptions,362 and the possibility that additional exceptions will take hold in the future363 conspire to invite the conclusion that the hoped-for benefits of Arcara’s supposedly bright-line rule have not been realized in practice. In these circumstances, it is fair to ask whether that rule should endure.364 It is fairer still to question the wisdom of expanding judicial use of Arcara’s hard-edged methodology, particularly by way of abandoning the Court’s intermediate-scrutiny approach to O’Brien-type incidental-burden cases.365 To be sure, there is room in the law for establishing and adhering to certainty-fostering doctrines. The play-out of Arcara illustrates, however, the difficulty of maintaining the integrity of doctrines designed to immunize generally applicable laws from any form of constitutional attack when those laws clash in powerful ways with core constitutional values.366

Fifth, the free-speech doctrines that have taken hold in this area—and particularly the no-review rule of Arcara—may prove to be vulnerable to change in light of the shifting membership of the Supreme Court. The departure in the post-Arcara period of three Justices from the Court seems especially significant. Chief Justice Rehnquist, who joined Arcara and penned the majority opinion in Alexander, earned a reputation over many years as a largely “statist” thinker, including with respect to free-speech issues.367 Justice Scalia was the architect of Smith, a staunch patron of Arcara, and the great champion within the Court of embracing a rules-over-standards jurisprudence.368 Finally, Justice Stevens, though often viewed as a First Amendment “liberal,” led the charge for establishing the no-exemption approach of Smith369 and endorsed the bright-line rule applied in both Arcara and Alexander.370 In contrast, none of the Court’s four now-sitting “liberals” have signed on to Justice Stevens’s hands-off approach to cases that involve generally applicable laws. And Justice Kennedy—as we have seen—passionately opposed application of the Arcara rule in Alexander.371 Perhaps most important of all, members of the so-called “conservative” wing of the current Court have signaled concerns about endorsing a no-review approach to generally applicable laws. Indeed, there are signals in the case law that this branch of the Court may look with special concern, rather than unflinching reserve, at generally applicable anti-discrimination laws insofar as they burden particularly important expressive-association rights.372 What all of this means for the future is unclear. One thing, however, is clear enough: A reading of the tea leaves gives reason to suspect that the now-sitting Justices may hesitate to keep traveling down the same path marked by their predecessors for dealing with generally applicable laws.

The final question looms: What is there to say about the three-part scheme laid out here in light of overarching First Amendment theory? Drawing on different conceptions of that Amendment’s purpose, distinguished commentators—one of whom now sits on the Supreme Court373—have reflected fruitfully on the proper treatment of generally applicable laws.374 Much of this work reflects a difference in philosophical outlook. Some analysts see the Free Speech Clause as centered on rooting out restrictions that stem from, or facilitate actions based on, speech-suppressive motivations. In their view, the chief aim of the clause is to facilitate judicial watchdogging of self-interested and majority-minded political officials,375 thus properly steering doctrine in the direction of ensuring that those officials do not purposefully act to rein in offbeat and state-critical expression.376 Other analysts focus less on speech-suppressive motivations and more on real-world effects. They urge that such First Amendment goals as facilitating the wide-open search for truth and accommodating valued exercises of personal self-expression require a practical focus on speech-suppressive impacts, regardless of the state-actor motivations that brought those impacts about.377

Some commentators who favor a purpose-centered approach have faulted the Court’s treatment of generally applicable laws, particularly under the O’Brien test, which triggers a measure of free-speech review even in the seeming absence of wrongful motivations.378 This Article may provide at least a partial response to this line of criticism by highlighting that O’Brien does not stand alone. Instead, (1) it operates to trigger only a limited form of scrutiny even when a law’s as-applied, speech-inhibiting impact is profound,379 and
(2) it constitutes only one part of a broader three-part framework, which includes a treatment of doubly incidental burdens that precludes Free Speech Clause review in a broad array of cases that involve generally applicable laws.380 Of no less importance, even analysts who take a purpose-centered view of things recognize that O’Brien-based intermediate scrutiny may help to “smoke out” wrongful, but hidden, legislative motives.381 Of importance, too, is the fact that there is no inconsistency in paying attention to problematic purposes and paying attention to problematic effects at the same time in assessing free-speech problems. As it turns out, the Court—neither surprisingly nor unwisely—has taken account of both the purpose-centered and effect-centered theories in crafting its own overarching vision of how the First Amendment should work.382

* * *

Against this backdrop, two questions move to center stage: Does it make sense for courts to scrutinize laws that impose direct-in-effect burdens on speech more aggressively than laws that impose incidental burdens on speech? And does it also make sense for courts to look at laws that impose incidental burdens on speech more carefully than laws that impose doubly incidental burdens?

The answer to the first question is clearly “yes” for two main reasons. First, incidental burdens on speech by definition have nothing to do with regulating speech because of its content or speech-related impact. The law at issue in O’Brien, for example, operated against draft-card destroyers without one iota of regard to how audience members were being influenced or affected by the card-burner’s expression. In contrast, concerns about speech-related effects lie at the heart of challenges to generally applicable laws that impose direct-in-effect burdens on speech for the most straightforward of reasons: When generally applicable laws impose direct-in-effect speech burdens, they do so precisely because listener reactions to the particular expressions at issue have triggered the law’s operation. For this reason, concerns centered on regulating speech based on content and speech-specific effects—which constitute core drivers of First Amendment law383—operate in the two contexts in dramatically different ways.384

Second, laws that impose incidental burdens on speech in their very nature foreclose only one particular mode of communicating a message, whereas laws that place direct-in-effect burden on speech can and often do foreclose communication of a disfavored message altogether. In O’Brien, for example, the protester remained free to express his message in many ways, including not only by speaking or writing but also (for example) by burning an identical facsimile of his draft card. In telling contrast, when the government enacts an across-the-board prohibition on obstructing the draft, vigorous anti-draft communications become subject to government control in all their forms.385 Put simply, such a restriction renders unlawful otherwise protected speech against the draft, whatever form it takes, no less than it renders unlawful (for example) all forms of obstructing entrance to a draft-board office. The core distinction in First Amendment law between forbidding dissemination of certain messages, as opposed to forbidding only certain means of disseminating those messages, thus strongly reinforces the Court’s choice to apply the strictest form of review in direct-in-effect-burden cases, while ratcheting downward the level of review in incidental-burden cases such as O’Brien.386

So, what of the Court’s drawing of a distinction between those laws that impose incidental burdens on speech as opposed to those laws that impose doubly incidental burdens? The argument in favor of this distinction begins with a simple idea: At least as a rule, to the extent that a proposed generally applicable restriction will “hit” speech itself, a greater likelihood should exist that the speech-inhibiting effect of the restriction will be visible to lawmakers and thus, at least potentially, play a problematic role in driving the restriction’s enactment. O’Brien is the poster-child for this line of thinking. There, after all, federal legislators well knew when they enacted the challenged law that many (if not most) then-ongoing acts of draft-card destruction were coming from war-protesting critics of government policy. With this speech-related linkage in clear view, it became a possibility (indeed, a very distinct possibility) that Congress would pass the law (at least in part) to hamper and punish this especially provocative form of government-critiquing expression.

Even more important than this point, however, is a more basic matter: Cases such as O’Brien arise because the government has compelled the speaker to make a state-favored, speech-specific choice. In other words, for speakers who desired to express opposition to the draft, the law at issue in O’Brien forced them to use a mode of speech that they did not want to use precisely because they viewed draft-card burning as the most effective and/or most meaningful means of personal self-expression. In telling contrast, rules that impose doubly incidental burdens on speech do not compel any speaker to choose one mode of expression over another (or, for that matter, to make any speech-related choice at all) for the simple reason that, by definition, they do not prohibit speech-related activity. Furthermore, doubly-incidental-burden-imposing laws always have something of a fairness-based, sleep-in-the-bed-you-made quality. The late-running reporter could have left for work earlier in the day; the frustrated taxpayer-publisher did not have to pursue a profit-making business (with all the ordinary burdens that running such a business brings); and the defendants in Arcara did not have to permit prostitution and sex acts to occur on their business premises. Each of those cases thus involves what might be called a consequential, down-the-line, or (if you will) doubly incidental burden on speech. Such a burden in its nature impinges on speech in a less direct way than does imposing a punishment—as necessarily occurs in all incidental-burden cases—for engaging in expression itself.

To say these things is not to say that courts should reject out of hand all challenges to laws that impose doubly incidental burdens on expression. (Indeed, as we have seen, the Court has eschewed this approach by recognizing many exceptions to the no-review rule laid down in Arcara.) It is to say instead that placing incidental-burden cases and doubly-incidental-burden cases into different analytical categories comports with a common-sense idea. The idea is that, within a legal regime focused on protecting “freedom of speech,” there is a built-in logic in distinguishing between laws that operate against primary conduct that is itself speech and laws that target primary conduct that does not involve speech at all.387 In short, the Court—albeit without ever explaining why—has walked on sturdy ground in moving toward a calibrated system of free-speech review that takes account of whether generally applicable laws impose direct-in-effect, incidental, or doubly incidental burdens on speech.

 VII.     Conclusion

In a famous song of the 1960s, John Sebastian of The Lovin’ Spoonful reflected in these terms on a mental process often experienced by thoughtful legal analysts: “But the more I see, the more I see there is to see.”388 There is a pull in working with the law to reduce its operation to simple-sounding verbal dictates. This tendency is understandable because the goals of any sound legal system include fostering predictability and efficiency and ensuring the equal treatment of individuals who find their way into the courts. At least in theory, hard-edged and uncomplicated doctrinal formulations operate to advance these values.

One doctrinal formulation that has attracted a receptive audience declares without more that “generally applicable laws” should be exempt from constitutional attack.389 As this Article shows, any effort to describe current free-speech law in these terms is not only mistaken, but wildly off the mark. In essence, the Court has resisted so simplistic a sorting mechanism because life is complex and free-speech values rooted firmly in the constitutional text are sufficiently weighty that they require more protection than this dismissive approach would afford.

The Court instead has erected a doctrinal structure that in fact has structure because it divides generally applicable laws into three distinct categories for purposes of Free Speech Clause analysis. This tripartite framework can and does produce hard questions. Courts will often have to wrestle with whether a generally applicable law burdens the free-speech right in a direct-in-effect, incidental, or doubly incidental way. When dealing with laws that impose direct-in-effect burdens, courts will have to apply strict means-end analysis in a serious way—sometimes such a serious way that (as Humanitarian Law Project itself shows) the challenged law will be upheld. With regard to laws that impose incidental burdens, courts will have to decide whether O’Brien’s intermediate-scrutiny test is operative and whether it is satisfied if it does apply. Even when a statute imposes only a doubly incidental burden on speech—thus bringing the Arcara no-review rule onto the scene—the court will have to determine whether any of the already-recognized exceptions to that rule are applicable or whether a new exception should take hold.

Some observers may find themselves unsettled by the need to address these sorts of often-difficult questions. But categorization choices, exceptions to generally operative First Amendment doctrines, and case-specific applications of strict and intermediate scrutiny are all common features of free-speech law. As a result, the Court’s incorporation of these features into the body of constitutional rules that govern cases concerning generally applicable laws is in line with themes and practices that mark—indeed, routinely mark—judicial work with the Free Speech Clause.

All of this helps to underscore the central message of this Article. At bottom, the analysis offered here shows that the existing doctrinal landscape provides thoughtful lawyers with rich opportunities to craft innovative arguments in advancing free-speech challenges to generally applicable laws, and in defending those laws as well. At the same time, those opportunities will be available as a practical matter to those lawyers only if they come to grasp the rhetoric around which the Court’s three-part doctrinal structure is built, the different styles of scrutiny that different categories of generally applicable laws trigger, the proper scope of the principles established by key Supreme Court precedents, and the push that underlying free-expression-related policies can exert as courts make their way through particular disputes. These many opportunities offer a reminder of a deeper point too—that, as Justice Cardozo once observed, the work of those who toil in the law “in its highest reaches is not discovery, but creation.”390


  1. [1]. See Erwin Chemerinsky, Constitutional Law: Principles and Policies §§ 11.1–11.4.3 (5th ed. 2011).

  2. [2]. Michael C. Dorf, Incidental Burdens on Fundamental Rights, 109 Harv. L. Rev. 1175,
    1176–78 (1996). An expansive literature touches on the sweeping variety of issues raised by such laws. See, e.g., Laurence H. Tribe, American Constitutional Law §§ 12-2 to 12-7, at 789–832,
    § 12-23, at 977–86 (2d ed. 1988); Larry A. Alexander, Trouble on Track Two: Incidental Regulations of Speech and Free Speech Theory, 44 Hastings L.J. 921 (1990); Ashutosh Bhagwat, Producing Speech, 56 Wm. & Mary L. Rev. 1029 (2015); David Bogen, Generally Applicable Laws and the First Amendment, 26 Sw. L. Rev. 201 (1997); Wesley J. Campbell, Speech-Facilitating Conduct, 68 Stan. L. Rev. 1 (2016); Joshua P. Davis & Joshua D. Rosenberg, The Inherent Structure of Free Speech Law, 19 Wm. & Mary Bill Rts. J. 131 (2010); David S. Day, The Incidental Regulation of Free Speech, 42 U. Miami L. Rev. 491 (1988); Daniel A. Farber, Content Regulation and the First Amendment: A Revisionist View, 68 Geo. L.J. 727 (1980); Alan E. Garfield, The Mischief of Cohen v. Cowles Media Co., 35 Ga. L. Rev. 1087 (2001); Raleigh E. Hansman, Doctrinal Development or Devolution?: An Examination of the Incidental Regulation Test from Texas v. Johnson Through Holder v. Humanitarian Law Project, 57 San Diego L. Rev. 122 (2012); Elena Kagan, Private Speech, Public Purpose: The Role of Governmental Motive in First Amendment Doctrine, 63 U. Chi. L. Rev. 413 (1996); Wendy K. Olin, Constitutional Survival Camp: What are the Chances that the General Applicability Test Will Make It?, 68 S. Cal. L. Rev. 1029 (1995); Robert Post, Recuperating First Amendment Doctrine, 47 Stan. L. Rev. 1249 (1995); Jed Rubenfeld, The First Amendments Purpose, 53 Stan. L. Rev. 767 (2001); Frederick Schauer, Categories and the First Amendment: A Play in Three Acts, 34 Vand. L. Rev. 265 (1981) [hereinafter Schauer, Categories]; Frederick Schauer, Cuban Cigars, Cuban Books, and the Problem of Incidental Restrictions on Communications, 26 Wm. & Mary L. Rev. 779 (1985) [hereinafter Schauer, Cuban Cigars]; Jeffrey M. Shaman, Rules of General Applicability, 10 First Amend. L. Rev. 419 (2012); Srikanth Srinivasan, Incidental Restrictions of Speech and the First Amendment: A Motive-Based Rationalization of the Supreme Courts Jurisprudence, 12 Const. Comment. 401 (1995); Geoffrey R. Stone, Content-Neutral Restrictions, 54 U. Chi. L. Rev. 46 (1987) [hereinafter Stone, Content-Neutral Restrictions]; Geoffrey R. Stone, Free Speech in the Twenty-First Century: Ten Lessons from the Twentieth Century, 36 Pepp. L. Rev. 273 (2009) [hereinafter Stone, Free Speech]; Eugene Volokh, Freedom of Speech and Workplace Harassment, 39 UCLA L. Rev. 1791 (1992) [hereinafter Volokh, Workplace Harassment]; Eugene Volokh, Speech as Conduct: Generally Applicable Laws, Illegal Courses of Conduct, “Situation-Altering Utterances,” and the Uncharted Zones, 90 Cornell L. Rev. 1277 (2005) [hereinafter Volokh, Speech as Conduct]; Keith Werhan, The O’Briening of Free Speech Methodology, 19 Ariz. St. L.J. 635 (1987); Susan H. Williams, Content Discrimination and the First Amendment, 139 U. Pa. L. Rev. 615 (1991); Note, Neutral Rules of General Applicability: Incidental Burdens on Religion, Speech, and Property, 115 Harv. L. Rev. 1713 (2002) [hereinafter Neutral Rules of General Applicability].

  3. [3]. See Garfield, supra note 2, at 1096; Volokh, Speech as Conduct, supra note 2, at 1294; see also Rodney A. Smolla, Law of Defamation § 11:11 (2d ed. 1999) (noting that the defamation tort is “different in kind” for this reason).

  4. [4]. See Kagan, supra note 2, at 492.

  5. [5]. See infra note 11 and accompanying text.

  6. [6]. See, e.g., Day, supra note 2, at 492 (describing the “incidental regulation standard” as “one of the three pillars of the modern free speech doctrine”); Dorf, supra note 2, at 1176–77 (explaining that direct and incidental restraints give rise to the two principal ways that “[f]undamental constitutional rights are burdened”); Kagan, supra note 2, at 491 (characterizing “the distinction between direct and incidental restrictions on speech” as “a distinction as important as any in First Amendment law”).

  7. [7]. Stone, Content-Neutral Restrictions, supra note 2, at 105 See Alexander, supra note 2, at 931 (arguing that “all laws have information effects” and thus are potentially susceptible to First Amendment review); Schauer, Cuban Cigars, supra note 2, at 784 (highlighting the “enormous range of government decisions” and the “host of government actions that” incidentally affect speech); id. at 779 (providing examples); see also Stone, Free Speech, supra note 2, at 290 (noting that generally applicable laws “come in many shapes and sizes”). This body of law has also taken on greater importance “in light of the pervasiveness of government regulation in the modern state.” Dorf, supra note 2, at 1201.

  8. [8]. Shaman, supra note 2, at 423.

  9. [9]. Day, supra note 2, at 495 n.15.

  10. [10]. Srinivasan, supra note 2, at 403.

  11. [11]. See, e.g., Dorf, supra note 2, at 1200 (“No unified constitutional doctrine of incidental burdens currently exists.”); id. at 1251 (noting that the Court’s work in this area has occurred “haltingly and sometimes unwittingly”); Garfield, supra note 2, at 1105–06 (noting different meanings of the term “incidental”); Olin, supra note 2, at 1030 (“The Court ... has not clearly articulated a definition of a generally applicable law ....”); id. at 1039 (expanding on this point by highlighting difficulties of analysis and prediction that result); Shaman, supra note 2, at 423 (“[T]he Supreme Court’s treatment of rules of general applicability has been anything but consistent.... [It] has wavered from one extreme to the other ....”); Srinivasan, supra note 2, at 402–03 (noting “doctrinal confusion” associated with “‘generally applicable’ laws ... that in some applications ‘incidentally’ restrict speech” and that “the Supreme Court has paid ... little attention” to the significance of these terms); Volokh, Speech as Conduct, supra note 2, at 1294 (noting different uses of the term “generally applicable law”). The confusion sown by the Court’s own struggles is well-evidenced by the inconsistent use of relevant terminology in the scholarly commentary. See, e.g., Kagan, supra note 2, at 492 (treating together as “general law[s],” which involve an “incidental restraint,” tax laws and labor laws that burden publishers, vandalism laws as applied to “draw[ing] swastikas on a synagogue wall” and the statute at issue in OBrien); Schauer, Cuban Cigars, supra note 2, at 779–82 (labeling as “generally applicable government regulations” with “incidental effect[s]” both antitrust and labor laws subjected to strict scrutiny and the sleeping-in-the-park regulation that triggered only intermediate scrutiny); Shaman, supra note 2, at 438–39 (defending the strict-scrutiny approach to “generally applicable law” at issue in Humanitarian Law Project because it “was being applied to regulate expressive conduct,” but not indicating why OBrien—to which the same description applied—resulted in only intermediate scrutiny); Williams, supra note 2, at 722–23 (arguing that the “incidental” label should not apply to cases in which the Court has applied it). Compare, e.g., Werhan, supra note 2, at 650–52 (asserting that a leafletting case involved “a direct restriction on expression”), with Members of the City Council of L.A. v. Taxpayers for Vincent, 466 U.S. 789, 808 (1984) (characterizing a sign-posting ban as involving only an “incidental restriction on expression”). On occasion, frustration with the Court’s work in this area surfaces within the Court itself. See Day, supra note 2, at 511 (noting Justice Stewart’s objection to the majority’s application of the “incidental regulation test” to a military rule that involved “purposeful” suppression of expressive petition-circulating activities).

  12. [12]. See infra Part II.

  13. [13]. See infra Part II.

  14. [14]. Emp’t Div. v. Smith, 494 U.S. 872 (1990).

  15. [15]. Id. at 878–79.

  16. [16]. Holder v. Humanitarian Law Project, 561 U.S. 1 (2010).

  17. [17]. United States v. O’Brien, 391 U.S. 367 (1968).

  18. [18]. Arcara v. Cloud Books, Inc., 478 U.S. 697 (1986).

  19. [19]. Humanitarian Law Project, 561 U.S. at 8–9 (referring to the prohibitions described in 18 U.S.C. § 2339B (2006) and relevant definitions contained in 18 U.S.C. § 2339A (2006)).

  20. [20]. Id. at 4–5, 10–11.

  21. [21]. Id. at 26–29.

  22. [22]. Id. at 27.

  23. [23]. Id. at 26–27.

  24. [24]. United States v. O’Brien, 391 U.S. 367, 369–70 (1968).

  25. [25]. See id. at 376.

  26. [26]. Arcara v. Cloud Books, Inc., 478 U.S. 697 (1986).

  27. [27]. Id. at 698–702.

  28. [28]. Id.

  29. [29]. Id. at 700–02.

  30. [30]. Id. at 705.

  31. [31]. See id. at 705–06.

  32. [32]. Id. at 706.

  33. [33]. Holder v. Humanitarian Law Project, 561 U.S. 1, 28 (2010).

  34. [34]. United States v. O’Brien, 391 U.S. 367, 382 (1968).

  35. [35]. Arcara, 478 U.S. at 706–07.

  36. [36]. Cohen v. Cowles Media Co., 501 U.S. 663 (1991).

  37. [37]. Hustler Magazine, Inc. v. Falwell, 485 U.S. 46 (1988).

  38. [38]. Volokh, Speech as Conduct, supra note 2, at 1286–87.

  39. [39]. See, e.g., supra note 11 and accompanying text.

  40. [40]. See Lovell v. City of Griffin, 303 U.S. 444, 452 (1938) (“Liberty of circulating is as essential ... as liberty of publishing ....” (quoting Ex parte Jackson, 96 U.S. 727, 733 (1878))).

  41. [41]. See supra notes 24–25 and accompanying text.

  42. [42]. United States v. O’Brien, 391 U.S. 367, 377 (1968).

  43. [43]. See infra notes 217–28 and accompanying text.

  44. [44]. See infra notes 217–28 and accompanying text.

  45. [45]. See, e.g., Hughes v. Oklahoma, 441 U.S. 322, 336 (1979) (noting that assessing whether a statute has merely “incidental effects on interstate commerce” requires evaluation both of the statute “on its face” and its “practical effect”).

  46. [46]. Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 894 (1992) (plurality opinion).

  47. [47]. Texas v. Johnson, 491 U.S. 397, 399 (1989).

  48. [48]. Id. at 400 n.1.

  49. [49]. Id. at 407.

  50. [50]. Flag Protection Act of 1989, Pub. L. 101–131, 103 Stat. 777 (as amended in 18 U.S.C. § 700).

  51. [51]. See Kagan, supra note 2, at 492 (describing the law in Eichman as “generally applicable”); Geoffrey R. Stone, Flag Burning and the Constitution, 75 Iowa L. Rev. 111, 119 (1989) (defending the law on this ground).

  52. [52]. United States v. Eichman, 496 U.S. 310, 314 (1990) (citing United States v. O’Brien, 391 U.S. 367, 377 (1968)). See also Day, supra note 2, at 499 (“[T]he adverse impact for [OBrien-type] regulations is considered incidental because it is nonpurposeful.”).

  53. [53]. A specialized set of limits, including with regard to generally applicable laws, applies to government restrictions that limit speech in so-called public forums. See Post, supra note 2, at 1260–61 (critiquing the Court’s conflation of review for time, place, and manner restrictions, including in regard to use of public forums, and statutes that, like the one in OBrien, do not focus on speech). Cases that involve such limits raise specialized concerns and therefore do not receive focused attention here.

  54. [54]. Cohen v. Cowles Media Co., 501 U.S. 663 (1991).

  55. [55]. Id. at 666.

  56. [56]. Id. at 665–66.

  57. [57]. Id. at 665.

  58. [58]. Id. at 665–66.

  59. [59]. Id. at 666.

  60. [60]. Id. at 667 (citing Cohen v. Cowles Media Co., 457 N.W.2d 199, 205 (Minn. 1990)).

  61. [61]. Id. at 671 (distinguishing Cohen from Florida Star v. B.J.F., 491 U.S. 524 (1989)).

  62. [62]. Id. at 669.

  63. [63]. Id. at 669–70 (citations omitted).

  64. [64]. Id. at 670.

  65. [65]. Id. at 671.

  66. [66]. Id. at 669.

  67. [67]. Id. at 677 (Souter, J., dissenting) (quoting Emp’t. Div. v. Smith, 494 U.S. 872, 901 (1990) (O’Connor, J., concurring)).

  68. [68]. Id. at 676–77.

  69. [69]. Id. at 677.

  70. [70]. Id. at 678.

  71. [71]. Id. at 679.

  72. [72]. Id. at 674 (Blackmun, J., dissenting).

  73. [73]. Hustler Magazine, Inc. v. Falwell, 485 U.S. 46 (1988).

  74. [74]. Cohen, 501 U.S. at 674 (Blackmun, J., dissenting).

  75. [75]. Hustler Magazine, 485 U.S. at 48.

  76. [76]. Cohen, 501 U.S. at 674–75 (Blackmun, J., dissenting) (citing Hustler Magazine, 485 U.S. at 50 n.3).

  77. [77]. Id. at 675.

  78. [78]. Id. at 675–76.

  79. [79]. Id. at 671 (majority opinion).

  80. [80]. Id. at 675 (Blackmun, J., dissenting). See also Garfield, supra note 2, at 1093 (summarizing the majority’s decision in Cohen v. Cowles Media Co.).

  81. [81]. Cohen, 501 U.S. at 674–75 (Blackmun, J., dissenting) (citing Hustler Magazine, 485 U.S. at 50 n.3).

  82. [82]. Id. at 675 n.3

  83. [83]. Id.

  84. [84]. See Volokh, Speech as Conduct, supra note 2, at 1295–96 (detailing this point).

  85. [85]. See id.

  86. [86]. Cohen, 501 U.S. at 670.

  87. [87]. Id.

  88. [88]. See Garfield, supra note 2, at 1123 (decrying Justice White’s suggested distinction between economic and reputational damages as “specious”).

  89. [89]. Cf. id. at 1088 (condemning the majority’s “brusque treatment” of the defendants’ First Amendment argument).

  90. [90]. Another problem is that this reading of the case clashes with well-settled Supreme Court precedent. That precedent, after all, has never even hinted that a monetary-injury exception limits the long-categorically-stated principle “that liability for the defamation of a public official or figure requires proof that defamatory statements were made with knowledge or reckless disregard of their falsity.” United States v. Alvarez, 132 S. Ct. 2537, 2563 (2012). The Court has stated no less categorically that “[t]his same requirement applies when public officials and figures seek to recover for the tort of intentional infliction of emotional distress.” Id. (emphasis added). See also Church of Scientology Int’l v. Behar, 238 F.3d 168, 177 n.3 (2d Cir. 2001) (observing that defamation plaintiff’s “status as a public figure means that it was required to demonstrate actual malice, whatever remedy it sought”).

  91. [91]. See United States v. O’Brien, 391 U.S. 367, 377 (1968) (using “substantial governmental interest” test that the Court later characterized of entailing intermediate scrutiny).

  92. [92]. Cohen, 501 U.S. at 677 (Souter, J., dissenting).

  93. [93]. See Bogen, supra note 2, at 204 (viewing Cohen v. Cowles Media Co. as holding “that the incidental effects of a generally applicable law do not violate the First Amendment” in derogation of OBrien); Garfield, supra note 2, at 1106–07 (seeing a “conundrum” in the Court’s failure to “mention OBrien or use the test announced in that case”); Srinivasan, supra note 2, at 419 (indicating that Cohen v. Cowles Media Co. “seems to violate” the overarching principle of OBrien).

  94. [94]. See Cohen, 501 U.S. at 675 (Blackmun, J., dissenting).

  95. [95]. Brandenburg v. Ohio, 395 U.S. 444, 456 (1969) (Douglas, J., concurring).

  96. [96]. Broadrick v. Oklahoma, 413 U.S. 601, 615 (1973) (emphasis added); see also Texas v. Johnson, 491 U.S. 397, 406 (1989) (“The government generally has a freer hand in restricting expressive conduct than it has in restricting the written or spoken word.”). See generally Bogen, supra note 2, at 229 (“One would think the Court would be embarrassed to give nude dancing more scrutiny than it gave the press, yet it did so [with its generally-applicable-law rulings in Barnes and Cohen v. Cowles Media Co.].”).

  97. [97]. Cohen, 501 U.S. at 675 (Blackmun, J., dissenting).

  98. [98]. See Bogen, supra note 2, at 227 (asserting that liability in Cohen v. Cowles Media Co. hinged on the “content of the speech”); Garfield, supra note 2, at 1096 (noting Justice White’s “failure” to consider this point); William E. Lee, The Unusual Suspects: Journalists as Thieves, 8 Wm. & Mary B. Rts. J. 53, 132 (1999) (“[T]he communicative impact of the publication was the heart of the plaintiff’s damage claim.”).

  99. [99]. Cohen, 501 U.S. at 674–75 (Blackmun, J., dissenting) (citing Hustler Magazine, Inc. v. Falwell, 485 U.S. 46, 50 n.3 (1988)).

  100. [100]. See N.Y. Times Co. v. Sullivan, 376 U.S. 254, 279–80 (1964) (setting forth actual malice test).

  101. [101]. See Veilleux v. Nat’l Broad. Co., 206 F.3d 92, 127 (1st Cir. 2000) (reconciling Cohen v. Cowles Media Co. with Hustler on this ground); Garfield, supra note 2, at 1121 (describing Hustler as impeding the “end-running” of New York Times).

  102. [102]. See supra notes 3–5 and accompanying text.

  103. [103]. See Boos v. Barry, 485 U.S. 312, 321 (1988) (indicating that a focus on “listeners’ reactions” correlates with content discrimination).

  104. [104]. Restatement (Second) of Torts § 46 cmt. d, illus. 3 (Am. Law Inst. 1965).

  105. [105]. See Texas v. Johnson, 491 U.S. 397, 404 (1989) (holding that, for conduct to qualify as speech, there must be “[a]n intent to convey a particularized message” and a “likelihood [that is] great that the message would be understood by those who viewed it”) (citation omitted).

  106. [106]. Assume, for example, that an individual secretly and maliciously causes the death of a neighbor’s beloved dog. Obviously, the heartless dog killer is not subject to a sanction on the ground that the individual has “spoken” to the dog owner; indeed, the dog owner’s suffering occurs when the dog dies, an event that may well occur long before the owner even learns of the brutal neighbor’s involvement.

  107. [107]. See Volokh, Workplace Harassment, supra note 2, at 1830.

  108. [108]. See generally Volokh, Speech as Conduct, supra note 2.

  109. [109]. See id. at 1287.

  110. [110]. Police Dep’t v. Mosley, 408 U.S. 92, 101–02 (1972).

  111. [111]. Volokh, Speech as Conduct, supra note 2, at 1286–87.

  112. [112]. Id. Others (albeit with less elaboration) have endorsed approaches similar to Professor Volokh’s. See Tribe, supra note 2, at 848 n.56 (“[H]owever a law is written, it may not constitutionally be applied to punish speech on content-related grounds ....”); Post, supra note 2, at 1257–58 (noting that the breach-of-the-peace statute in Cohen v. California “appeared on its face to be a proper and legitimate regulation of conduct” but that the Court focused on the statute’s “distinct and separate application to ... the language of political discourse”). Indeed, this approach can be seen as having its origins in the Court’s earliest encounters with the Free Speech Clause. See Volokh, Speech as Conduct, supra note 2, at 1287–88 (discussing the Schenck and Debs cases, which involved “a generally applicable provision of the Espionage Act”).

  113. [113]. See Boos v. Barry, 485 U.S. 312, 321 (1988).

  114. [114]. See Volokh, Speech as Conduct, supra note 2, at 1291 (discussing Hustler in this same way).

  115. [115]. See generally United States v. O’Brien, 391 U.S. 367 (1968).

  116. [116]. Id. As a matter of legal labeling, one might deal with laws that impose direct-in-effect burdens on speech by declaring them, for that reason, not to be “generally applicable” in the first place. Because these restrictions do not single out speakers, however, others have taken a different approach. See, e.g., Shaman, supra note 2, at 438 (describing Humanitarian Law Project as involving “a law of general applicability”); Volokh, Speech as Conduct, supra note 2, at 1288–94 (describing content-based-as-applied laws as “generally applicable”). The more important point is that, any debate about which of these labels to apply is a tempest in a teapot. The key point is that the Court should employ consistently whatever terminology it opts to use in light of the functional distinctions that separate laws that place (1) direct-in-effect; (2) incidental; and
    (3) doubly incidental burdens on speech.

  117. [117]. Volokh, Speech as Conduct, supra note 2, at 1336, 1347.

  118. [118]. See id. at 1284, 1347.

  119. [119]. See id. at 1287, 1348. To be sure, the government has broader freedom to regulate the content of speech in certain specialized settings, such as in government-run schools and workplaces. The problems presented by specialized speech regulation in these settings are beyond the scope of this Article.

  120. [120]. Cohen v. California, 403 U.S. 15, 20 (1971).

  121. [121]. Volokh, Speech as Conduct, supra note 2, at 1287, 1305.

  122. [122]. See generally United States v. O’Brien, 391 U.S. 367 (1968).

  123. [123]. See infra notes 145–60 and accompanying text.

  124. [124]. Some observers, for example, have suggested that the direct-in-effect component of the three-part structure advanced here is inconsistent with the Court’s treatment of speech-based sexual harassment in R.A.V. v. City of St. Paul, 505 U.S. 377 (1992). In fact, however, R.A.V. by its terms addressed only sexual harassment in the form of unprotected fighting words. See Volokh, Speech as Conduct, supra note 2, at 1293 n.74; Volokh, Workplace Harassment, supra note 2, at 1829–32.

  125. [125]. See supra notes 11–13 and accompanying text.

  126. [126]. Holder v. Humanitarian Law Project, 561 U.S. 1 (2010).

  127. [127]. Id. at 7–8.

  128. [128]. Id. at 27.

  129. [129]. Id. at 26–27.

  130. [130]. Id. at 28. The unsettled character of the law in this area is illustrated by the fact that lower courts had agreed with the government’s OBrien-based argument. See, e.g., Humanitarian Law Project v. Reno, 205 F.3d 1130, 1135 (9th Cir. 2000) (“[T]he material support restriction here does not warrant strict scrutiny because it is not aimed at interfering with the expressive component of their conduct but at stopping aid to terrorist groups. Compare OBrien ... with Johnson ....”).

  131. [131]. Humanitarian Law Project, 561 U.S. at 27–28 (alteration in original) (citations omitted) (quoting Texas v. Johnson, 491 U.S. 397, 401 (1989)).

  132. [132]. David Cole, The First Amendments Borders: The Place of Holder v. Humanitarian Law Project in First Amendment Doctrine, 6 Harv. L. & Pol’y Rev. 147, 152 (2012). See Cox Broad. Corp. v. Cohn, 420 U.S. 469, 495 (1975) (noting that a newspaper report involved “pure expression” in overturning application of a rape shield law).

  133. [133]. Notably, the challengers of the material-support statute in Humanitarian Law Project themselves relied on this distinction in arguing their case to the Supreme Court. See Brief for Petitioners at 28, Holder v. Humanitarian Law Project, 561 U.S. 1 (2010) (Nos. 08-1498 & 09-89), 2009 WL 5177142, at *28 (“Courts are somewhat more skeptical of government regulation of ‘pure speech’ than of expressive conduct ....” (citing Broadrick v. Oklahoma, 413 U.S. 601, 615 (1973))).

  134. [134]. See, e.g., Kagan, supra note 2, at 492 n.212 (emphasizing the lack of difference between verbal and nonverbal expression beyond the obvious point that nonverbal expression “more often falls within the terms of generally applicable regulations”); Williams, supra note 2, at 708 (rejecting the pure-speech/non-pure-speech distinction in part because it has been used by the Court “almost exclusively as a makeweight”). Indeed, as we have seen, the Court’s classic content-discrimination ruling came in a case that involved picketing—a form of “speech brigaded with action.” Brandenburg v. Ohio, 395 U.S. 444, 456 (1969) (Douglas, J., concurring); see also supra note 110 and accompanying text (discussing Mosley case).

  135. [135]. See supra note 11; see also Post, supra note 2, at 1257 (rejecting the pure-speech/non-pure-speech divide, which has “sometimes confused the Court’s First Amendment doctrine”).

  136. [136]. Humanitarian Law Project, 561 U.S. at 27–28.

  137. [137]. See Kathleen M. Sullivan & Noah Feldman, Constitutional Law 1150 (18th ed. 2013) (raising the question whether “it matter[s] that the statute specifically refers to ‘advice’ as a form of material support”). Notably, the challengers of the material-support statute emphasized in their brief to the Court that it “contains provisions that are content-based on their face.” Opening Brief for Humanitarian Law Project, et al. at 50, Holder v. Humanitarian Law Project, 561 U.S. 1 (2010) (No. 08-1498 & 09-89), 2009 WL 3865433, at *50 (citing Turner Broad. Sys. Inc. v. FCC, 512 U.S. 622, 642–43 (1994) for the proposition that “assertion of a content neutral purpose [is not] enough to save a law which, on its face, discriminates based on content”). Notably, this distinction might gain some traction from the academic work of then-Professor and now-Justice Elena Kagan. Kagan suggested that, when lawmakers enact a general prohibition that applies only in some instances to speech, they will be less likely to bring into play content-discriminatory purposes than when they pass laws that target speech per se, even in a content-neutral way. Kagan, supra note 2, at 496. The latter type of law, according to Kagan, will at least focus the legislative mind on speech effects, thus potentially piquing “hostility or sympathy toward particular messages.” Id. The former type of law, in contrast, will have a more “covert” effect on expression, thus reducing the risk of purposeful content discrimination. Id. Based on this idea, one might say that statutes (such as the material support statute) that deal with speech in express terms merit more intensive judicial scrutiny, because that treatment raises the odds that legislators will focus on the resulting chance to suppress unpopular messages.

  138. [138]. See Humanitarian Law Project, 561 U.S. at 27–28; see also Volokh, Speech as Conduct, supra note 2, at 1292 (noting the Court’s similar treatment of other breach-of-the-peace cases, including in the landmark Edwards v. South Carolina (1963) and Cantwell v. Connecticut (1940) cases). On this score, then-Professor Elena Kagan argued that breach-of-the-peace statutes raise special dangers that might make content-based-as-applied analysis distinctly justifiable for them. She noted, in particular, that the “vague standards” established by these laws permit enforcement officials “to take action based on their views of ideas” and because peace disturbances often result from speech activities “that raise the ire of the public,” thus creating a de facto (and improper) “delegation of authority to the public to suppress messages it disfavors.” Kagan, supra note 2, at 462–63. In addition, according to Kagan, history teaches that these laws are especially prone to abuse when compared to other laws that impose direct-in-effect burdens of “numerous, disparate, and crosscutting” kinds. Id. at 462. All of these points, it merits emphasis, might have provided a basis for distinguishing Humanitarian Law Project from Cohen v. California, and for declining to apply Volokh’s content-based-as-applied principle in the former case. But that is exactly what the Court did not do in dealing with the material-support statute. (And notably, in Hustler, no less than in both Humanitarian Law Project and Cohen v. California, the Court paid no attention to any statutory focus on speech—which obviously was absent with regard to the operative common-law emotional-distress tort—in finding unconstitutional the state’s application of a generally applicable law that imposed a direct-in-effect burden on speech. See Hustler Magazine, Inc. v. Falwell, 485 U.S. 46, 46–47 (1988).)

  139. [139]. See supra note 131 and accompanying text.

  140. [140]. Humanitarian Law Project, 561 U.S. at 28 (emphasis added).

  141. [141]. Id.

  142. [142]. Id. (emphasis added).

  143. [143]. See Reply Brief for Petitioner at 23–27, Holder v. Humanitarian Law Project, 561 U.S. 1 (2010) (Nos. 08-1498, 09-89), 2010 WL 302209.

  144. [144]. See Cole, supra note 132, at 152–55 (describing the Court’s analysis in these terms).

  145. [145]. See United States v. O’Brien, 391 U.S. 367, 367 (1968).

  146. [146]. Rodney A. Smolla, 1 Smolla and Nimmer on Freedom of Speech § 9:17 (2017); see also Turner Broad. Sys., Inc. v. F.C.C., 512 U.S. 622, 662 (1994) (citing OBrien as setting forth an “intermediate level of scrutiny” standard).

  147. [147]. OBrien, 391 U.S. at 377.

  148. [148]. See, e.g., Rubenfeld, supra note 2, at 771.

  149. [149]. See supra note 14 and accompanying text.

  150. [150]. See Emp’t Div. v. Smith, 494 U.S. 872, 874 (1990).

  151. [151]. Id. at 884. For a treatment of the difficulties the Court faced in sidestepping earlier authorities, see Michael W. McConnell, Free Exercise Revisionism and the Smith Decision, 57 U. Chi. L. Rev. 1109, 1120–28 (1990).

  152. [152]. Barnes v. Glen Theatre, Inc., 501 U.S. 560, 579 (1991) (Scalia, J., concurring).

  153. [153]. Cohen v. Cowles Media Co., 501 U.S. 663 (1991).

  154. [154]. See Bogen, supra note 2, at 222 (questioning whether the Court’s free-speech jurisprudence has been “moving from an intermediate level of scrutiny toward the absence of scrutiny adopted in its free exercise decisions”).

  155. [155]. See Dan T. Coenen, Where to, O’Brien? (unpublished manuscript) (on file with author).

  156. [156]. Emp’t Div. v. Smith, 494 U.S. 872, 890 (1990).

  157. [157]. Id. at 885 (citing Reynolds v. United States, 98 U.S. 145, 167 (1878)).

  158. [158]. See infra notes 312–15 and accompanying text.

  159. [159]. Smith, 494 U.S. at 878.

  160. [160]. See Barnes v. Glen Theatre, Inc., 501 U.S. 560, 579–80 (1991) (Scalia, J., concurring).

  161. [161]. See, e.g., Bogen, supra note 2, at 232–33 (suggesting that OBrien seems “inconsistent with Smith,” particularly because of concerns in both contexts “about balancing as a judicial technique ... [and] treating people unequally as a result of their subjective intentions.”); Rubenfeld, supra note 2, at 809, 810 n.96 (reasoning that Smith forecloses elevated scrutiny for generally applicable laws “not only for free exercise but also for free speech” cases, especially because in “many [cases] free-exercise claims can ... be recast as [free speech claims],” thus undermining the Smith rule). For a fortiori arguments for abandoning OBrien based on Smith, see Bogen, supra note 2, at 233 (arguing that differential approaches of OBrien and Smith are “anomalous” because prohibiting religious acts—including sacramental acts—“has more serious consequences” than limiting particular modes of expression, which can take other forms), and Dorf, supra note 2, at 1215 (noting the particularly draconian burden placed on the free-exercise right in Smith in comparison to the less onerous “particular means” burden imposed in OBrien) (quoting Kent Greenawalt, Free Speech in the United States and Canada, 55 Law & Contemp. Probs. 5, 27 (1992)). There is also a precedent-based consideration, rooted in pre-Smith, free-exercise law that may lend support to these a fortiori contentions. This is so because, prior to Smith, the Court applied the First Amendment even more aggressively in incidental-effect, free-exercise cases than incidental-effect, free-speech cases. Compare Wisconsin v. Yoder, 406 U.S. 205 (1972) (applying strict scrutiny), with United States v. O’Brien, 391 U.S. 367 (1968) (applying intermediate scrutiny). All of these a fortiori arguments are weakened, however, by the Court’s recognition of significant exceptions to its no-review rule in Smith itself, and still more restrictions in later cases. See generally Hosanna–Tabor Evangelical Lutheran Church & Sch. v. Equal Emp’t Opportunity Comm’n, 565 U.S. 171 (2012); Church of Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520 (1993). In the end, Smith at most calls for a thoughtful reexamination of the relation between free-exercise rights and free-speech rights in incidental-burden cases, and such a reexamination is set forth in the text that follows.

  162. [162]. See generally Volokh, Speech as Conduct, supra note 2, at 1298 (“The Free Exercise Clause and the Free Speech Clause protect different private interests, and courts have long interpreted them differently.”).

  163. [163]. Accord Tex. Monthly, Inc. v. Bullock, 489 U.S. 1, 8–9 (1989) (reasoning that government may not “favor the adherents of any sect or religious organization”); Wallace v. Jaffree, 472 U.S. 38, 60 (1985); Comm. for Pub. Educ. & Religious Liberty v. Nyquist, 413 U.S. 756, 792–93 (1973) (insisting that “the State ... pursue a course of ‘neutrality’ toward religion”).

  164. [164]. See, e.g., Good News Club v. Milford Cent. Sch., 533 U.S. 98, 114 (2001).

  165. [165]. In particular, the government is free to promote speech—for example, by creating designated public forums, subsidizing art, encouraging citizen input with regard to government decision-making, or disseminating its own preferred messages. On the other hand, whenever the government promotes religion, it creates tension with the Establishment Clause norm of neutrality. To be sure, First Amendment principles restrict the government from promoting some speech over other speech through the use of viewpoint-based or subject-matter-based restrictions. In contrast to the religion clauses, however, the Free Speech Clause does not condemn the government’s favoring speech over non-speech as a general matter.

  166. [166]. Speech—unlike religious practice—often specifically targets with criticism majoritarian practices and viewpoints, as well as existing political institutions. Given this fact, there is special reason to fear that political officials—who tend to be political officials precisely because they adhere to majoritarian viewpoints—will overregulate speech. See, e.g., Stone, Free Speech, supra note 2, at 277 (arguing that the risk of unfairly disadvantaging minorities is “especially potent” in “the realm of free speech”); id. (“[P]ublic officials will often be sorely tempted to silence dissent in order to insulate themselves from criticism and preserve their own authority.”).

  167. [167]. See generally Tribe, supra note 2, at §§ 12-27, 12-31, 12-36.

  168. [168]. See supra note 162 and accompanying text.

  169. [169]. See The Federalist No. 10 (James Madison) (noting that, if “you take in a greater variety of parties and interests; you [will] make it less probable that a majority of the whole will have a common motive to invade the rights of other citizens”); The Federalist No. 51 (James Madison) (noting the importance of the expanded republic for this reason, including in safeguarding “religious rights”).

  170. [170]. Palko v. Connecticut, 302 U.S. 319, 327 (1937); see also Post, supra note 2, at 1272 (noting that “speech is special,” because it “uniquely serves as the precondition to the very existence of social institutions and practices”).

  171. [171]. Two other considerations may also contribute to the Court’s distinctive treatment of free-speech and free-exercise claims in this context. First, application of the no-scrutiny rule in the free-exercise context greatly reduces the need to make rulings about whether activities stem from the exercise of “religion.” See Neutral Rules of General Applicability, supra note 2, at 1716–17 (implying that the Smith rule may reflect in part the particular difficulties of “defining religion”). Second, a regime of compulsory exemptions for religious practices pushes in the direction of requiring judicial sorting between practices that are “central” or not “central” to one’s religious life. See id.; see also United States v. Lee, 455 U.S. 252, 263 n.2 (1982) (Stevens, J., concurring) (“[T]he principal reason for adopting a strong presumption against [recognizing free-exercise-based exemptions to generally applicable laws] is not a matter of administrative convenience. It is the overriding interest in keeping the government—whether it be the legislature or the courts—out of the business of evaluating the relative merits of differing religious claims.”).

  172. [172]. See supra note 21 and accompanying text.

  173. [173]. See, e.g., Boy Scouts of Am. v. Dale, 530 U.S. 640, 645, 659 (2000) (distinguishing OBrien in invalidating a state public accommodations law as applied to the Boy Scouts’ decision to exclude an openly gay man from serving as a scoutmaster).

  174. [174]. See supra notes 54–90 and accompanying text.

  175. [175]. Turner Broad. Sys., Inc. v. FCC, 512 U.S. 622, 640 (1994).

  176. [176]. The most important feature of the relevant context is that the full citation following the quoted assertion reads as follows: “compare Cohen v. Cowles Media Co. with Barnes v. Glen Theatre, Inc.Id. (citations omitted). This “compare” cite suggests a focus on incidental burden cases because Barnes involved a generally applicable ban on nudity applied to dancers whose nudity was deemed to be express behavior, thus triggering analysis by the controlling four-Justice plurality pursuant to OBrien.

  177. [177]. See Garfield, supra note 2, at 1097 (noting that Cohen v. Cowles Media Co. “leaves the impression that, at least in some instances, a generally applicable law can be immune from First Amendment scrutiny even if it regulates speech”).

  178. [178]. See id. at 1088 n.6 (collecting cases, including breach-of-contract cases, in which courts invoked Cohen v. Cowles Media Co. in rejecting free-speech claims); Kagan, supra note 2, at 491 (positing the unavailability of any First Amendment defense to trespass-law prosecutions); Post, supra note 2, at 1251–52 (questioning applicability of United States v. OBrien in some incidental-burden cases, such as when a bus passenger carves political messages onto seats).

  179. [179]. For an explanation of independently illegal actions and their potential protection under the First Amendment as free speech, see Volokh, Speech as Conduct, supra note 2, 1315–16.

  180. [180]. See, e.g., Post, supra note 2, at 1252, 1257 (suggesting that OBrien does not apply to “violent crime,” including the act of “a terrorist who is prosecuted for a murder that successfully communicated a political message”); Schauer, Cuban Cigars, supra note 2, at 787 (“We would not want to open for [F]irst [A]mendment analysis every criminal prosecution, ... [such as when] the defendant claimed that ... [a] murder was [committed] for political and communicative purposes.”); see also Kagan, supra note 2, at 498 n.229 (stating she is “sure [that] courts would decline to apply First Amendment scrutiny to the conviction ... of a political terrorist for blowing up the Statue of Liberty”).

  181. [181]. Lawrence v. Texas, 539 U.S. 558, 562, 595 (2003).

  182. [182]. Branzburg v. Hayes, 408 U.S. 665, 698 (1972) (relying in part on tradition to reject constitutionally grounded reporter’s privilege); see also McIntyre v. Ohio Elections Comm’n, 514 U.S. 334, 375 (1995) (Scalia, J., dissenting) (“A governmental practice that has become general throughout the United States, and particularly one that has the validation of long, accepted usage, bears a strong presumption of constitutionality.”).

  183. [183]. Indeed, in OBrien itself the Court concluded that a law prohibiting the destruction of a draft card was sustainable under intermediate-scrutiny review. To say the least, it should follow a fortiori that a law prohibiting the destruction of a human being would satisfy that same test. See Dorf, supra note 2, at 1245 (noting that, even if the First Amendment requires review of murder law as applied to ritual human sacrifice, “the government will easily meet its justificatory burden”).

  184. [184]. Cohen v. Cowles Media Co., 501 U.S. 663, 671–72 (1991).

  185. [185]. E. Allan Farnsworth, Contracts, § 2.19 (3d ed. 1999) (laying out “the common law’s traditional view of reliance”).

  186. [186]. See supra note 183 and accompanying text. Under another possible view of the Cohen v. Cowles Media Co. limit on access to OBrien-based review, such review would not extend to a law (whether traditional or not) so long as it reaches mainly non-speech behavior. The justification for such an approach is that lawmakers are unlikely to enact such laws out of animus or insensitivity to speakers; rather, because non-speakers will predictably oppose the passage of such a restriction, it is unlikely that the legislature will act unless strong justifications for its intervention exist. See, e.g., Sullivan & Feldman, supra note 137, at 1211 (noting that operation of generally applicable laws on non-speakers may create “political process” dynamics that “will help protect unpopular speakers” against the enactment of laws that otherwise would burden them); see also Kagan, supra note 2, at 496 (describing vandalism, tax laws, and zoning laws as having a “breadth [that] makes them poor vehicles for censorial designs; they are instruments too blunt for effecting, or even reflecting, ideological disapproval”). Much the same set of problems flagged in the text for the traditional-prohibition exception, however, would apply to this posited sweeping-prohibition exception. See also supra note 131 (noting the Court’s willingness in Humanitarian Law Project to reaffirm its prior acceptance of as-applied challenge to a breach-of-the-peace statute, notwithstanding its wide-ranging application to non-speech cases). Another possible approach—which has a close kinship to the would-be sweeping-prohibition exception—would distinguish between generally applicable laws based “on the likelihood of a speech-suppressive administrative motivation.” Srinivasan, supra note 2, at 420. In developing this idea, Srikanth Srinivasan distinguished Cohen v. Cowles Media Co. from Barnes and OBrien on precisely this ground, emphasizing that a private person—and not the government—had initiated the promissory estoppel suit. Id. (positing that the “unstated underpinning” of the Cohen v. Cowles Media Co. “decision may well be the impossibility of an illicit administrative motive”); see also Bogen, supra note 2, at 230 n.111 (offering additional reasons for detecting a greater risk of speech-suppressive motives in Barnes than Cohen v. Cowles Media Co.). Yet another approach to limiting OBrien would be to distinguish between generally applicable laws based on whether they impose “substantial” or “insubstantial” burdens on speech. See Dorf, supra note 2, at 1182. Both the would-be insubstantial-danger-of-wrongful-motive approach and the insubstantial-burden approach to limiting OBrien are discussed below. See infra notes 324–42 and accompanying text. These approaches, not surprisingly, pose problems of their own, in part because they would present courts with difficult sorting tasks and because they are hard to reconcile with key Supreme Court precedents. See id.

  187. [187]. See Volokh, Speech as Conduct, supra note 2, at 1297.

  188. [188]. Garfield, supra note 2, at 1087 (describing Cohen v. Cowles Media Co. as “sloppily reasoned” and thus exerting an “insidious influence on First Amendment law”); id. at 1125 (describing the Court’s opinion as a “hodgepodge of poorly reasoned explanations”).

  189. [189]. Cohen, 501 U.S. at 665.

  190. [190]. Gardebring v. Jenkins, 485 U.S. 415, 430–31 n.17 (1988).

  191. [191]. United States v. Carlton, 512 U.S. 26, 39 (1994) (Scalia, J., concurring).

  192. [192]. Arnett v. Kennedy, 416 U.S. 134, 153–54 (1974) (plurality opinion).

  193. [193]. For articles that touch on this idea, see Jerome A. Barron, Cohen v. Cowles Media and Its Significance for First Amendment Law and Journalism, 3 Wm. & Mary Bill of Rts. J. 419, 456 (1994), and Garfield, supra note 2, at 1087.

  194. [194]. See Srinivasan, supra note 2, at 419–20 (focusing on private, as opposed to state, enforcement decision in Cohen v. Cowles Media Co.).

  195. [195]. See Garfield, supra note 2, at 1126 (expressing worries about a waiver principle that would “extend a welcome mat for private parties, and particularly businesses, to shut down the flow of embarrassing information by wrapping that information in a veil of property and contract rights”). Some support for this view comes from Snepp v. United States, 444 U.S. 507 (1980). There, a former CIA agent breached a promise to allow the agency to review any written materials concerning the agent’s work prior to publication. Id. at 508. The Court’s analysis focused on remedial issues. Id. at 514. The agent, however, had also challenged enforcement of the promise itself as inconsistent with the First Amendment. See id. at 509 n.3. Notably, the Court did not respond to this argument by declaring, without more, that Snepp had waived his rights “voluntarily.” Id. It went on to emphasize that the particular agreement was supported by “a compelling interest in protecting ... the secrecy of information important to our national security” and that “[t]he agreement that Snepp signed is a reasonable means for protecting this vital interest.” Id.

  196. [196]. See, e.g., Garfield, supra note 2, at 1119–20 (raising the contention that, despite Cohen v. Cowles Media Co., waiver law should give way in free-speech cases because at stake are not only “the interests of the speaker but also those of the public which is deprived of the information”); id. at 1114 (noting that enforcement of nondisclosure agreements is especially problematic when there are no “alternative channels for communication of the information”).

  197. [197]. United States v. Carlton, 512 U.S. 26, 39 (1994) (Scalia, J., concurring).

  198. [198]. See supra notes 160–65 and accompanying text.

  199. [199]. Barnes v. Glen Theater, Inc., 501 U.S. 560, 572, 57679 (1991) (Scalia, J., concurring).

  200. [200]. Id. at 579.

  201. [201]. See supra notes 152–88 and accompanying text.

  202. [202]. Barnes, 501 U.S. at 56672.

  203. [203]. See Alexander, supra note 2, at 945 (making the case that elimination of review for noncommunicative-effect cases—including cases such as OBrien—“would not have a great impact because [such] challenges ... almost never succeed”); Neutral Rules of General Applicability, supra note 2, at 1726 n.76 (describing OBrien test as now “functionally useless”).

  204. [204]. See, e.g., United States v. Albertini, 472 U.S. 675 (1985); Clark v. Cmty. for Creative Non-Violence, 468 U.S. 288 (1984).

  205. [205]. See supra notes 52–90 and accompanying text.

  206. [206]. See Kagan, supra note 2, at 495 (arguing with respect to “all laws incidentally restricting speech,” that “[i]f, as seems likely, most of the laws would pass constitutional muster, incurring [the] costs” of individualized judicial assessments of constitutionality “does not seem worthwhile,” so that courts should “assume from the outset that these laws raise no serious problem”).

  207. [207]. See, e.g., Bogen, supra note 2, at 255–58 (citing Turner Broadcasting and commercial speech cases as illustrative of how the Court has applied the OBrien test in an “invigorated” way and “can examine laws closely” under that standard); Dorf, supra note 2, at 1201 n.101 (noting the Court’s recent receptiveness to “a more stringent form of intermediate scrutiny” for content-neutral laws). For some rulings that have invalidated laws pursuant to OBrien-type intermediate-scrutiny analysis, see Turner Broad. Sys., Inc. v. FCC, 512 U.S. 622, 662 (1994) (applying OBrien test to “must carry” rule applied to cable television operators and remanding for more careful review); and Procunier v. Martinez, 416 U.S. 396, 419 (1974) (invalidating prison mail restriction pursuant to the OBrien test).

  208. [208]. See Kagan, supra note 2, at 497 n.226 (noting that OBrien-based intermediate scrutiny “has the potential to matter”); Srinivasan, supra note 2, at 407 (noting that “[t]he Court would likely engage in a more searching inquiry of incidental effects in some circumstances”—a point of importance because lower courts will have more opportunities than the Supreme Court to operate in this way); Williams, supra note 2, at 708 n.344 (asserting that: (1) “OBrien, on its face,” does not set forth a “toothless” test; and (2) “the Court often has allowed lower federal courts to apply the OBrien test fairly strictly”).

  209. [209]. Arcara v. Cloud Books, Inc., 478 U.S. 697 (1986).

  210. [210]. See id. at 698–99.

  211. [211]. Id. at 698.

  212. [212]. Id. at 699.

  213. [213]. Id. at 701.

  214. [214]. Id. at 701–02.

  215. [215]. Id. at 706–07.

  216. [216]. Id. at 706.

  217. [217]. Id. at 705.

  218. [218]. See id. at 706.

  219. [219]. Id.

  220. [220]. Id. at 708 (O’Connor, J., concurring).

  221. [221]. See id.; see also Dorf, supra note 2, at 1178 (noting the “obvious constitutionality” of some laws that have doubly incidental effects on expression, including environmental and minimum wage laws applied to newspaper publishers).

  222. [222]. See Arcara, 478 U.S. at 708–09 (Blackmun, J., dissenting).

  223. [223]. Id. at 711–12 (Blackmun, J., dissenting) (citations omitted).

  224. [224]. Id. at 706–07. See generally infra Part IV.A.1 (discussing the inevitable-effect exception to Arcara).

  225. [225]. United States v. Albertini, 472 U.S. 675 (1985).

  226. [226]. See infra Part V.A.2 (discussing the intimate-relation exception).

  227. [227]. See infra Part V.A.3 (discussing the improper-purpose exception).

  228. [228]. See infra Part V.C (discussing the hybrid-rights exception).

  229. [229]. See infra Part V.B.1 (discussing the lawful activity exception).

  230. [230]. See infra Part V.B.2 (discussing the unprotected-speech exception).

  231. [231]. See infra Part V.D.2 (discussing the substantial-effects exception proposed by Dorf).

  232. [232]. See infra Part V.D.1 (discussing bad-purpose-prophylaxis exception proposed by Srinivasan).

  233. [233]. Arcara v. Cloud Books, Inc., 478 U.S. 697, 707 (1986).

  234. [234]. Minneapolis Star & Tribune Co. v. Minn. Comm’r of Revenue, 460 U.S. 575 (1983).

  235. [235]. Arcara, 478 U.S. at 704 (citation omitted).

  236. [236]. Id. at 704.

  237. [237]. Id. at 705.

  238. [238]. Id. at 704.

  239. [239]. Id. at 705.

  240. [240]. For example, in Forbes v. City of Seattle, the Washington Supreme Court considered a First Amendment challenge to a local ordinance that levied a five percent tax on all admission fees collected by local businesses. Forbes v. City of Seattle, 785 P.2d 431, 432–33 (Wash. 1990). The court examined, in particular, whether the ordinance “targets a subgroup of First Amendment activities (here, motion picture theaters) for taxation,” in a way that made it subject to scrutiny under Minneapolis Star. Id. at 435 (citations omitted). In the end, the court found the Minneapolis Star principle inapplicable, but distinguished one earlier case in which a generally phrased admission tax in practice reached only the challenger’s business, and another case in which “90 percent of the admissions tax was borne by four businesses, all of which were engaged in protected speech.” Id. at 436 (citations omitted) (emphasis in original). The court emphasized that the law in Forbes operated in a different manner because under it several hundred businesses were subject to taxation, only two dozen of which were movie theaters, which accounted for only 20% of total tax revenues. Id. What if, however, speaker-taxpayers in Forbes had accounted for 60%, 70%, or 80% of tax payments? The question highlights the inherent difficulty of applying the inevitable-effect principle.

  241. [241]. Cal. Civ. Code § 1708.8 (West 2016).

  242. [242]. Raef v. Superior Court, 193 Cal. Rptr. 3d 159, 166 (Cal. Ct. App. 2015).

  243. [243]. Id. at 167.

  244. [244]. Perhaps a better explanation of the result in the case rests on the long-recognized distinction between gathering information and communicating information to others. See, e.g., Barry P. McDonald, The First Amendment and the Free Flow of Information: Towards a Realistic Right to Gather Information in the Information Age, 65 Ohio St. L.J. 249, 265–72 (2004) (noting decisions that afford more First Amendment protection to communicative activities than to news-gathering activities). The court in Raef, however, did not make this distinction the centerpiece of its analysis.

  245. [245]. In another case, for example, the Pennsylvania Supreme Court applied Arcara’s no-scrutiny rule after a local government wielded its eminent domain power to close an adult theatre as part of an effort to rehabilitate a blighted three-block area located downtown. In re Condemnation by Urban Redevelopment Auth., 913 A.2d 178, 186–87 (Pa. 2006). The court, however, noted that it would have reached a different result—apparently under the Minneapolis Star principle—if no other adult theatres were able to operate in the city. Id. at 186. But how far, in spinning out this thread, would the court go? What if, for example, one other theatre was in operation but it lay on the outskirts of a large city in which cross-town driving was uncommon? What if the city were extremely large—the size, perhaps, of Seattle or Milwaukee—and only two or three such establishments operated within its boundaries? What if the city were small, no other theatres operated within its boundaries, but theatres did operate in neighboring towns? With regard to this last question, in Schad v. Borough of Mount Ephraim, 452 U.S. 61 (1981), a town in New Jersey sought to sanction a business that featured nude dancing for breaching a local zoning restriction that prohibited “live entertainment.” Id. at 63–64. On appeal, one justification offered by the city in support of the ordinance was that nude dancing venues operated in surrounding areas. Id. at 64. The Court, however, rejected this argument, stating that “[o]ne is not to have the exercise of his liberty of expression in appropriateplaces abridged on the plea that it may be exercised in some other place.” Id. at 76–77 (quoting Schneider v. State, 308 U.S. 147, 163 (1939)).

  246. [246]. Arcara v. Cloud Books, Inc., 478 U.S. 697, 704 (1986).

  247. [247]. Id. at 707.

  248. [248]. Id. at 704.

  249. [249]. Id. at 706 n.3 (emphasis added) (citations omitted).

  250. [250]. Id. at 706 (emphasis added). See also id. at 702 (describing OBrien as a case in which “the otherwise unlawful burning of a draft card was to ‘carry a message’”); id. at 705 (distinguishing the law challenged in Arcara from the one challenged in OBrien on the ground that “the sexual activity carried on in this case manifests absolutely no element of protected expression”).

  251. [251]. United States v. Albertini, 472 U.S. 675, 677 (1985). As the Court noted, the Air Force issued the bar letter against Albertini because he and a companion “obtained access to secret Air Force documents and destroyed the documents by pouring animal blood on them.” Id.

  252. [252]. Id. at 677–78.

  253. [253]. Id. at 678–79.

  254. [254]. Id. at 679.

  255. [255]. Id. at 688–89.

  256. [256]. Arcara v. Cloud Books, Inc., 478 U.S. 697, 703 (1986).

  257. [257]. See Srinivasan, supra note 2, at 412 (noting that, because “Albertini did not begin his protest until after he had illegally reentered the base, the conduct that drew application of the law (the initial reentry) was not itself expressive”); Stone, supra note 2, at 110 (observing that reentry, not speech, triggered the sanction in Albertini).

  258. [258]. Kagan, supra note 2, at 498 n.228 (noting that “the reentry ... was not itself expressive”).

  259. [259]. See Dorf, supra note 2, at 1206 (discussing the facilitation of speech as a trigger for First Amendment scrutiny); Srinivasan, supra note 2, at 413 (noting that one could “interpret the Arcara rule so that it would accommodate Albertini” by holding that a law is worthy of scrutiny when it “facilitates some expressive activity”); Williams, supra note 2, at 723–24, 727 (suggesting that elevated review should sometimes apply to interferences with “facilitative aspects of speech”).

  260. [260]. See Srinivasan, supra note 2, at 413 (challenging the proposed facilitation-based exception).

  261. [261]. See supra note 204 and accompanying text.

  262. [262]. See Schauer, Cuban Cigars, supra note 2, at 787 (questioning the application of intermediate scrutiny in “every trespass prosecution ... solely because the trespasser trespassed for the purpose of communicating”).

  263. [263]. See, e.g., Williams, supra note 2, at 724 (“There is ... no clear dividing line between facilitative aspects of speech and other activities.”).

  264. [264]. See, e.g., Hodgkins ex rel. Hodgkins v. Peterson, 355 F.3d 1048, 1058–59, 1064 (7th Cir. 2004) (invalidating a youth curfew law under the Free Speech Clause in part because leaving home was a “necessary precursor” to engaging in speech activities).

  265. [265]. Arcara v. Cloud Books, Inc., 478 U.S. 697, 706–07 (1986); see also id. at 708 (O’Connor, J., concurring) (“If ... a city were to use a nuisance statute as a pretext for closing down a bookstore because it sold indecent books ... the case would clearly implicate First Amendment concerns and require analysis under the appropriate First Amendment standard of review.”).

  266. [266]. State of Ohio ex. rel. Konstam v. Video Visions, Inc., No. 93-CA-38, 1994 WL 167925, at *1–2 (Ohio Ct. App. Apr. 28, 1994).

  267. [267]. Id. at *3.

  268. [268]. Id. at *4.

  269. [269]. Id.

  270. [270]. See id.

  271. [271]. See id.

  272. [272]. United States v. O’Brien, 391 U.S. 367, 383 (1968).

  273. [273]. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2228 (2015); see also Ward v. Rock Against Racism, 491 U.S. 781, 791 (1989). For additional Free Speech Clause authorities along these lines, see Dan T. Coenen, A Constitution of Collaboration: Protecting Fundamental Values with Second-Look Rules of Interbranch Dialogue, 42 Wm. & Mary L. Rev. 1575, 1760 n.784 (2001); Theodore Eisenberg, Disproportionate Impact and Illicit Motive: Theories of Constitutional Adjudication, 52 N.Y.U. L. Rev. 36, 99–156 (1977) (analyzing in detail how the Court approaches the issue of impermissible legislative motives); Farber, supra note 2, at 745 n.94 (critiquing OBrien’s failure to recognize improper legislative purpose as a basis for invalidating statutes); Kagan, supra note 2, at 427 n.43 (citing recent motive-based regimes); and Schauer, Cuban Cigars, supra note 2, at 781 n.15 (endorsing wrongful-purpose-based invalidations based in part on past Supreme Court authority). The most recent treatment of purpose-based constitutional challenges is Richard H. Fallon, Jr., Constitutionally Forbidden Legislative Intent, 130 Harv. L. Rev. 523 (2016).

  274. [274]. See, e.g., Hunter v. Underwood, 471 U.S. 222, 229 (1985).

  275. [275]. See, e.g., Edwards v. Aguillard, 482 U.S. 578, 608 (1987).

  276. [276]. See, e.g., Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520, 540 (1993).

  277. [277]. See generally Coenen, supra note 273, at 1755–73 (discussing a wide array of cases in which the Court has invalidated legislation on the ground that it was enacted in furtherance of unconstitutional purposes). Notably, the Court’s free-exercise ruling in Church of the Lukumi Babalu Aye may be especially significant in this regard. After all, as we have seen, the Court’s ruling in the Smith case reflects a decision to apply the Free Speech Clause more aggressively than the Free Exercise Clause as a general matter in cases involving generally applicable laws. See supra notes 153–65 and accompanying text. It thus would seem anomalous for the Court to apply the Free Speech Clause less aggressively than the Free Exercise Clause in assessing the constitutionality of generally applicable laws based on claims of impermissible purpose.

  278. [278]. See, e.g., United States v. O’Brien, 391 U.S. 367, 383–84 (1968).

  279. [279]. Other considerations may support this same hesitance as well. See generally Coenen, supra note 273, at 1758. Moreover, courts may find reasons to be more aggressive in responding to wrongful-purpose challenges in some constitutional settings than others. See Stormans, Inc. v. Wiesman, 136 S. Ct. 2433, 2437 n.3 (2016) (Alito, J., dissenting from denial of cert.) (noting that different approaches might apply in the equal-protection and free-exercise contexts).

  280. [280]. See note 268 and accompanying text (quoting State of Ohio ex. rel. Konstam); see also Sanchez v. Warden, 570 A.2d 673, 677 (Conn. 1990) (upholding a rule that required headphones to listen to radios in prison; but investigating, in so doing, whether prison authorities had acted with the goal of limiting overall access to communications); Williams, supra note 2, at 724 (“[R]egulation derived from an actual intent to silence certain content categories is subject to strict review ....”).

  281. [281]. In re Condemnation by Urban Redevelopment Auth. of Pittsburgh, 913 A.2d 178, 180–81 (Pa. 2006).

  282. [282]. Id. at 183.

  283. [283]. Id. at 186–87.

  284. [284]. Id.

  285. [285]. Id. at 191 (Saylor, J., dissenting).

  286. [286]. Arcara v. Cloud Books, Inc., 478 U.S. 697, 705 (1986) (emphasis added).

  287. [287]. Id. at 705–06.

  288. [288]. In re Condemnation by Urban Redevelopment Auth., 913 A.2d at 191 (Saylor, J., dissenting).

  289. [289]. Id. at 187–88 (majority opinion).

  290. [290]. Id. at 191 (Saylor, J., dissenting).

  291. [291]. People v. Sequoia Books, Inc., 537 N.E.2d 302, 309 (Ill. 1989).

  292. [292]. Id. at 304.

  293. [293]. Id. at 307–08.

  294. [294]. Id. at 309 (quoting Arcara v. Cloud Books, Inc., 478 U.S. 697, 707 (1986)).

  295. [295]. Id. at 309.

  296. [296]. See, e.g., Miller v. California, 413 U.S. 15, 23 (1973).

  297. [297]. Sequoia Books, 537 N.E.2d at 309.

  298. [298]. Alexander v. United States, 509 U.S. 544 (1993).

  299. [299]. Id. at 548.

  300. [300]. Id. at 552–53.

  301. [301]. Id. at 557 (citation omitted).

  302. [302]. Id. (citation omitted).

  303. [303]. Id. at 575 (Kennedy, J., dissenting).

  304. [304]. Arcara v. Cloud Books, Inc., 478 U.S. 697, 707 (1986).

  305. [305]. Alexander, 509 U.S. at 573 (Kennedy, J., dissenting).

  306. [306]. Id. at 576. See also Schauer, Cuban Cigars, supra note 2, at 790 (noting “the special character of books” in arguing that the Court might logically apply more exacting scrutiny to government restrictions that impair their distribution because “the relationship between books and the principles of freedom of speech and freedom of press need not be insulted by a citation”).

  307. [307]. Alexander, 509 U.S. at 575 (Kennedy, J., dissenting).

  308. [308]. Id. at 578.

  309. [309]. Id. at 575.

  310. [310]. See, e.g., Reed v. Town of Gilbert, 135 S. Ct. 2218, 2224 (2015) (applying strict scrutiny in striking down a municipal law that sought to regulate signs related to meetings of nonprofit groups, reasoning that the provisions were “content-based regulations of speech”); United States v. Alvarez, 132 S. Ct. 2537, 2548 (2012) (invalidating a statute that prohibited the act of falsely claiming receipt of military medals); Snyder v. Phelps, 562 U.S. 443, 460 (2011) (holding that the First Amendment barred a plaintiff from recovering damages for intentional infliction of emotional distress against individuals affiliated with Westboro Baptist Church, who picketed the funeral of the plaintiff’s son, a military veteran killed in Iraq); United States v. Stevens, 559 U.S. 460, 482 (2010) (holding that a federal statute that outlawed depictions of animal cruelty violated the First Amendment). See generally Dan T. Coenen, Freedom of Speech and the Criminal Law, 97 B.U. L. Rev. 1533 (2017).

  311. [311]. See, e.g., Alexander, 509 U.S. at 578 (Kennedy, J., dissenting) (condemning the majority’s approach as “deplorable” in its “abandonment of fundamental First Amendment principles” and as embodying a “flagrant violation of the right of free speech”).

  312. [312]. Murdock v. Pennsylvania, 319 U.S. 105 (1943).

  313. [313]. Emp’t Div. v. Smith, 494 U.S. 872, 881 (1990).

  314. [314]. Murdock, 319 U.S. at 112.

  315. [315]. Smith, 494 U.S. at 881.

  316. [316]. Hodgkins ex rel. Hodgkins v. Peterson, 355 F.3d 1048 (7th Cir. 2004).

  317. [317]. Id. at 1058–59.

  318. [318]. Id. at 1051 (noting that the parents raised due process arguments as well). For cases applying intermediate scrutiny to children’s curfew laws that parents challenged on the ground that such laws violate the First Amendment, the Fourth Amendment, and the Due Process Clause, see Ramos v. Town of Vernon, 353 F.3d 171, 173, 176 (2d Cir. 2003), and Hutchins v. District of Columbia, 188 F.3d 531, 535, 541 (D.C. Cir. 1999).

  319. [319]. Smith, 494 U.S. at 881–82 (discussing Wisconsin v. Yoder, 406 U.S. 205 (1972), among other cases).

  320. [320]. See, e.g., Shaman, supra note 2, at 443–44 (“Commentators have been routinely disdainful of the hybrid-rights exception, viewing it as a bungled attempt to distinguish disagreeable precedent, and the Sixth Circuit refused to follow it on the ground that it was ‘completely illogical.’” (footnote omitted) (quoting Kissinger v. Bd. of Trs. of the Ohio State Univ., Coll. of Veterinary Med., 5 F.3d 177, 180 (6th Cir. 1993))).

  321. [321]. See Michael Coenen, Combining Constitutional Clauses, 164 U. Pa. L. Rev. 1067, 1077–86 (2016) (analyzing cases in which the Court has applied different hybrid-rights doctrines); David L. Faigman, Measuring Constitutionality Transactionally, 45 Hastings L.J. 753, 772–78 (1994) (discussing the aggregation of rights in constitutional analysis); Stephen Kanter, The Griswold Diagrams: Toward a Unified Theory of Constitutional Rights, 28 Cardozo L. Rev. 623, 638–40 (2006) (discussing cases in which the Court has employed hybrid-rights analysis).

  322. [322]. See supra notes 148–62 and accompanying text.

  323. [323]. Consider, for example, a variation of the facts of Arcara that involve the closing not of an adult bookstore, but of a church (in which the since-fired pastor had engaged in prostitution). In part because of the combination of communicative and religious activities that occur within church communities, the tug of hybrid-rights analysis might well prove hard to resist in such a setting.

  324. [324]. Srinivasan, supra note 2, at 420.

  325. [325]. See supra notes 265–80 and accompanying text.

  326. [326]. Srinivasan, supra note 2, at 417.

  327. [327]. Id. at 416 n.77 (contrasting the proposed test, which focuses on “the danger of a speech-suppressive administrative motive,” from a test founded on proof of an “outright . . . intent to restrict speech”).

  328. [328]. Id. at 417–19. Albertini, in Srinivasan’s view, presented a “danger that the military applied the reentry law against Albertini in order to suppress his [anti-nuclear-arms] speech, or at least that the decision to apply the law was a reaction to his demonstration.” Id. at 417. More specifically, the presence of that danger rendered Albertini distinguishable from the hypothetical case, posited by Justice O’Connor in Arcara, involving application of a speeding law to the late-running reporter. Thus, “if Albertini had been cited for speeding while traveling to the military base, it would have been extremely unlikely that a desire to restrict his expressive activities provoked his arrest.” Id. at 416. Indeed, “[e]xcept under peculiar circumstances, his plans to go to the base and his intent to engage in a protest there would have been unknown to the arresting agent.” Id. In the actual Albertini case, however, there were “special reasons,” id. at 420, that pointed to a “danger of improper administrative motivation.” Id. at 418. Accordingly, while it would “normally be the case” that the no-scrutiny rule of Arcara should operate when “a distinct, nonexpressive violation . . . triggers application of a law,” in some cases—such as Albertini—it would be proper for courts to apply a heightened style of means-ends scrutiny. Id. at 418–19.

  329. [329]. See, e.g., Bogen, supra note 2, at 231 (asserting that Srinivasan’s reading of Cohen v. Cowles Media Co. “was not stated as a standard for decision” in that case, and that application of this standard defies “easy judgment”).

  330. [330]. Srinivasan, supra note 2, at 420.

  331. [331]. See Kagan, supra note 2, at 499 (“[T]he Court’s decision to apply intermediate review to certain incidental restrictions may result not so much from use of the Arcara test as from a visceral sense that an illicit factor entered into a governmental decision—whether legislative, administrative, or judicial.”).

  332. [332]. Dorf, supra note 2, at 1179.

  333. [333]. Id. (emphasis added). Professor Dorf does not address with crystalline clarity how his “substantially burdening” approach to “incidental effect” problems intersects with Arcara and the doubly-incidental-burden category of cases it exemplifies. His own articulation of the “substantially burdening” test, however, suggests that it applies to all forms of incidental-effect cases, and he states that in Arcara “the Court subjected an incidental burden on speech to no First Amendment scrutiny at all.” Id. at 1205 (emphasis added). For this reason, it appears that Dorf would apply his “substantially burdening” limiting principle not only in OBrien-type incidental burden cases, but also to doubly-incidental-burden cases like Arcara. See Campbell, supra note 2, at 4 (indicating that Dorf’s approach applies to “incidental burdens on speech, whether falling on expressive or nonexpressive conduct”). For the suggestion of an approach akin to Professor Dorf’s, see Stone, Content-Neutral Restrictions, supra note 2, at 112–13 (pointing to a difference between incidental-burden cases that involve a “significant effect” and a “modest effect” on speech rights, and placing OBrien in the latter category).

  334. [334]. Dorf, supra note 2, at 1240.

  335. [335]. See supra Part V.B.2. For additional critiques of Dorf’s approach, see Bogen, supra note 2, at 249 (challenging the “substantial burden” test on the ground that “it is arbitrary” and in effect applies “balancing ... to all laws impacting speech”); Campbell, supra note 2, at 9, 58–62 (expressing concerns that this approach might well prove “unmanageable and underprotective of speech interests”); and Srinivasan, supra note 2, at 414–15 (challenging the “‘substantial’ speech-restrictive effect” approach, in part on the ground that it would produce uncertainties in “countless cases”).

  336. [336]. See, e.g., Boy Scouts of Am. v. Dale, 530 U.S. 640, 659 (2000) (applying strict scrutiny on an expressive-association theory based on the conclusion that “a state requirement that the Boy Scouts retain Dale ... would significantly burden the organization’s right to oppose or disfavor homosexual conduct”). See also Stone, Free Speech, supra note 2, at 298 (explaining NAACP v. Alabama and Brown v. Socialist Workers ‘74 Campaign Committee as cases involving the invalidation of “laws of otherwise general application” because their effects on speech were “particularly severe”); Stone, Content-Neutral Restrictions, supra note 2, at 63–64 (suggesting that the theme of the Court’s earlier decisions in Button, Brown, and Jaycees involved the presence or absence of a “significant” or “serious” burden on free speech rights); Williams, supra note 2, at 713 n.360 (noting the Court’s decisions in Brown and Bates and attributing its invalidation of generally applicable laws in those cases to “the exceptionally severe impact of disclosure requirements on unpopular groups”).

  337. [337]. See Alexander v. United States, 509 U.S. 544, 576 (1993) (Kennedy, J., dissenting) (arguing that, while “the sanction in Arcara did not involve a complete confiscation or destruction of protected expression,” the state’s action here resulted in the forfeiture “of hundreds of original titles and thousands of copies”).

  338. [338]. Arcara v. Cloud Books, Inc., 478 U.S. 697, 705 (1986).

  339. [339]. Dorf, supra note 2, at 1198–99 (“[O]ne must ... confront the practical problem of how to recognize incidental burdens without invalidating all legislation—that is, the floodgates concern.”); Stone, Content-Neutral Restrictions, supra note 2, at 107 (noting the “Pandora’s box” and the “nightmare of judicial administration” arguments against judicial scrutiny of generally applicable laws).

  340. [340]. See supra notes 226–32 and accompanying text (detailing actual and potential exceptions identified by courts to the Arcara rule). See also Stone, Content-Neutral Restrictions, supra note 2, at 107 (defending the Court’s openness to reviewing generally applicable laws on free-speech grounds because “some laws that have only an incidental effect on free expression may have a substantial restrictive effect”). Also supporting the pull toward greater flexibility is that idea that the “balancing inquiry need not be an involved one in every case.” Srinivasan, supra note 2, at 406 (challenging “First Amendment overload” argument on this ground). For another proposed limit on the operation of Arcara proposed by a First Amendment scholar, see generally Campbell, supra note 2 (advocating an exception focused on “speech-facilitating conduct”).

  341. [341]. See Dorf, supra note 2, at 1199 (“[C]urrent Supreme Court doctrine attempts to reconcile these theoretical and practical considerations by treating incidental burdens as infringements on constitutional rights in some, but not all, circumstances.”).

  342. [342]. Pre-Arcara rulings also suggest the tendency of courts to consider the degree and proportionality of the burden imposed on speech in assessing the constitutionality of generally applicable laws. See, e.g., Commonwealth v. Croatan Books, Inc., 323 S.E.2d 86, 87 (Va. 1984) (noting the suitability of the store-closure remedy in light of the “continuous and pervasive nature” of the illegal conduct); Commonwealth ex rel. Lewis v. Allouwill Realty Corp., 478 A.2d 1334, 1338 (Pa. Super. Ct. 1984) (upholding closure order in part because it allowed for remedial adjustment based on a demonstration “that the premises are going to be used for proper activities”); People ex rel. Sorenson v. Randolph, 160 Cal. Rptr. 69, 72 (Cal. Ct. App. 1979) (noting that the burden placed on the abated business was rightly proportionate in light of “a history of police problems and ... prior warnings [that] had gone unheeded”). See generally Victor V. Vicinaiz, The Content Distinction and Freedom of Expression: Arcara v. Cloud Books, Inc., 20 Creighton L. Rev. 893, 904–07 (1986) (discussing these and other cases). In addition, courts might well take account of the context-specific substantiality of burdens on speech in applying state constitutional protections of free expression—particularly since this is exactly what happened in Arcara itself. See People ex rel. Arcara v. Cloud Books, Inc., 503 N.E.2d 492, 495 (N.Y. 1986) (finding a violation of the state constitution following remand; distinguishing the speeding reporter case on the ground that the burden on free speech presented there was “slight and indirect” compared to the burden created by “closing a bookstore for a year”).

  343. [343]. See supra notes 312–23 and accompanying text.

  344. [344]. Notably, in Stanley v. Georgia, 394 U.S. 557 (1969), the Court gave protection to supposedly unprotected obscene speech when viewed within the home. See id. at 559. Building on Stanley, courts might view obscenity as meriting a measure of Free Speech Clause protection, at least when (as in Alexander) its suppression carries with it a far-reaching forfeiture of wholly protected speech materials (and especially so when that taking of property is wholly uncompensated).

  345. [345]. See generally Dan T. Coenen, Free Speech and the Law of Evidence (Jan. 6, 2017) (unpublished manuscript) (on file with author). Notably, thoughtful commentators have investigated how constitutional free-speech law and generally applicable rules of evidence should intersect in cases that involve politically controversial speech, but they have done so without considering the possible relevance of the principles laid down in OBrien and Arcara. See generally Robert P. Faulkner, Evidence of First Amendment Activity at Trial: The Articulation of A Higher Evidentiary Standard, 42 UCLA L. Rev. 1 (1994); Peter E. Quint, Toward First Amendment Limitations on the Introduction of Evidence: The Problem of United States v. Rosenberg, 86 Yale L.J. 1622 (1977).

  346. [346]. See Coenen, supra note 345.

  347. [347]. By way of example, Professor Cole has argued that the Court in Humanitarian Law Project departed from the analytical approach normally directed at content-discriminatory laws by applying “deferential strict scrutiny.” Cole, supra note 132, at 158 (citation omitted). It remains to be seen whether any such tweaking of operative doctrine will take hold, either as a formal matter or as a practical matter, in future cases.

  348. [348]. The Court’s ruling in Thomas v. Chicago Park Dist. illustrates the complexities raised by public-forum cases. In that case, the Court rejected a Speech Clause challenge to a law that required a permit for holding any “large-scale events” in a municipal park, whether the event involved a sit-in, a company picnic, or a crowd-producing rugby match. Thomas v. Chicago Park Dist., 534 U.S. 316, 318–22 (2002). The opinion is a curious one, in part because it was authored by Justice Scalia, who generally viewed generally applicable laws as properly subject to no special constitutional limits of any kind, even when particular applications of them “hit” speech activities.

  349. [349]. See supra Part IV.A.

  350. [350]. See, e.g., Daryl J. Levinson, Framing Transactions in Constitutional Law, 111 Yale L.J. 1311, 1313, 1318 (2002) (noting the inevitable need for courts to “slice [a] ceaseless and complex course of dealings into adjudicative transactions for the purpose of evaluating whether government has inflicted a constitutionally cognizable harm” and that “the problem of transactional framing is pervasive in constitutional law”).

  351. [351]. See supra Part V.D.1–2 (discussing analyses of Srinivasan and Dorf).

  352. [352]. See generally Kathleen M. Sullivan, The Justices of Rules and Standards, 106 Harv. L. Rev. 22 (1992).

  353. [353]. Oliver Wendell Holmes, Jr., The Common Law 1 (1881).

  354. [354]. See supra notes 52–53 and accompanying text.

  355. [355]. See supra notes 127–39 and accompanying text.

  356. [356]. See supra notes 23–24, 145–48 and accompanying text.

  357. [357]. See supra notes 316, 318–19 and accompanying text.

  358. [358]. See supra notes 218–342 and accompanying text.

  359. [359]. See, e.g., supra notes 251–64 and accompanying text (discussing the Court’s ruling in Albertini).

  360. [360]. See, e.g., Shaman, supra note 2, at 433 (noting Arcara’s effort to distinguish OBrien but also concluding that this distinction was “misguided”); Stone, Content-Neutral Restrictions, supra note 2, at 112 (questioning whether Arcara is fairly distinguishable from earlier incidental-burden cases).

  361. [361]. See supra notes 215–311 and accompanying text.

  362. [362]. See id.

  363. [363]. See supra notes 312–42 and accompanying text (discussing possible substantial-burden and wrongful-purpose-prophylaxis exceptions); Stone, Content-Neutral Restrictions, supra note 2, at 110 (raising the question whether limits noted in Arcara “exhaust the circumstances in which the Court will review a law that has only an incidental effect”); Williams, supra note 2, at 726–27 (proposing an exception to Arcara’s no-review rule under which states’ speech-related response to a generally-applicable-law violation must be “related” to that violation); see also Neutral Rules of General Applicability, supra note 2, at 1735 (reasoning that developments in other related constitutional fields may push forward efforts “for examining the most extreme burdens on First Amendment rights” in generally-applicable-law cases).

  364. [364]. See generally Schauer, Categories, supra note 2, at 307 (“The risk of misapplication of numerous subcategories leads us to eschew subcategories within the first amendment....”).

  365. [365]. See supra notes 149–64 and accompanying text. Of interest in this regard is the Court’s recent recognition of an additional exception to the free-exercise rule of Smith, the Court’s seminal effort to apply a bright-line-type no-review rule to generally applicable laws. See Hosanna–Tabor Evangelical Lutheran Church & Sch. v. E.E.O.C., 132 S. Ct. 694, 702–07 (2012) (precluding suits by ministers against churches based on generally applicable employment-discrimination laws).

  366. [366]. See Stone, Content-Neutral Restrictions, supra note 2, at 58 (noting the draw toward reviewing content-neutral laws when those laws have a “severe effect” on speech). See generally Frederick Schauer, The Tyranny of Choice and the Rulification of Standards, 14 J. Contemp. L. Issues 803 (2005) (discussing the breakdown of bright-line rules).

  367. [367]. Elliot Mincberg, A Look at Recent Supreme Court Decisions: Judicial Prior Restraint and the First Amendment, 44 Hastings L.J. 871, 872 (1993) (noting that Chief Justice Rehnquist “exemplified” a “‘statist’ view” of the judicial role); Kathleen M. Sullivan, Two Concepts of Freedom of Speech, 124 Harv. L. Rev. 143, 152 (2010) (suggesting that Chief Justice Rehnquist gravitated toward “a consistent statist position” at least in free-speech cases).

  368. [368]. See, e.g., Barnes v. Glen Theatre, Inc., 501 U.S. 560, 580 (1991) (Scalia, J., concurring) (“[W]e should avoid wherever possible ... analysis that requires judicial assessment of the ‘importance’ of government interests ....”); see also David Boling, Comment, The Jurisprudential Approach of Justice Antonin Scalia: Methodology over Result?, 44 Ark. L. Rev. 1143, 1197 (1991) (noting Justice Scalia’s “aversion to applying ‘balancing modes’ and ‘totality of the circumstances’ tests”). See generally Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. Chi. L. Rev. 1175, 1187 (1989) (arguing that standards should “be avoided where possible [and rules should] be extended as far as the nature of the question allows”).

  369. [369]. See Emp’t Div. v. Smith, 494 U.S. 872, 879 (1990) (relying on Justice Stevens’s earlier concurring opinion in United States v. Lee, 455 U.S. 252, 263 n.3 (1982)).

  370. [370]. See supra notes 200–09, 279–90 and accompanying text.

  371. [371]. See supra notes 303–11 and accompanying text.

  372. [372]. See, e.g., Boy Scouts of Am. v. Dale, 530 U.S. 640, 656 (2000) (invalidating application of an antidiscrimination law to require organization’s use of a gay scoutmaster based on freedom of expressive association). See also Stormans, Inc. v. Wiesman, 136 S. Ct. 2433, 2433–35 (2016) (Alito, J., joined by Roberts, C.J., and Thomas, J., dissenting from denial of certiorari) (arguing that the Court should consider whether a state regulation requiring pharmacies to provide emergency contraceptives violates the Free Exercise Clause); cf. Obergefell v. Hodges, 135 S. Ct. 2584, 2625 (2015) (Roberts, C.J., joined by Scalia, J., and Thomas, J., dissenting) (expressing concern that the majority’s decision might prevent persons sincerely opposed to same-sex marriage from exercising religious freedoms). For a detailed development of the Boy Scouts case—including a critique based on the notion that the Court failed to take proper account of its generally applicable character—see Rubenfeld, supra note 2, at 810–17.

  373. [373]. See Kagan, supra note 2.

  374. [374]. See supra note 2 (collecting key scholarly treatments).

  375. [375]. See, e.g., Vincent Blasi, The Checking Value in First Amendment Theory, 1977 Am. B. Found. Res. J. 521, 528, 538–44, 631–48 (arguing that the First Amendment serves as a democracy-fostering check on the government by the governed).

  376. [376]. See, e.g., Kagan, supra note 2, at 414, 443 (“First Amendment law, as developed by the Supreme Court over the past several decades, has as its primary, though unstated, object the discovery of improper governmental motives” and “rules ... devised to flush out illicit purpose ... constitute the foundation stones of First Amendment doctrine”); id. at 423, 427–37, 505–17; Rubenfeld, supra note 2, at 775 (focusing on the importance of wrongful motivation, including by noting that “an impermissible legislative purpose undoubtedly can render an otherwise valid law unconstitutional”); see also id. at 767–70, 775–78; Srinivasan, supra note 2, at 415–16 (arguing that “the dangerof a speech-suppressive administrative motive rather than the degree of speech-restrictive effect” properly explains “the Supreme Court’s entire approach to incidental restrictions of speech”).

  377. [377]. See, e.g., Dorf, supra note 2, at 1176, 1232–33, 1242–43 (suggesting that judicial review of laws that incidentally burden speech should focus on speech-burdening effects); Stone, Content-Neutral Restrictions, supra note 2, at 57–80 (endorsing judicial review of content-neutral laws even when not motivated by an improper purposes).

  378. [378]. See, e.g., Rubenfeld, supra note 2, at 786 (“The OBrien test runs into trouble ... because it presents itself as ... unmoored from an inquiry into legislative purpose.”).

  379. [379]. See supra notes 184–91 and accompanying text.

  380. [380]. See supra notes 178–82 and accompanying text.

  381. [381]. See Davis & Rosenberg, supra note 2, at 182; see also Kagan, supra note 2, at 457–59 (noting the importance of impact-related prophylactic rules to guard against the exercise of wrongful lawmaker purpose); Srinivasan, supra note 2, at 420 (describing OBrien as “fairly coherent and sensible” notwithstanding its lack of an overt focus on wrongful purpose).

  382. [382]. See, e.g., Kagan, supra note 2, at 415 (“Some aspects of First Amendment law resist explanation in terms of motive ....”); id. at 455 (noting relevance, including in assessing content-neutral laws, of an effects-driven “speaker-based perspective”). See generally Volokh, Speech as Conduct, supra note 2, at 1339 n.317 (“The Supreme Court has been notoriously reluctant ... to settle on any theory—self-government, the search for truth, self-expression, and so on—as being the sole foundation of First Amendment law.”). For one example of effect-centered reasoning, see NAACP v. Alabama ex rel. Patterson, 357 U.S. 449, 461 (1958) (“In the domain of these indispensable liberties, whether of speech, press, or association, the decisions of this Court recognize that abridgement of such rights, even though unintended, may inevitably follow from varied forms of governmental action.”).

  383. [383]. See supra note 92 and accompanying text.

  384. [384]. See Volokh, Speech as Conduct, supra note 2, at 1301. Given these considerations, speech-effect-minded thinkers should have little difficulty applying strict scrutiny to generally applicable laws that place direct-in-effect burdens on expression. After all, the practical impact of such laws is exactly the same as that of laws that directly regulate speech—that is, the speaker is silenced because (and only because) of speech-based effects on listeners. To be sure, Professor Volokh has observed that applying strict scrutiny to direct-in-effect-burden-imposing generally applicable laws might seem “in some tension” with purpose-centered theories. Id.; see, e.g., Kagan, supra note 2, at 461–62 (noting that generally applicable breach-of-the-peace statutes “preselect[] no particular ideas as posing a danger” and that the application of such statutes “not only to speech, but also to conduct posing a risk of disorder,” carries with it a “breadth [that] usually decreases ... the chance[s] of illicit purpose”). Perhaps, however, this perceived tension is overstated. Why? Partly because untangling bad purposes is a complex matter. For example, the legislature’s failure to build a speech-related exemption into a generally applicable breach-of-the-peace law may result, to some degree, from impatience with maverick speakers. Such inaction might also flow from systemic institutional failures to rightly weigh the rights of such persons and the value of their exercise of rights, including to the broader society. See Werhan, supra note 2, at 657 (noting that government officials “can be expected to overregulate expression”). In light of all this, a thoughtful, purpose-minded free-speech theorist might at least think twice before rejecting a principle of careful judicial review for laws that impose direct-in-effect burdens on speech—which is exactly how the Court has dealt with cases such as Cohen v. California.

  385. [385]. See Volokh, Speech as Conduct, supra note 2, at 1287­–91, 1301 (developing this point in detail, including with regard to the Court’s now-discredited treatment of the early Schenck and Debs cases).

  386. [386]. See, e.g., Bogen, supra note 2, at 248 (endorsing OBrien’s “[p]rophylactic standards” in part because they do not call on “the Court to use its most stringent test to review the incidental restrictions on speech”); id. at 205 (defending “OBrien ... because it assures both the government’s ability to accomplish its legitimate functions and the protection of speech ... from unnecessary restriction”); Shaman, supra note 2, at 462–63 (defending OBrien as sensibly steering a “middle course” when a case “combines elements of speech and action”).

  387. [387]. See, e.g., Tribe, supra note 2, § 12-7, at 825–32. To be sure, the drawing of the incidental-burden/doubly-incidental-burden distinction might be particularly suspect in a case such as Arcara itself, given the likelihood of a foreseeable impact of anti-lewdness laws on off-color bookstore operators. See Kagan, supra note 2, at 480 (noting that “hostility toward certain ideas about sexual mores—otherwise stated, the desire to maintain status quo ideas about sexuality free from challenge—[is] likely to color [an otherwise nonideological evaluation of resulting harms] or trump it entirely”). But this possibility does not change the fact most doubly-incidental-burden cases do not raise this problem, so that the Arcara rule is at least justifiable as a rule-of-thumb “administrative device” that sensibly operates “to cabin the circumstances in which incidental restrictions can raise first amendment questions.” Stone, Content-Neutral Restrictions, supra note 2, at 110.

  388. [388]. The Lovin’ Spoonful, She is Still a Mystery, on Everything Playing (Kama Sutra Records 1967).

  389. [389]. Notably, this bright-line-rule rhetoric has surfaced in areas of law untethered to First Amendment doctrine. See, e.g., Exxon Corp. v. Eagerton, 462 U.S. 176, 191–92 (1983) (finding the balancing approach applied in earlier Contract Clause cases entirely inapplicable to the challenged statute because it “imposed a generally applicable rule of conduct”).

  390. [390]. Benjamin N. Cardozo, Lecture IV, Adherence to Precedent, The Subconscious Element in the Judicial Process, Conclusion, delivered at Yale Law School (1921) in The Nature of the Judicial Process 166 (1921).


University Professor & Harmon W. Caldwell Chair in Constitutional Law, University of Georgia School of Law.

The author thanks Richard Fallon, Mark Tushnet, Sonja West, and Michael Coenen for valuable comments on drafts and Jared Batte, Grayson Chambers, Kate Ekstrand, and Ashley Klein for helpful research assistance.