Article

In Defense of Content Regulation

I.     Introduction

Since its 1972 decision in Police Department of Chicago v. Mosley, the central tenet of the Supreme Court’s free speech doctrine has been that “above all else, the First Amendment means that government has no power to restrict expression because of its message, its ideas, its subject matter, or its content.”1 In modern, doctrinal terms this means that content-based regulations are subject to “strict scrutiny,” while content-neutral laws receive only intermediate scrutiny. Moreover, this classification is essentially outcome determinative; in only one modern case has a majority of the Court unambiguously2 upheld a content-based law under strict scrutiny,3 while it has almost always sustained laws subject to intermediate scrutiny.4 This principle has been mitigated by the Court’s recognition of categories of unprotected speech, and low-value speech, where greater regulation is permitted. But in the realm of fully protected speech, the Supreme Court has held firmly to its doctrinal framework and its implications. Indeed, in recent years it has expanded the reach of its doctrine by extending full protection to speech previously thought to be at the margins of First Amendment protections.5

Looking at the Court’s decisions, then, one is left with the impression of a now well-established and stable free speech doctrine, providing robust protection to most expression. If one expands one’s gaze beyond the Supreme Court, however, to lower courts, this cozy vision collapses. The circuit courts are rife with divisions over questions such as the constitutionality of local ordinances regulating signs,6 and the constitutionality of bans on panhandling or other forms of solicitation.7 These courts disagree not just over the correct result in such cases, but also over the more basic issue of how to define content neutrality. Moreover, their disagreement reflects inconsistencies in the Supreme Court’s own statements regarding the meaning of content discrimination, as well as in areas of law—notably the “secondary effects” doctrine—where the Court has itself been unwilling to live with the full consequences of its doctrinal framework.

This Article contends that these developing fault lines are not merely a reflection of poor reasoning or unclear doctrine. Instead, they reflect a growing discontent with a basic assumption underlying current doctrine—that all fully-protected speech is equal and must be treated equally by regulators. Furthermore, this discontent is justified. The all-speech-is-equal principle is in deep tension with another basic premise of modern First Amendment jurisprudence—that the primary purpose of the First Amendment is to advance democratic self-governance, and therefore speech relevant to self-governance (notably political speech and speech on public issues) deserves the highest level of First Amendment protection. Implicit in that premise is a hierarchy of speech, which the rule against content discrimination simultaneously appears to condemn. The problem with modern First Amendment doctrine, then, is not one of confusion or lack of clarity, it is an existential one.

This Article takes the position that the reason why lower courts disagree about the definition of content discrimination, and why the Supreme Court itself has not been consistent on this question, is an unstated discomfort with the implications of the all-speech-is-equal premise. The truth is that this premise simply does not coincide with the instincts of most citizens and—importantly—most judges. As a result, when a law that regulates fully protected speech that seems less socially valuable than speech at the core of First Amendment’s protections is coupled with a strong, albeit likely not “compelling,” government reason to restrict the speech, judges regularly look to avoid labeling the law as content-based, even when it is clearly so.8

All of this suggests that it is time to rethink our hostility to all forms of content regulation, and to consider whether a more nuanced approach is required. This Article seeks to begin that process. By no means is this article in favor of abandoning all attention to content discrimination; but the blanket assumption that all distinctions among categories of protected speech are presumptively invalid must be abandoned. The Article identifies a specific situation—conditions of physical scarcity—where the all-speech-is-equal rule is clearly irrational. It then proceeds to a discussion of additional contexts such as sexually-oriented speech, panhandling, privacy, and commercial speech, where a blind application of principles that were developed in cases involving political or religious speech leads to noticeably problematic results, which in turn produces judicial evasion. The Article also considers whether in these areas as well an explicit toleration of content-based regulation might not make sense.

Part II discusses various classes of cases in which the lower courts, or in some instances the Supreme Court, have struggled with the definition of content discrimination. Part III identifies a number of doctrinal developments in recent years that have tended to extend full First Amendment protections to seemingly peripheral speech. It also notes how this trend towards proliferating protection places great pressure on the all-speech-is-equal premise. Part IV develops the now well-accepted principle that the First Amendment, above all else, is intended to advance democratic self-governance, and also how this principle tends to undermine the doctrinal assumption of speech equality. Part V explores a variety of contexts, notably conditions of physical scarcity, where the presumption of speech equality produces irrational or incongruous results. Finally, Part VI closes by identifying a few obvious concerns raised by the abandonment of a firm rule against content discrimination, and offers some preliminary thoughts about how clear doctrinal rules might alleviate those concerns.

II.     Discontent with Content

As noted above, the seemingly monolithic edifice of modern free-speech doctrine hides some very significant cracks, which in turn raise questions about the stability of that very edifice. In this Part, I will examine various specific contexts in which these cracks have been particularly present, and have led to serious judicial confusion and disagreement.

A.     Signs

The Supreme Court recently decided a case titled Reed v. Town of Gilbert.9 The case addressed the constitutionality of an ordinance adopted by the Town of Gilbert that restricted the placement of outdoor signs, both on private property and in public ways.10 The ordinance established a default rule that required a permit to display any outdoor sign, but also established a series of exemptions for 23 specific types of signs.11 Among the exempted signs were “Ideological Sign[s],” “Political Sign[s],” and “Temporary Directional Signs Relating to a Qualifying Event” (“Qualifying Event Signs”).12 Ideological Signs were defined as signs that conveyed a message for noncommercial purposes, but which were neither Political nor Qualifying Event signs.13 Political Signs were defined as temporary signs in support of a candidate, or otherwise “designed to influence the outcome of an election.”14 Finally, Qualifying Event Signs were temporary signs to provide directions to an event sponsored by a religious or other nonprofit group.15 Critically, however, the ordinance imposed very different restrictions on Qualifying Event Signs. Political Signs and Ideological Signs were permitted to be much larger than Qualifying Event Signs, and were permitted to remain in place for longer period of time.16 In particular, Qualifying Event Signs could only be erected 12 hours before the event at issue, and had to be removed within an hour following the event.17 Political Signs, meanwhile, could be erected 60 days before a primary election (though they had to be removed within 15 days after the general election), and Ideological Signs faced no time limit at all.18 Finally, the ordinance also limited the number of Qualifying Event Signs permitted on a piece of property, but did not so limit Political or Ideological Signs.19

The Gilbert sign ordinance was challenged by Clyde Reed, the pastor of a small Christian congregation named the Good News Community Church.20 Good News met in rented locations, including one elementary school in Gilbert and another school in a neighboring city.21 Reed and Good News had for years erected signs in Gilbert directing members of the community to their services, which they viewed as part of their religious obligation to evangelize.22 Gilbert’s ordinance (which Good News regularly violated) limited the number and size of their signs, and more onerously, effectively only permitted signs to be in place overnight. Since Good News’ services started at 9 a.m., the ordinance meant that legally their signs had to be erected after 9 p.m. the previous night.23 Reed’s lawsuit claimed the ordinance violated his and Good News’ First Amendment rights, because the law discriminated against particular speech (“Qualifying Event Signs”) on the basis of content.24 A complex sequence of litigation ensued, but the long and short of it is that the District Court, and two separate panels of the Ninth Circuit, consistently concluded that the ordinance was content-neutral and upheld the ordinance’s restrictions as reasonable time, place, and manner rules under intermediate scrutiny.25 The Supreme Court unanimously reversed the Ninth Circuit, holding that the ordinance was content-based, and therefore unconstitutional.26

The lower court decisions in Reed are peculiar. As Judge Watford’s dissent points out, on its face the Gilbert sign ordinance seemed to unambiguously favor certain categories of speech—political and ideological speech—over others—signs that promote and direct the public to events.27 As Judge Watford wrote, the only plausible explanation for the Gilbert ordinance was “Gilbert’s apparent determination that ‘ideological’ and ‘political’ speech is categorically more valuable, and therefore entitled to greater protection from regulation, than speech promoting events sponsored by non-profit organizations. That is precisely the value judgment that the First and Fourteenth Amendments forbid Gilbert to make.”28 Why exactly the first Ninth Circuit panel and the majority on the second panel rejected this seemingly overwhelming argument is unclear. At several points, both panels appear to equate content discrimination with laws motivated by a governmental purpose to suppress messages with which it disagrees.29 Thus the second panel said that the ordinance was content-neutral because “Gilbert did not adopt its regulation of speech because it disagreed with the message conveyed.”30 The first panel, however, had shied away from this language,31 instead emphasizing that the definition of “Qualifying Event Signs” turned on “‘who’ is speaking and ‘what event’ is occurring” rather than the sign’s content.32 The second panel similarly (and even more opaquely) stated that the Gilbert ordinance was content-neutral because “each classification and its restrictions [were] based on objective factors relevant to Gilbert’s creation of the specific exemption from the permit requirement.”33

In short, both Ninth Circuit opinions were as confused as they were illogical. As such, it came as no surprise that the Supreme Court reversed. Justice Thomas’s majority opinion (for six justices) began by emphasizing that a law can be content-based either if it “‘on its face’ draws distinctions based on the message a speaker conveys,”34or if the law “cannot be ‘justified without reference to the content of the regulated speech,’ or . . . [was] adopted by the government ‘because of disagreement with the message [the speech] conveys.’”35 In other words, once a law is determined to be content-based on its face, strict scrutiny will apply regardless of the government’s benign justifications for the law.36 Having reached this conclusion, the Court’s analysis became trivial, since as noted above and recognized by the Court, the Gilbert ordinance was obviously facially content-based. The sole basis upon which it differentially regulates signs was their message (as opposed to say their size or location).37 Finally, the Court held that the ordinance could not survive strict scrutiny because the Gilbert’s stated regulatory interests—preserving aesthetic appeal and traffic safety—had no discernable connection to the distinctions they had drawn.38

The majority also noted that the Ninth Circuit’s conclusion that the Gilbert ordinance was content-neutral was simply incorrect.39 Insofar as the Ninth Circuit equated content-discrimination with purpose, its reasoning conflicted with numerous decisions by the Supreme Court that reject this view and find content discrimination on the face of statutes, without regard to intent.40 If anything, the Reed majority actually understated the frequency with which the Court has previously reached this conclusion.41

Second, the Court emphasized that if the Ninth Circuit had instead equated content discrimination with hostility towards specific viewpoints, that too would be contrary to precedent, which makes it clear that content discrimination encompasses distinctions based on subject matter as well as viewpoint.42 Finally, the Court concluded that the Ninth Circuit’s alternative argument that the speaker and event, not content, triggered the definition of Qualifying Event Signs was equally flawed.43 While certainly the definition of a “Qualifying Event” depends on who speaks and what the event is, the definition of a “Sign” turns on what the sign says about that event.44 Moreover, the Court’s conclusion in this regard must be correct—a prohibition on signs advertising a protest is not a regulation of protests, it is a regulation of speech about protests.

It should also be noted that the Supreme Court was in fact only unanimous in its conclusion that the Gilbert ordinance was content-based. The concurring justices parted ways with the majority on what legal consequences should flow from that classification. In particular, they argued that the classification of the ordinance as content-based should not automatically result in the application of strict scrutiny.45 Instead, they argued, the existence of content discrimination should be treated as a “rule of thumb” rather than an automatic trigger for strict scrutiny;46 or alternatively, that the rule forbidding content discrimination should be administrated “with a dose of common sense, so as to leave standing laws that in no way implicate its intended function.”47 But none of the justices endorsed the idea—advanced by the Ninth and (as we shall see) other lower courts—that a law which is clearly on its face content-based might nonetheless be described otherwise.

If the Ninth Circuit’s Reed decisions were merely an example of a single court making a logical error, this case would seem insignificant, and certainly not worthy of certiorari. But they were not. Instead, as Reed’s Petition for Certiorari extensively argued, the Ninth Circuit decision widened an already deep, three-way circuit split over how to define content-neutrality in the context of sign ordinances.48 Several circuits, including seemingly the Ninth, defined content discrimination as a government motive to suppress particular ideas.49 Others decided the issue more in line with Supreme Court precedent, identifying instances of content discrimination on the face of the statute.50 And the Third Circuit had followed its own idiosyncratic multifactor test.51 To make matters worse, the Ninth Circuit did not even seem to be consistent, fluctuating between facial and purpose-based analyses depending on the context.52

Signs, then, had been a source of great doctrinal confusion in the appellate courts. Presumably, the Supreme Court’s Reed decision resolved this division and confusion. Reed’s clear implication is that the Ninth Circuit (and presumably others such as the Fourth) were simply, and inexplicably, confused. This conclusion, however, is unfair to those Circuits. The Court itself had been far from clear on this issue prior to Reed. While the Court had often eschewed motive inquiries, at other times it had suggested exactly the opposite. Thus in Ward v. Rock Against Racism, the Court said that “[t]he principal inquiry in determining content neutrality, in speech cases generally and in time, place, or manner cases in particular, is whether the government has adopted a regulation of speech because of disagreement with the message it conveys.”53 Admittedly, in Ward this statement was dictum, since the law at issue was facially content-neutral as well as lacking any invidious intent. But in Hill v. Colorado, the Court crafted this approach into a holding.54 In the course of upholding a statute regulating expressive activities near health care facilities—in practice abortion clinics—the Court held, after quoting Ward, that a restriction on “oral protest, education, or counseling” was content-neutral even though it clearly required an examination of the content of communication in order to determine if it fell afoul of the law.55 In support of its conclusion the Court cited to a string of examples where content is taken into account, without noting, however, that each of those examples involved unprotected or low-value speech.56 Justice Scalia, in dissent, objected to this approach, arguing that bad motive is an additional ground for finding a law content-based, not a substitute for facial analysis.57 The majority, however, did not respond to this argument, nor acknowledge the tension between its current approach and that of other cases where the Court had rejected bad intent the definition of content discrimination. This article will examine Hill in more detail below, but suffice it to say that the doctrinal confusion among lower courts in this area was built on inconsistent signals from the Supreme Court itself. And it is far from certain that the Court will adhere to the Reed approach when the restriction on speech involves a more complex issue than sign ordinances.

B.     Buskers and Panhandlers

Parallel to the jurisprudence of signs, an extensive body of case law has also developed in the Circuits regarding the regulation of solicitation. Most of these cases review restrictions on panhandling—meaning the noncommercial solicitation of money or goods for immediate receipt—though the Ninth Circuit’s contribution involves the regulation of “active solicitation” by street performers, or “buskers.”58 The solicitation cases reveal a similar confusion regarding the nature of content analysis.

A good starting point is the Seventh Circuit’s decisions evaluating the City of Springfield’s panhandling ordinance. In its first decision in the litigation, Norton v. City of Springfield, the court upheld locational restrictions adopted by the City of Springfield on panhandling, defined as oral requests for an immediate donation of money.59 The court, in an opinion by Judge Frank Easterbrook, acknowledged that the circuits are divided on whether such laws are content-based,60 but ultimately adopted the content-neutral analysis. The court upheld the law as valid, defining content discrimination as a restriction based on governmental motive.61 The following, somewhat remarkable statement sums up the court’s approach: “When the Supreme Court writes that rules regulating speech by content” must survive strict scrutiny, “it does not mean that all classification of speech is (effectively) forbidden. Government regularly distinguishes speech by subject-matter, and the Court does not express special concern.”62 Like the Supreme Court in Hill, the Seventh Circuit then defended this statement with numerous references to prior cases that all involved low-value speech.63 In a sharp dissent, Judge Manion criticized the majority’s analysis, as well as the analysis in a D.C. Circuit opinion that also equated content discrimination with intent.64 Judge Manion wrote that both courts’ failed to ask if the law was facially content-based and were therefore deficient in their analysis.65 Interestingly, the Seventh Circuit deferred its consideration of a petition for rehearing until after the Supreme Court decided Reed. After Reed was issued, the Seventh Circuit (with Judge Easterbrook again writing) reversed course and concluded that, because the Springfield ordinance distinguished between different messages on its face, it was—based on the analysis from Reed—content-based and therefore unconstitutional.66

The Seventh Circuit’s original approach in Norris can be contrasted to that of the en banc Ninth Circuit decision in Berger v. City of Seattle.67Berger did not involve panhandling, but instead street performers whose activities in Seattle Center, an urban park (and home of the Space Needle, inter alia), had been limited.68 Among the restrictions at issue was a prohibition on the “active solicitation” of donations.69 The rule permitted the performer to have a receptacle with a sign seeking donations, but essentially nothing else.70 Critically, the majority held that a law “is content-based if either the underlying purpose of the regulation is to suppress particular ideas . . . or if the regulation, by its very terms, singles out particular content for differential treatment.”71 Having adopted this view, the majority easily concluded that the active solicitation ban was content-based, given that “[i]t specifically restrict[ed] street performers from communicating a particular set of messages—requests for donations.”72 A dissenting opinion took issue with this analysis, concluding instead that the active solicitation ban was content-neutral because it was not badly motivated,73 and a concurring opinion agreed (though it found the ban content-based on other grounds).74

The Berger majority built on a previous Ninth Circuit decision, ACLU of Nevada v. City of Las Vegas, which did invalidate a panhandling restriction.75 The ACLU of Nevada court announced the same bifurcated content analysis that it later followed in Berger.76 The court then drew a distinction between regulations of acts of solicitation (meaning the actual exchange of money), which are content-neutral, and regulations of words that solicit money, which are not.77 Because Las Vegas barred all solicitation, not just immediate exchanges of money, it fell within the latter camp and was invalid.78

One last solicitation case worth mentioning is Clatterbuck v. City of Charlottesville.79 In Norton, both the majority and the dissent cited Clatterbuck as an example of a decision invalidating a panhandling prohibition as content-based.80 Interestingly, however, this characterization is not quite accurate. The district court in Clatterbuck had found Charlottesville’s anti-panhandling law to be content-neutral on its face.81 The Fourth Circuit reversed, however, because “[t]he Ordinance plainly distinguishe[d] between types of solicitations on its face.”82 It then remanded the case to determine if the law was motivated by a “censorial purpose.”83 If not, the court stated that the law should be found content-neutral.84 Interestingly, the district court on remand did find that the law was improperly motivated and therefore content-based, in contrast to decisions such as Norton.85

What these cases reveal is that, like sign ordinances, panhandling ordinances have imposed a dilemma upon the courts of appeals. The most straightforward approach to defining content regulation would seemingly doom almost all such laws. But, as discussed later, such laws are seen by many as both innocuous and valuable.86 If the Seventh Circuit was correct in its most recent Norton decision and Reed does resolve this dilemma, that would spell the end of panhandling ordinances. But the truth is that Reed cannot eliminate the underlying policy conundrum, and it is not at all clear that other courts will follow Norton’s lead in condemning panhandling ordinances essentially across the board based on Reed’s holding.

C.     Abortion Protestors

Another area where the judiciary—in this instance, the Supreme Court itself—has struggled with the definition and application of content neutrality is in the regulation of abortion protestors. The most salient example of this division can be found in the Court’s 2000 decision in Hill v. Colorado.87 In that case, the majority upheld a statute forbidding any person within 100 feet of the entrance of a health care facility, “to ‘knowingly approach’ within eight feet of another person, without that person’s consent, ‘for the purpose of passing a leaflet or handbill to, displaying a sign to, or engaging in oral protest, education, or counseling with such other person.’”88 The majority found the law content-neutral, despite its targeting of “oral protest, education, [and] counseling” because, according to the Court, this provision was merely directed at determining the speaker’s purpose, not at suppressing a particular viewpoint or subject matter.89 It also rejected the argument that because the law only targeted speech at specific locations, it was necessarily directed at specific speakers.90 Finally, and most tellingly, the Court endorsed as legitimate and content-neutral one of the underlying purposes of the statute: To protect “[t]he unwilling listener’s interest in avoiding unwanted communication.”91

In dissent, Justice Scalia, joined by Justice Thomas, took sharp exception to the majority’s reasoning. He pointed out that in previous cases, the Court had explicitly rejected as content-based any state interest in protecting listeners from unwanted or unwelcome communications, if the reason for the communication being unwanted was its message.92 He also pointed out, quite convincingly, that the targeting of protest, education and counseling seems obviously content-based on its face, since it permits some types of messages but not others.93 Finally, Justice Kennedy in his dissent made the equally obvious point that given the nature of this statute, the Colorado government surely knew that the probable effect was the restriction of anti-abortion speech and no other, thereby making it content-based in practice.94 Reading the dissent, it is very difficult to avoid the conclusion that, under the Court’s precedent, the Colorado statute was content-based, but the majority was unwilling to accept the full implications of that conclusion and so simply avoided it. In this sense, Hill is the necessary, and acknowledged, progenitor of the previously discussed appellate decisions that found seemingly content-based sign ordinances to be content-neutral.

In 2014, in McCullen v. Coakley, the Court returned to the topic of abortion protestors, bringing the Court’s struggles and divisions in this area to a head.95 In McCullen, the Court unanimously invalidated a Massachusetts statute creating a fixed buffer zone around abortion clinics by making it a crime to knowingly enter or remain on a “‘public way or sidewalk’ within 35 feet of the entrance or driveway” of an abortion clinic.96 Critically, however, a narrow five-justice majority (consisting of Chief Justice Roberts joined by the traditionally liberal justices) found the statute to be content-neutral.97 As in Hill, the majority rejected the view that the Massachusetts law was necessarily content-based because it restricted speech only near abortion clinics.98 It also concluded that the buffer zone statute was “justified without reference to the content of the regulated speech” because the statute’s purpose was to protect patient safety and unobstructed access to abortion clinics, not to silence speech.99 Importantly, however, the majority backed off from the suggestion in Hill that protecting patients from unwanted speech was a content-neutral justification, saying instead that “the Act would not be content neutral if it were concerned with undesirable effects that arise from ‘the direct impact of speech on its audience’ or ‘[l]isteners’ reactions to speech.’”100 In his concurring opinion, Justice Scalia (joined by Justices Kennedy and Thomas) argued that through this statement the majority “sub silentio (and perhaps inadvertently) overruled Hill.”101 Adhering to his views in Hill, Justice Scalia also argued that several aspects of the Massachusetts ordinance—that the statute applied only to speech outside abortion clinics, and that it applied to speech outside all abortion clinics, including those with no history of disruptive protests—suggested a censorial intent on the government’s part.102 The exemption in the statute for clinic employees, he argued, serves the same purpose by favoring speakers with predictably pro-abortion views.103

In short, the Court’s abortion protestor decisions demonstrate profound differences among the justices in how to identify content-based statutes.104 The majority opinion in Hill in particular seems to avoid fairly clear doctrinal rules because of discomfort with their consequences. Why this might be the case will be considered later, but for now it should be noted that it remains far from clear whether the Court will view Reed as resolving this tension, and/or overruling Hill (the tension between Reed and McCullen is less inescapable).

D.     Sexually Oriented Businesses and “Secondary Effects”

Another area of law where the Court has failed to consistently adhere to its doctrinal framework regarding content-neutrality (and where the justices have sharply disagreed with each other) is the evaluation of zoning regulations directed at sexually oriented businesses. The problematic nature of the Court’s analysis in these cases has been widely acknowledged in the commentary,105 and by a number of the justices;106 but the Court has failed to date to forthrightly address those contradictions.

The Court’s first encounter with zoning of sexually oriented businesses (“SOBs”) was in Young v. American Mini Theatres.107 In Young, a five-justice majority upheld a Detroit zoning regulation that placed restrictions on the location of adult theaters.108 Though acknowledging that the law distinguished between speech “on the basis of content,” the Court declined to subject the law to strict scrutiny because the law purportedly did not single out for suppression any specific “social, political, or philosophical message.”109 The majority then concluded that the law was a reasonable means to accomplish the city’s legitimate goal of “preserving the character of its neighborhoods.”110 A four-justice dissent sharply criticized the majority’s analysis as inconsistent with bedrock First Amendment principles.111

Ten years later, the Court returned to the issue in City of Renton v. Playtime Theatres.112 As in Young, the issue in Renton was the constitutionality of a zoning law that imposed special restrictions on the location of adult theaters.113 Once again, the Court upheld the ordinance, but this time via a different (albeit equally questionable) doctrinal move. The Renton majority held that despite appearing so on its face, the Renton zoning ordinance was not content-based at all.114 Rather, the law was content-neutral because it was “aimed not at the content of the films . . . but . . . at the secondary effects of such theaters on the surrounding community,” in particular, the increased crime and blight associated with SOBs.115 The majority therefore applied only intermediate scrutiny, and upheld the law.116 Again, a dissent sharply criticized this analysis as inconsistent with cases treating facial content discrimination as sufficient to trigger strict scrutiny,117 but again, to no avail.

Finally, the Court’s most recent decision on this subject is City of Los Angeles v. Alameda Books.118Alameda Books involved a challenge to a Los Angeles ordinance prohibiting SOBs from locating within 1,000 feet of one another (or within 500 feet of religious institutions, schools, and parks) or co-locating two SOBs in the same building.119 Alameda Books, which operated an adult bookstore and adult video arcade at the same location, challenged the ordinance. Once again, the Court upheld the ordinance, albeit with more dissension this time.120 A four-justice plurality simply invoked the Renton “secondary effects” test to justify its decision.121 Justice Kennedy’s opinion concurring in the judgment (and providing the crucial fifth vote for the result) acknowledged that the secondary effects test was “something of a fiction,” but nonetheless concluded that intermediate, not strict scrutiny was the appropriate analysis because zoning ordinances are different from other regulations of speech.122 Finally, four justices dissented, though even they seemed to accept that the secondary effects doctrine remained viable, so long as a city can empirically demonstrate the existence of secondary effects (which they concluded Los Angeles had not).123

The secondary effects doctrine is in obvious and direct tension with numerous Supreme Court cases, including most obviously Reed. Reed clearly states that a law is necessarily content-based if it, on its face, distinguishes between speech based on content, without further analysis required.124 The saving grace here is that the Supreme Court has never applied the secondary effects doctrine to any case not involving zoning of SOBs125 and though lower courts have on occasion done so, even that is a rare event.126 The damage done by the secondary effects doctrine is thus limited; but the doctrinal inconsistency remains, and is in need of explanation. Moreover, once one recognizes, as this article demonstrates, that the secondary effects doctrine is not an isolated and unique deviation from the general content-neutrality doctrine, some larger framework is needed to explain these deviations.

E.     Disclosure and Privacy

Finally, one area of law which is likely to be exceedingly important in coming years, and where the full implications of the content-neutrality doctrine seem to produce highly problematic results, is the constitutionality of privacy laws regulating the disclosure of personal data. As has been extensively noted in both academic commentary and the popular press, we live in the era of Big Data. Today, both the government and private firms such as Google and Facebook are able to amass huge quantities of information about the behavior and preferences of private citizens.127 The disclosure or use of this data raises obvious and serious privacy concerns, as highlighted most sharply by Edward Snowden’s revelations about government surveillance.128 With respect to government surveillance, the primary restraints are likely to be constitutional (through the Fourth Amendment) and political, which are both beyond the scope of this Article. With respect to the private collection and storage of data, however, there are growing calls for restrictions on the public disclosure of personal data, as illustrated by the European Union’s recent adoption of a “right to be forgotten.”129 But can such restrictions survive constitutional scrutiny?

There is an extensive academic commentary on this subject of course, including notable works by Neil Richards, defending privacy laws,130 and by Jane Bambauer, arguing that privacy restrictions generally constitute content-based restrictions on speech.131 Under current law, Professor Bambauer is almost certainly correct, as the Supreme Court itself has strongly suggested (albeit in dictum) in its recent Sorrell v. IMS Health Inc. decision.132 The issue in Sorrell was the constitutionality of a Vermont statute that prohibited pharmacies from selling “prescriber-identifying information”—i.e., information about the prescribing habits of specific doctors—to pharmaceutical companies for use in marketing prescription drugs to doctors.133 In response to the argument by Vermont that the sale of such information is not speech at all, the Court, in an opinion by Justice Kennedy, had this to say:

This Court has held that the creation and dissemination of information are speech within the meaning of the First Amendment. Facts, after all, are the beginning point for much of the speech that is most essential to advance human knowledge and to conduct human affairs. There is thus a strong argument that prescriber-identifying information is speech for First Amendment purposes.134

Ultimately, the Court avoided the broader issue of whether facts and data are speech because it concluded that the Vermont law’s targeting of marketing by pharmaceutical companies violated the commercial speech doctrine.135 But the implications of the majority’s language are clear enough.

Justice Kennedy is surely correct. There is an extensive body of case law treating the disclosure of information as speech,136 and common sense supports that conclusion. When the New York Times publishes the names of fallen soldiers, that is the disclosure of pure information, and indeed personal information, but no one would seriously suggest that it is not speech that is protected by the First Amendment. That remains true whether the publication is in print or electronically transmitted. Not only is data/‌information speech, but it logically follows that most privacy laws must be considered content-based restrictions on speech. This is because privacy laws do not typically regulate all data, but only specific types of data considered especially sensitive—e.g., an individual’s viewing or reading habits.137 In other words, privacy laws single out specific content for suppression, and so the doctrine would require that they be subject to strict scrutiny. Of course, some privacy laws, such as bank and healthcare secrecy laws, might survive strict scrutiny—though as noted above, the only modern free speech case in which an unambiguous majority of the Court upheld a law after applying strict scrutiny involved national security138—but many surely will not.

It seems likely that, as with signs and panhandlers, lower courts are going to be discomfited by the implications of the above analysis and seek ways to avoid them. One recent Seventh Circuit decision suggests that this is indeed the case. In Dahlstrom v. Sun-Times Media, LLC, the court faced a challenge to a provision of the Driver’s Privacy Protection Act (“DPPA”) prohibiting individuals from obtaining or disclosing “personal information” from motor vehicle records.139 The Sun-Times newspaper challenged this provision, as applied to its publication of the birth date, height, weight, hair color and eye color of several police officers who had participated in a lineup.140 The court concluded that the challenged DPPA provision was content-neutral, because: (1) it was motivated by privacy concerns, not hostility to particular messages; and (2) because the law did not suppress the information completely, it only suppressed it if obtained from a particular source.141 This analysis is, of course, doctrinally indefensible. The DPPA provision singles out specific content—particular types of sensitive details about individuals—for suppression. That is the very definition of content discrimination. Moreover, while the fact that this content is only suppressed if derived from a specific source—motor vehicle records—might be relevant to tailoring analysis, it cannot refute the conclusion that the law is content-based. After all, surely a provision banning a newspaper from publishing the names of fallen soldiers if the names are obtained from military press releases would be analyzed as a content-based regulation, just as would a flat prohibition on publishing such information. Dahlstrom then strongly suggests that as privacy cases begin to percolate in the lower courts, these courts will once again evade rather than confront the meaning and implications of current free speech doctrine. And again, there is little indication that Reed will change this fact, as evidenced by the fact that the Supreme Court denied certiorari in Dahlstrom in the aftermath of Reed.142

III.     The Proliferation of Speech Protections

The discussion in the previous part demonstrates that in a number of areas of First Amendment law, courts are struggling with the meaning of content-neutrality, and are regularly evading the implications of the Court’s well-established doctrine that was reaffirmed by Reed. In this part, I discuss a different development in the Court’s free speech jurisprudence that is likely, over time, to exacerbate these problems.

During the 1970s and 1980s, when the Court was building the edifice of modern free speech law, including notably the prohibition on content discrimination, it also strongly embraced the view that not all speech is equal. Thus, in its leading cases extending protection to commercial speech, the Court also made clear that such speech was accorded less protection than political and other fully protected speech, and could be subject to even content-based regulation at times.143 Similarly, cases from this era assumed that false statements of fact were valueless, and so intentional falsehoods could be suppressed entirely.144 At the same time, the Court reaffirmed its view that “obscenity” was not constitutionally protected, and notably expanded its definition of what constitutes obscene speech.145 The Court also suggested in the cases involving SOBs,146 and elsewhere,147 that sexual speech might be subject to greater regulatory restraints than other forms of speech. Finally, using “categorical balancing,” under which the Court denies constitutional protection to entire categories of speech if it concludes that harm caused by speech of that sort “overwhelmingly outweighs” any social value the speech possess, the Court entirely rejected protecting certain forms of especially harmful speech, notably child pornography.148

These developments, in combination, appeared to adopt a view that many categories of speech are of “low value” based on First Amendment values, and so can be subject to greater regulation than high value speech—including with content-based restrictions.149 This approach tended to alleviate the effects of the content-neutrality doctrine, restricting its application to situations where the regulated speech was highly valuable, and the government’s reasons for limiting the speech suspect. Recent developments, however, seriously undermine this accommodation.

The Court’s retreat from broad assumptions about “low-value” speech is most obvious in the area of sexually oriented speech. It is true, as noted above, that the Court adheres to the “secondary effects” doctrine in evaluating zoning of SOBs.150 However, elsewhere the Court has extended full protection to non-obscene sexually oriented speech, including speech that is frankly pornography. For example, in United States v. Playboy Entertainment Group, Inc., the majority applied strict scrutiny to, and struck down, a statute requiring cable channels that were “primarily dedicated to sexually-oriented programming”—i.e., porn—to scramble their signals.151 Similarly, in Ashcroft v. ACLU, the Court used strict scrutiny to invalidate a statute imposing screening obligations on commercial websites containing information that is obscene as to minors.152 And finally, in Ashcroft v. Free Speech Coalition the Court invalidated a prohibition on “virtual child pornography”—sexually explicit images that appear to depict minors, but do not actually involve minors.153 These cases in combination appear to establish a simple principle: Non-obscene, sexually explicit speech receives full First Amendment protection, and may not be regulated based on its content (unless strict scrutiny is satisfied, which it never has been).

Another arguably more important development has been the death of categorical balancing, and thus the Court’s effective renunciation of the power to identify new categories of “low-value” speech. The key case is United States v. Stevens, decided in 2010.154 In Stevens, the Court struck down a statute criminalizing the sale or possession of depictions of animal cruelty.155 Crucially, the Court rejected the argument that depictions of animal cruelty should be treated as “low value,” and should therefore be classified as unprotected speech under categorical balancing.156 Indeed, the Court rejected the idea of categorical balancing altogether, holding instead that unprotected categories of speech must be rooted in history, not identified using free-form balancing.157 The next year, in Brown v. Entertainment Merchants Association, the Court reaffirmed its rejection of categorical balancing, striking down as a result a California statute restricting the sale to minors of violent video games.158 The sum of Stevens and Brown appears to be to freeze into place the existing list of low-value speech categories, absent the (unlikely) discovery of a historically unprotected category which somehow has escaped notice until now.

In addition to abandoning categorical balancing, Stevens and Brown also extend the coverage of the First Amendment in three other significant ways. First, they make clear that violent speech receives full First Amendment protection, no matter the grotesqueness of the violence.159 In other words, there is no obscenity-equivalent for violence. Indeed, in Brown the Court explicitly rejected the view that violence should be treated as lower value even when directed at minors (and so a fortiori for adults), concluding that there is no historical basis for such an approach.160 Second, Brown makes it clear that differences between media and forms of speech—in particular, the interactive nature of video games—do not justify differential constitutional treatment.161 Finally, also in Brown, the Court made it clear that speech directed solely at children normally receives full First Amendment protection—a result in some serious tension with earlier cases recognizing a compelling governmental interest in protecting children from the harm caused by non-obscene sexually oriented speech.162

Another area where the Court has abandoned its earlier, more restrictive views on the scope of First Amendment coverage concerns factual falsehoods. As previously noted, the Court’s early libel jurisprudence took the position that factual falsehoods lacked any constitutional value, and therefore intentional falsehoods could be freely punished (unintentional falsehoods did in fact receive constitutional protection, but only as a means to avoid chilling speech, not as an end in itself).163 In 2012, however, in United States v. Alvarez, the Court rejected this view.164 In Alvarez, the Court was faced with a challenge to the Stolen Valor Act, which criminalized false claims about receipt of military decorations or medals.165 Alvarez undoubtedly made such a false claim, and undoubtedly did so intentionally, but the Court nonetheless reversed his conviction.166 The plurality opinion (by Justice Kennedy, joined by three others) appeared to extend full First Amendment protection to knowing falsehoods, by applying strict scrutiny to the Act.167 Justice Breyer (joined by Justice Kagan) adopted a more narrow view, extending partial but apparently not complete protection to knowing falsehoods.168 Nonetheless, six justices unqualifiedly rejected previous statements rejecting protection for knowing falsehoods.

Finally, another area where the First Amendment appears to be expanding is commercial speech. As noted earlier, early commercial speech cases only extended partial protection to such speech, applying intermediate scrutiny to even content-based regulations of speech.169 More recently, however, the Court has applied a much more stringent level of review, and almost without fail struck down regulations of commercial speech.170 Moreover, in the recent Sorrell case discussed earlier, Justice Kennedy’s majority opinion strongly suggested that even though the Vermont statute protecting prescriber-identifying information was a regulation of commercial speech, it should be subject to “heightened judicial scrutiny.”171 While the Court ultimately backed off these suggestions and applied intermediate scrutiny172—albeit, the ultra-strict modern variety—the implications for the future are obvious.

As this discussion shows, in recent years First Amendment protections have spread far and wide. This may seem to be an entirely benign development, but it is in deep tension with continuing adherence to a strict insistence on content neutrality. The reason is simple enough—given that full First Amendment protection is extended to speech that most citizens, legislators, and judges are likely to view as of limited value, and peripheral to the purposes of the First Amendment (which will be further explored in the next Part), the pressure for regulation is likely to grow. So, concomitantly, will the pressure on the judiciary to permit such regulation. Moreover, for reasons discussed further in Part IV, there will also be increasingly frequent instances when governments will wish to prefer speech seen as more valuable, and more central to the First Amendment’s purposes, over peripheral speech. For example, cities may wish to grant preferential treatment to non-commercial over commercial speech.173 If, however, commercial speech is fully protected, then such preferentialism constitutes content discrimination, and is presumptively invalid. For reasons that we will now explore, however, such a flat prohibition on preferring high-value over low-value speech is neither theoretically required, nor practically possible.

IV.     Democracy and Speech Equality

Over the years, scholars and judges have identified three principle theories explaining why the First Amendment protects freedom of expression. The first is that free speech is necessary to advance the search for truth. This idea derived from the writings of John Stuart Mill,174 as well as the “marketplace of ideas” metaphor derived from Justice Holmes seminal dissent in Abrams v. United States.175 The second, associated mainly with scholars such as Thomas Emerson and Edwin Baker, is that speech is an essential component of individual autonomy and self-fulfillment.176 The third, rooted in Justice Brandeis’s also-seminal opinion in Whitney v. California,177 as well as the scholarship of Alexander Meiklejohn,178 is that the First Amendment is intended to enable democratic self-governance and participatory politics. In other words, free speech is about democracy.

Each of these approaches has of course had its adherents over the years. In recent decades, however, a fairly broad (albeit not universal) consensus has seemingly emerged. The consensus holds that the primary, though not necessarily the only, purpose of the First Amendment, both as a matter of original intent and as a matter of modern theory, is the third theory, to enable self-governance. Leading scholars as diverse as Robert Bork and Cass Sunstein have endorsed this position.179 In recent years, Robert Post and Jim Weinstein have also ably defended it.180 And finally, as I have argued elsewhere, the self-governance theory, unlike the other theories, has the distinct advantage of explaining the free speech clause in its context, accompanied as it is by the Press, Assembly, and Petition Clauses.181

It is not only scholars who have endorsed the self-governance understanding of the First Amendment. If one looks at its recent decisions, the Supreme Court has repeatedly reaffirmed the view that the First Amendment protects, above all else, speech relevant to democratic self-governance. Thus in its famous/infamous Citizens United decision, the Court commented that “[s]peech is an essential mechanism of democracy, for it is the means to hold officials accountable to the people.”182 Similarly, in Snyder v. Phelps (discussed below), the Court recognized that “[s]peech on matters of public concern . . . is at the heart of the First Amendment’s protection.”183 Indeed, as far back as its seminal decision in New York Times v. Sullivan, the Court recognized that the first and foremost purpose of the First Amendment was to protect political debate, in order to effectuate responsive, democratic government.184 Other, similar statements in recent cases are too numerous to count.185

Nor is the principle that speech relevant to democracy occupies a central role in free speech theory a merely theoretical one—the Court has repeatedly operationalized this idea in holdings. For example, in Snyder v. Phelps the Court was faced with a claim that the First Amendment prohibited a grieving father from recovering damages for Intentional Infliction of Emotional Distress (“IIED”) against individuals who had staged a grotesque and hurtful demonstration near the funeral of his son, who had been killed in action in Iraq.186 Critical to the Court’s holding in favor of the demonstrators was its conclusion that the subject matter of their speech—the treatment of homosexuality in the United States and in the military—was “a matter of public concern.”187 In other words, the Court held that the speech was immune from civil liability because, and only because, it was relevant to democratic self-governance.188 The strong, negative implication was that if the speech had been purely private, liability would have been permitted.189

The Court’s approach in Snyder to IIED claims draws from earlier jurisprudence dealing with privacy and the disclosure of private facts. In a series of cases beginning with Cox Broadcasting Corp. v. Cohn, the Court was faced with damages claims brought by individuals based on the disclosure of private facts about those individuals by the press.190 (Cox and one other case involved the names of rape victims,191 while other cases involved disclosure of the names of juvenile defendants.192) In each of these cases the Court rejected press liability.193 It did not, however, adopt a blanket rule of press immunity. Instead, the Court emphasized its conclusion that in each of the cases, the information disclosed was on a matter of public concern, and therefore highly protected.194 Presumably the disclosure of purely private gossip might lead to a different result.

The Court has taken a similar approach in the area of libel. In early cases, the Court held that the Constitution provided substantial protection to allegedly libelous statements made about public officials and public figures.195 Later, the Court extended protection, albeit not quite as stringent protection, to libel even about private figures, so long as the subject matter of the speech was a matter of public concern.196 When faced with a claim brought by a private figure where the subject matter of the speech was purely private—i.e., it was not even arguably relevant to self-governance—the Court denied all constitutional protection.197

Finally, the distinction between democratic and private speech also appears in the law regarding the free-speech rights of public employees. In its foundational decision in Pickering v. Board of Education, the Court held that public employees do possess free speech rights, and have some immunity from adverse employment actions based on their speech.198 In addition, Pickering establishes a balancing test for such claims.199 Critically, however, the Court subsequently made it clear that this rule will typically only be applied to speech on matters of public concern.200 If the speech that provoked the discipline was on a purely private matter, no First Amendment scrutiny, “absent the most unusual circumstances,” is required.201

The lesson from these cases and bodies of law is clear—not all speech is equal. The Court has regularly protected speech relevant to democratic self-governance in situations where purely private speech would not be protected.202 These cases thus establish that speech relevant to democratic self-governance is indeed of central First Amendment importance, and more importantly, that such speech sometimes receives greater constitutional protection than other forms of speech. In other words, the Court has created a First Amendment hierarchy based on the conclusion that certain types of speech, defined by subject matter, are more valuable than other types of speech.

The contradiction should now be clear. Recall Judge Watford’s dissent in Reed II, in which he criticized the Town of Gilbert because it determined “that ‘ideological’ and ‘political’ speech is categorically more valuable, and therefore entitled to greater protection from regulation, than speech promoting events sponsored by non-profit organizations. That is precisely the value judgment that the First and Fourteenth Amendments forbid Gilbert to make.”203 Judge Watford, quite reasonably, sees this as the bedrock principle driving the presumption against content-based regulations. Yet in the IIED, privacy, libel, and government employee contexts, this is precisely the “value judgment” that the Supreme Court has made based on its understanding of the purposes of the First Amendment. The obvious question that arises is this: If the Court is permitted to make such judgments and prefer speech relevant to self-governance over other forms of speech because of perceived First Amendment values, might there be circumstances when other government actors, for similar reasons, can do the same? It is to that question that we now turn.

V.     Scarcity and Other Justifications for Content Regulation

This Part explores specific circumstances in which, I argue, we should be more tolerant of distinctions based on the content of speech than current doctrine would seem to permit. I emphasize several caveats, however, at the outset. First, my argument should not be understood as calling for a wholesale abandonment of the presumption against content discrimination. To the contrary, as my discussion should make clear, in most “normal” situations, the presumption may be entirely appropriate. Second, the corollary of that conclusion is that tolerance of content preferences should be limited to certain distinct, objectively definable circumstances where the general presumption against content discrimination does not make sense. Finally, and most importantly, a tolerance for content distinctions does not mean a tolerance for any and all content distinctions. A preference for some speech over other speech must be rooted in First Amendment theory, including notably the high value of speech relevant to self-governance. On this view, for example, the preference for commercial over political speech that the Court upheld in Lehman v. City of Shaker Heights204 is utterly perverse, and cannot be justified on any theory. Careful judicial attention to content-based laws will still be necessary, and not all distinctions will be upheld. As will hopefully become obvious, the argument here is not to fully abandon current doctrine, or even careful scrutiny of content distinctions. It is instead an argument for nuance.

A.     Scarcity

As Owen Fiss memorably commented, traditional First Amendment law developed as a means to protect the “street corner speaker” from persecution by the state.205 When Fiss coined that phrase, he meant to contrast such problems with the very different ones raised by regulation of large media entities, or of money in politics. There is, however, another important distinction that can be drawn here. Most of the classic First Amendment cases involve the state acting as censor and regulator, generally through criminal prosecution of speakers,206 but also occasionally through the imposition of civil liability based on the allegedly harmful content of speech.207 Government, however, does not only operate in a regulatory capacity. It also has a managerial role, primarily as landowner and employer, but also sometimes regulating the private use of resources for managerial, as opposed to censorial, reasons. When the government operates in a managerial capacity, I posit, issues might arise that simply have no analog to situations where the government regulates to prevent harm.

Of course, current First Amendment doctrine is hardly silent regarding restrictions on the government as manager. As discussed earlier, there is a well-developed body of law governing the speech of public employees.208 More fundamentally, the public forum doctrine has as its very purpose the exposition of rules governing the power of the state to restrict speech on government property.209 Under current law, however, if a piece of government property is classified as either a traditional or designated public forum, the rules regarding the government’s power to regulate speech are the same as those governing censorial regulation: Content-based rules are presumptively unconstitutional, while content-neutral rules are subject to a more forgiving intermediate scrutiny standard.210 The assumption appears to be that, in such forums, the government has no particular need, or right, to regulate based on content.

That assumption, I would argue, is false. The reason is simple—when the government manages speech on its own property, it faces a constraint, one of scarcity, which is simply irrelevant to censorship of allegedly harmful private speech. Property is almost by definition a limited resource.211 The consequence of this is that necessarily, the amount of speech, and number of speakers, on any particular piece of property will have to be restricted. The traditional response to this insight is that content-neutral rules such as time, place, or manner regulations212 or permitting schemes213 should suffice to allocate scarce space. However, the facts of Reed, and other cases that involve the regulation of signs on public forums such as medians and sidewalks,214 suggest that this may not be true. It may be, instead, that a strictly content-neutral allocation of scarce resources is profoundly irrational.

The point here is a simple one. Basic economics tells us that scarce resources should be allocated to their highest value use. That is in fact the whole function of markets—to allocate scarce resources to their highest value use, value being defined by the level of demand driven by purchasers’ ability and willingness to pay. Consider now the problem of a city trying to decide how to allocate limited space on its sidewalks or medians to those wishing to erect signs. Note in this regard that space for signs is limited both because it is physically limited, and because an excess of signs might well pose traffic dangers. And, of course, too many signs might just be ugly. Regardless, the key point is that the opportunities to speak are scarce. How then, should those scarce speech opportunities be allocated? One possibility is first-come, first-served, with later-erected signs rejected or removed once a numerical limit within any particular area is reached. There is, however, no reason to think that early sign-erectors truly represent the highest value use of that property.215 If we wished to mimic the market, then a city might auction off the right to erect signs to the highest bidder. That would allocate the scarce resource to the highest value use, as measured by the market. But is a market-based measure of value, which favors ability and willingness to pay, truly the most rational allocation of speech rights? Such a system is likely to favor commercial messages and wealthy speakers. It is hard to imagine why First Amendment law should favor, or require, such a result.216 The First Amendment, after all, does not enact either Mr. Herbert Spencer’s Social Statics,217 or Adam Smith’s Wealth of Nations.218 Surely First Amendment doctrine should reflect First Amendment values rather than market values or some other measure of equity.

Now consider the solution adopted by the Town of Gilbert.219 Recall that Gilbert generally prohibited posting signs on public property without a permit, but permitted certain forms of signs the Town deemed important. Even within this group, however, the Town drew distinctions. While the Town permitted directional signs of the sort Pastor Reed wanted to erect, it subjected them to greater restrictions than political and ideological signs. Those distinctions, which as the Court correctly held are surely content-based, were the target of Reed’s lawsuit. Doctrinally, therefore, Reed should have won (as he of course did), since the Town was engaged in blatant preferentialism based on content in a traditional public forum. But was it really impermissible, as Judge Watford argued in his Reed II dissent and as the Court’s decision suggests, for the Town to conclude that political and ideological signs represent the highest value use of a scarce resource—speech opportunities on public property? Surely such speech, based on the democratic self-governance reading of the First Amendment, is the highest value.

A similar issue arose in City of Cincinnati v. Discovery Network, Inc.220 Cincinnati adopted an ordinance prohibiting news racks containing commercial handbills (advertising) on public property, but permitting news racks containing newspapers.221 The City’s justification was that any reduction in the number of news racks on sidewalks would advance its interests in safety and aesthetics, and so it made sense to target commercial speech since it was less valuable than newspapers.222 The Court disagreed, apparently based on its conclusion that the distinction between commercial and noncommercial speech was not related to the City’s aesthetic interests.223 But this is a non sequitur. While it is true that commercial news racks are no uglier than ones containing newspapers, Cincinnati was surely correct that any reduction in the number of newsstands improves safety and aesthetics. Given that, targeting commercial news racks might well be the “least cost” way to achieve the city’s objectives, since commercial news racks’ social value might well be lower than that of newspaper news racks because the speech they distribute contributes less to the goals of the First Amendment. After all, from the perspective of democratic self-governance, surely commercial advertising sits lower on the First Amendment hierarchy than newspapers? Indeed, political speech, ideological speech, and newspapers are precisely the types of speech that the Court has favored most in its “matters of public concern” jurisprudence.224 If the Court may do this, then the obvious question arises, why not the City of Cincinnati or the Town of Gilbert?

My basic point is simple—if we acknowledge that there is a hierarchy of speech under the First Amendment, as the Court and most commentators appear to do, and if we acknowledge the problem of scarcity, then at first cut, it seems entirely irrational to forbid the government from allocating scarce speech opportunities to the highest value speech, as measured by First Amendment values. There may be prophylaxis or slippery-slope arguments to the contrary, but they are second-order concerns, and the burden of proof should be on those wishing to prevent the facially most rational outcome.

Indeed, the Supreme Court’s own jurisprudence appears to implicitly recognize this set of concerns, through its rules regarding nonpublic, or limited public forums (the Court appears to use these terms interchangeably).225 Under the Court’s modern jurisprudence, in such forums the government may not engage in viewpoint discrimination, but it may impose other content-based restrictions on speech such as subject-matter restrictions, and it may impose restrictions based on the identity of speakers, so long as those rules are reasonable.226 For example, in Rosenberger v. Rector & Visitors of University of Virginia, the Court considered a university program that providing funding for student publications.227 The Court classified the program as a limited public forum, and therefore upheld the limitation of funding to student groups (though it struck down the exclusion of religious publications as invalid viewpoint discrimination).228 Similarly, in Perry Education Association v. Perry Local Educators’ Association, the Court upheld a rule permitting a certified teachers’ union as well as certain private nonprofit groups such as the Cub Scouts, but not a rival teachers’ union, to use an interschool mail system.229 Again, the Court’s reasoning was that because the system was at most a limited forum, reasonable restrictions were permitted so long as they were viewpoint neutral.230 Decisions like Rosenberger and Perry, it should be recognized, are fundamentally cases about scarcity. The reason funding in Rosenberger was limited to student groups was, of course, that money is not unlimited—i.e., it is a scarce resource. So too is the capacity of an interschool mail system; if all groups were permitted access, the system would have to be expanded at substantial cost, so the school district limited access to messages on subjects—official school business and activities of interest to students—which are the highest value use of the system. The insight I offer is that scarcity is not a problem restricted to limited public forums, it can also be present in traditional and designated ones.

Another example of scarcity is public libraries. The Supreme Court has famously struggled with how to assess decisions by public officials regarding the addition or removal of materials in public libraries.231 Libraries, of course, suffer from the problem of scarcity—of shelf space, and of funds—which means choices must be made regarding what to stock.232 One seemingly “neutral” method of choice would be to prefer popular materials; but that is not a truly content-neutral approach, since the views of the public are themselves shaped by the content of books and other reading materials. And in any event, it seems utterly mad to completely ignore content in choosing what books to purchase, and which to forgo—after all, not all literature is created equal. Again, this is an area where content discrimination can and must be, and in fact is, tolerated. I would suggest, however, that First Amendment theory can provide insights into what forms of content discrimination should be permitted.

Indeed, the problem of scarcity does not arise only in the context of managing government property. Consider in this regard the Turner Broadcasting litigation.233 In these cases, the Court upheld the “must-carry” rules issue by the Federal Communications Commission (“FCC”) against a First Amendment challenge.234 The rules required cable television operators to dedicate up to one-third of their channel capacity to carry, free of charge, the signals of local over-the-air broadcast television stations.235 The Court found the rules to be content-neutral, 236 and ultimately concluded they survived intermediate scrutiny.237 Importantly, in the first Turner case, four justices dissented from the finding of content-neutrality.238 Justice O’Connor, writing for the dissenters, noted that the reasons why Congress favored local broadcast stations was because they carried programs about local events, including notably local news, and because they provided a diversity of viewpoints.239 Both of these, O’Connor convincingly argued, are at heart content-based justifications, making the must-carry rules themselves content-based.240 The majority rejected this conclusion, but without a convincing rebuttal.

The Turner cases are about scarcity. The number of channels on any cable system is limited, especially in the pre-digital days when the typical capacity was just 30 or so channels. Congress and the FCC faced the question of what the highest value use of this limited capacity was. Congress concluded that dedicating at least some channels to broadcast signals is a higher-value use than adding yet more cable networks, because local broadcast channels advance viewpoint diversity, and ensure that all citizens (including ones who cannot afford cable) continue to have access to television programming, including news. These are policies rooted in the First Amendment’s self-governance goals, since they promote the creation of an educated and informed citizenry. Therefore, the FCC’s choice can be defended as a rational response to scarcity. Note, that the scarce resource in Turner was not public, it was privately owned; however, that did not stop the Court from upholding the must-carry rules. The Court’s conclusion that the must-carry rules were content-neutral was undoubtedly a stretch, but it was driven by the doctrine that condemns essentially all content-based preferences. Turner is better understood, however, to uphold a benign content-based preference in the context of scarcity.

The problem of private scarcity is broader than Turner or cable television. Indeed, in the Reed case and in other litigation over sign ordinances, private scarcity is at play. The ordinance in Reed did not just regulate signs on public property (though it did impose special restrictions on such signs), but it also restricted purely private signs. Why? The standard reason for limiting private signs is aesthetic concerns, and traffic safety concerns. These policies, however, can be rephrased as problems of scarcity. Basically, because of aesthetic and safety goals, there are a limited number of signs that can be aesthetically or safely displayed in any particular location. This means that even private signage space is scarce. As in Turner, the application of a sign ordinance to private signs can be understood as the legislature choosing to allocate this scarce resource to its highest value use as judged by First Amendment standards. Of course, that preference must be viewpoint-neutral, but not necessarily content-neutral.241

Scarcity, then, represents probably the strongest example of a problem that is best solved by countenancing some content-based governmental preferences, so long as those preferences further First Amendment values. There are undoubtedly concerns, and counter-arguments, raised by this assertion, which will be addressed in the next part. For now, the point is simply that as a preliminary matter, it seems perfectly rational to permit scarce speech resources to be allocated to their highest social-value use, reflecting well-established First Amendment policies.

B.     Discomfort and Emotional Harm

One of the most fundamental principles of First Amendment law is that speech may not be silenced simply because it causes discomfort or offense to those exposed to it. That principle drives the result in classic cases such as Cantwell v. Connecticut, where the Court reversed the conviction of a Jehovah’s Witness who had been playing a record on the street attacking the Catholic Church.242 That same principal also motivated the Court in Terminiello v. City of Chicago, where the conviction of a far-right speaker who described his opponents as “slimy scum,” “snakes,” and “bedbugs” was reversed.243 More recently, it was the basis for the Court’s conclusion in Cohen v. California, that the defendant Paul Robert Cohen could not be punished for walking around the Los Angeles Municipal courthouse wearing a jacket with “Fuck the Draft” written on the back, protesting the Vietnam War.244 Moreover, this principle must be correct if the First Amendment is to accomplish its primary goal of ensuring that speech is not suppressed simply because the message it conveys is unpopular or out of the mainstream. After all, many minority positions, such as support for Communism or, for that matter, for racial equality, have at times in our history been considered deeply offensive by substantial parts of society. To permit a Heckler’s Veto is to permit the radical truncation of political discourse.

The rule that offense or discomfort alone cannot justify suppression of speech is thus an essential tool in accomplishing the First Amendment’s general goal of ensuring that the government does not suppress speech that it (or the majority of citizens) dislikes. But is this principle a universal one? Consider the panhandler and busker cases discussed earlier.245 Why is it that cities seek to restrict solicitations of money in public places? Some might argue, as Justice Kennedy did in his separate opinion in International Society for Krishna Consciousness v. Lee, Inc., that the purpose is to prevent the disruption, coercion, and fraud associated with immediate exchange of money.246 Indeed, this reasoning led Kennedy to label a solicitation ban as a regulation of conduct, and therefore permissible.247 This, however, is not entirely convincing. The laws at issue in cases like Berger and Norton do not forbid the exchange of funds, they forbid oral requests for funds—indeed, in Berger, the Seattle restriction on street performers explicitly permitted passive solicitation (through a written sign) and receipt of funds.248 There can be little doubt that at least part of the driving force behind solicitation bans is the discomfort begging causes to the targets of panhandling. It is surely no coincidence that such rules tend to target solicitation in tourist-friendly areas such as Seattle’s “Seattle Center,”249 Springfield’s “downtown historic district,”250 the Fremont Street Experience in Las Vegas,251 and the Downtown Mall in Charlottesville.252 Moreover, these rules do not typically ban only aggressive panhandling, they ban all solicitation of funds.253

If we accept, however, that at least part of the motivation for solicitation bans is the discomfort caused to listeners, certain doctrinal consequences flow. First, there can be no doubt that these laws are content-based (though in truth, there should be no doubt in any case, given that the laws suppress a specific message: Please give me money).254 The discomfort caused by solicitation is not produced just by the manner of the speaker (or else the laws would target only those types of solicitation), it is also a product of the message itself. The solicitation acts as a reminder that others are in need, thereby inducing guilt—the last emotion that a tourist wants to feel, and that people who want tourists’ money want them to feel. Second, under the reasoning of cases like Cantwell and Cohen, it would seem that such laws must also be flatly unconstitutional, since as noted, these cases establish that discomfort associated with a message cannot be grounds to suppress the message.255

The question we should ask, however, is whether this particular doctrinal result makes any sense. Certainly, Cantwell and Terminiello must be correct that offense from political, religious, or ideological viewpoints cannot be grounds for suppression—the contrary result would utterly undermine the First Amendment’s role in fostering a vibrant democracy. But does discomfort with panhandling really raise the same concerns? I would argue perhaps not. After all, panhandling is hardly speech closely tied to democratic self-governance or the shaping of citizens’ political and cultural values. Furthermore, the discomfort caused by solicitation is arguably also not political or ideological in nature, it is rather personal. Moreover, the discomfort is a real harm, which perhaps the government should have some discretion to address, even if the harm is content-based, as I think we must acknowledge it is. If we accept these premises then perhaps this is another area where we should permit the state to draw content-based distinctions, so long as those distinctions adhere to the hierarchy of values implicit in the First Amendment and do not target viewpoints. On this view, restricting solicitation because of discomfort or offense is permissible, because the discomfort does not relate to ideology or otherwise implicate democratic values. On the other hand, the suppression of political speech because of concerns about discomfort or offense that was upheld in the Lehman should not have been permitted, because it turns First Amendment values on their head.256

I must confess, however, that my conclusion regarding solicitation bans is only tentative. This is because there is an argument to be made that discomfort with panhandling is not in fact apolitical. Rather, it is closely tied to feelings about class and the distribution of resources in our society, all of which are of course profoundly political subjects, highly relevant to self-governance. Arguably, just as tourists do not want to be reminded of poverty when on vacation, they also do not want to be reminded of contentious foreign policy issues, yet surely a ban on antiwar protests in Seattle Center could not be upheld. Furthermore, unlike in the case of scarcity, where speech opportunities must be allocated, there is no necessity of selectively suppressing speech here. The question is undoubtedly a close one, but on balance I am inclined to believe that there are strong grounds to at least qualify the general rule that offense at the content of speech may not be a basis for suppressing it, even if such an exception might not apply to solicitation bans for reasons I will come to.

Consider now abortion protestors. As noted earlier, in Hill v. Colorado a majority of the Court appeared to accept that the government has a legitimate, even an important interest in shielding individuals from anti-abortion speech because of the discomfort it causes them. Admittedly, the Court has retreated from this position in its more recent McCullen decision, but to my mind not entirely convincingly. As Justice Scalia points out in dissent, the sheer scope of the restriction on speech in McCullen seems inconsistent with its purported targeting of only intimidation or physical interference with access to clinics; rather, it seems likely that at least part of the motivation was also to protect patients from the discomfort caused by the content of abortion protestors’ speech.257

Anti-abortion speech is, of course, core political speech, and an active part of the process of democratic self-governance. There can be no general principle, therefore, that such speech can be suppressed because it causes discomfort or offense, especially in public forums such as sidewalks. The same principle, of course, drove the Court’s decision in Snyder v. Phelps dismissing Snyder’s IIED based on the defendants’ extraordinarily offensive protest at his son’s funeral.258 Again, there can be no doubt that the emotional harm to Snyder was caused by the content and viewpoint of the defendants’ speech, and so seemingly under the Cantwell/Terminiello principle the result in the case followed a fortiori. That may well be correct (and Hill therefore incorrect), but there is one possible counterargument. Even if we adhere to the view that generally, discomfort and offense caused by ideological speech must be endured, might we create an exception for especially vulnerable settings such as patients at healthcare facilities, or family members at private funerals?259 We must, I think, abandon the fiction that laws seeking to protect against discomfort and offense are content-neutral—they are not, since whatever their wording, their clear and undeniable purpose and effect is to suppress specific messages because of the discomfort or offense they engender. This is as true of prohibitions on protesting near abortion clinics as it is of restrictions on protests near funerals. But it may be that this is another instance in which our across-the-board condemnation of content discrimination may go too far. Perhaps in such vulnerable settings some content distinctions are permissible, so long as they genuinely target genuinely serious emotional harms.260 Of course, any such rule would have to highly restrictive—a long list of “especially vulnerable settings” could gut political dialogue. With close judicial oversight, however, one can imagine tolerating some content discrimination—i.e., making some subjects, such as abortion near clinics and the character of the decedent near funerals, off limits—without seriously undermining democratic self-governance.

I should note at this point that the concern I have just identified, regarding the need to define vulnerable settings narrowly, cuts against upholding the bans on solicitation I discuss above. Those bans tend, as I noted, to focus on tourist-friendly areas of cities. But it is hard to take seriously the idea that tourists are an especially vulnerable population, or that touristy locales such as Seattle’s City Center constitute vulnerable settings. As such, even if the discomfort caused by begging is apolitical (which, as I noted, there is reason to doubt), nonetheless the state interest in preventing that discomfort pales compared to the facts of Snyder v. Phelps and the abortion protest cases. On balance, therefore, because of doubts about the apolitical nature of solicitation bans, and because of the absence of any especially vulnerable population, I think that even if we permit speech-regulation to prevent emotional harms, I would not extend that exception to solicitation bans.

Finally, we should briefly consider Cohen v. California and the problem of offensive words and imagines. California’s regulatory goal in prosecuting Cohen for his “Fuck the Draft” jacket was clearly to protect onlookers from offense and to shield any children present from inappropriate words. Clearly, this is a content-based goal, since the offense and inappropriateness are tied to a specific word. Moreover, Cohen’s protest against the draft and the Vietnam War was obviously core, political speech. However, California’s objection was not to his ideological message, it was to Cohen’s specific language choice. Is the objection to the word “Fuck” more like an objection to a political message, something we should almost never tolerate,261 or like the personal objection I posit towards panhandling? The question is frankly a close one, making Cohen a genuinely hard case. On the one hand, one could argue that offense at foul language, and a desire to shield children from it, is purely personal, and has little connection to ideology, politics, or democratic participation. Support for this position might be found in the fact that people of many different political and ideological persuasions are offended by such language, or at least wish to shield their children from it. On the other hand, it might alternatively be argued that feelings about the use of “inappropriate” language are deeply tied to cultural and political values, and were very much a part of the culture wars of the 1960s and 1970s. I am inclined to the latter perspective, and therefore conclude that Cohen was rightly decided because suppressing the word “Fuck” is a political and cultural choice that the state cannot make consistent with First Amendment values. After all, there was a time when it was thought that exposing children to even the concept of homosexuality was inappropriate; yet today, surely most would agree that such a preference is highly political, and therefore may not be effectuated by the state. I am also inclined to this view because of serious slippery slope problems. If the state can suppress the word “Fuck,” can it also suppress pictures of aborted fetuses, held up by abortion protestors in public places? Such pictures are also deeply offensive to many, and would be considered inappropriate for children by most. Permitting such selective suppression, however, has the potential to have a significant impact on political dialogue, and therefore the price is too high to pay for the social gains at issue.

C.     Secondary Effects and Low-Value Speech

As noted in Part II, one of the most visible departures in the Supreme Court from general principles of content neutrality is its use of the “secondary effects” doctrine to uphold zoning regulations aimed at sexually oriented businesses.262 There can be no doubt that such zoning regulations are on their face content-based, as numerous dissenting and concurring justices, as well as many commentators, have pointed out.263 Yet common knowledge and sense also tells us that SOBs are in fact associated with blight and crime, for any number of reasons. Should the existence of secondary effects therefore be another instance where we should forthrightly acknowledge that some deviation from content neutrality is permissible?

As a general matter, my answer is no. The problem is that secondary effects are ubiquitous. It is true that SOBs tend to attract crime and blight.264 It is also true that KKK rallies tend to attract bottle-throwing objectors, and that anti-police-violence rallies have sometimes attracted violent nihilists (and, unfortunately, occasionally violent police officers). Can we therefore prohibit such rallies, based not on their content but on their secondary effects? Surely not. It is also no use to say that the latter situations are different because it is the content of the speech that triggers the violence. There can be no doubt that it is the sexual content of the books and movies sold by SOBs that attract crime and blight. That is why regulations of SOBs are facially content-based.

The broader point here is that the exceptions to the general principle of content neutrality that I have proposed here are designed to be narrow and objectively definable. Any other approach would undermine the broader presumption against content-based regulations that I am not challenging here. The secondary effects doctrine, however, is not such a narrow exception; it is potentially expansive. As Mark Rienzi and Stuart Buck discuss, in lower courts the doctrine has indeed proven expansive and systematically undercut the general requirement of content neutrality.265

It is true that the secondary effects doctrine has not had such a broad impact in the Supreme Court itself, but that is because the Court has strictly limited the doctrine to zoning of SOBs.266 But what is the justification for doing so? The Court has provided none. Nor, under current doctrine, can the limitation be defended on the grounds that sexual speech is “low value,” since as discussed above the Court has clearly rejected that argument in non-zoning contexts.267 Moreover, the Court’s current jurisprudence limits low-value categories to those historically recognized as such.268 Geoffrey Stone points out, however, that there is no historical basis for treating sexual speech as low-value or unprotected.269 All of this suggests that the entire secondary effects doctrine is illegitimate and should be abandoned. The difficulty with this is that for four decades, a consistent majority of the Court has struggled to uphold zoning of SOBs, and shows no signs of backing off this commitment. Why might that be?

The answer must lie in a combination of two factors. First, as Justice Kennedy emphasized in the Alameda Books case: Zoning rules applied to SOBs generally do not—and in Kennedy’s view cannot—entirely suppress sexual speech.270 Instead, they merely move the speech around physically, so that the impact of such zoning laws is muted. But that cannot be enough—surely a zoning rule excluding Marxist bookstores from one specific neighborhood would be invalid. The second factor is that whatever the Court says, no one really believes that pornographic books and movies contribute to First Amendment values in the same way or to the same extent as an important political work like Marx’s Das Kapital. If one is perfectly honest, it is hard to deny that zoning regulations of SOBs, though no doubt primarily motivated by concerns about secondary effects, are also motivated to some extent by hostility to such speech because of its content. Such hostility towards Marxism would be considered an entirely illegitimate basis for regulation. But do we really feel the same way about pornography? I think not, which suggests a way forward.

Based on the above considerations, I would suggest we reframe the secondary effects doctrine in this way. First, we should acknowledge that regulations targeting secondary effects are indeed content-based, not content-neutral. Nonetheless, we should permit such regulations so long as the following conditions are met: (1) the state must make an evidentiary showing that a particular, well-defined category of speech does indeed create harmful secondary effects; (2) the category of speech must be one with little or no connect to democratic self-governance; and (3) a regulation must not have the practical impact of entirely eliminating, or even severely restricting, such speech within the relevant jurisdiction. If these fairly stringent conditions are met (I for one cannot think of any class of laws aside from zoning of SOBs that could satisfy all three), perhaps we could recognize another, narrowly defined exception to the background requirement of content neutrality.

D.     Privacy and Private Facts

As noted earlier, as a strictly doctrinal matter, essentially all laws which seek to preserve privacy by prohibiting the disclosure of specific, sensitive facts about individuals, are content-based regulations of fully protected speech and so are presumptively unconstitutional.271 Furthermore, privacy regulations cannot generally be saved on the rationale that they target secondary effects or are otherwise not truly content-based. Most privacy regulations are not concerned with secondary effects; they are concerned with how others will judge, and use, certain sorts of sensitive details about individuals if they became available to third parties or to the world. In other words, privacy laws are designed to regulate the communicative impact of personal facts.

How then are we to rationalize privacy legislation from a constitutional perspective? Jane Bambauer argues we should not bother; rather, we should accept that most privacy legislation is unnecessary and counterproductive, and so permit such regulation only when the government can satisfy strict scrutiny.272 I, however, think that this might go too far. The reason is, quite simply, that loss of privacy is felt as a very serious harm by many individuals, and in the vast majority of instances, the speech, and the information disclosed, have little or no relevance to democratic self-governance. This, then, seems like another tailor-made situation for recognizing a narrow exception to the general prohibition on content discrimination.

If a privacy exception is to be recognized, it must of course be defined narrowly and carefully. First of all, any such exception must be limited to disclosure of facts about individuals.273 Any expansion of this exception to cover opinions, facts about companies or other non-individual facts, or worst of all facts about the government, would threaten to substantially impact political debate, and so must be resisted.

Second, the exception should apply only to facts, or rather the disclosures of facts, that have essentially no relationship to democratic self-governance. Even with respect to entirely private individuals, facts about their political affiliations, or political contributions, might well be a part of an ongoing debate, so that disclosure of such facts cannot be suppressed. However, facts about, say, medical conditions or shopping preferences are different. Of course, it matters who the fact concerns. It is hard to comprehend how a disclosure that I suffer from a heart murmur (I don’t) has any relevance to democratic self-governance. On the other hand, similar information about Hillary Clinton or Donald Trump surely would have been relevant to their fitness to serve as President during the 2016 campaign. Context matters, and in case of doubt courts must err in favor of finding speech to be relevant to self-governance.

Finally, privacy laws must be limited to situations where there exists a broad societal consensus that the protected facts are indeed highly sensitive, and so their disclosure would do substantial emotional harm to a reasonable person. Without this important limitation, privacy law could devolve into an excuse for extensive regulation of factual disclosures. As Justice Kennedy pointed out in Sorrell v. IMS Health Inc., however, “[f]acts . . . are the beginning point for much of the speech that is most essential to advance human knowledge and to conduct human affairs.”274 Any widespread legislative efforts to suppress facts, even private facts about private individuals, would in the long term threaten to hinder academic research, and more broadly the compilation of valuable information about citizens and our society. Since such information is an essential input into the public debates that undergird democratic self-governance, any such expansion must be prevented.

VI.     Concerns and Ways Forward

Up to this point in the Article, I have argued in favor of modifying the general doctrinal presumption against content-based regulations of speech to recognize some narrow, well-defined exceptions where some such restrictions might make sense. In this Part, I consider some objections to the abandonment of strict content-neutrality and suggest why they are not inconsistent with my proposal. In particular, I consider two important counterarguments: Geoffrey Stone’s argument that subject-matter restrictions should be viewed suspiciously because they can be a guise for viewpoint discrimination,275 and Seth Kreimer’s argument that the prohibition on content discrimination is an essential means to control the petty censorship of what he calls “village tyrants.”276

In Stone’s classic article Restrictions of Speech Because of Its Content: The Peculiar Case of Subject-Matter Restrictions,277 he considers the problem of subject-matter restrictions in the context of the broader presumption against content discrimination. Subject-matter restrictions, he notes, are undoubtedly content-based, but they are viewpoint neutral.278 As such, they are less likely to distort political debate, and are also less likely to be motivated by legislative hostility, than other content-based laws.279 Nonetheless, Stone argues that even subject-matter restraints should be viewed suspiciously, albeit not as suspiciously as viewpoint-based laws. The primary reason for this is that many facially neutral subject-matter restrictions will have an obvious, predictable, and intended viewpoint-differential impact, making them indistinguishable from a viewpoint-based law—he gives as one example “banning street demonstrations concerning the propriety of the Vietnam War.”280 He also notes that relatively narrow subject-matter restrictions are especially likely to have this intended impact, suggesting the need for particularly searching judicial scrutiny of such laws.281 While raising the question whether suspicion should be limited to laws creating “a demonstrable viewpoint-differential effect,” he ultimately rejects such a limitation because of the likelihood of such an effect, combined “with the general presumption in favor of clarity and ease of administration.”282 One particular example he gives is laws that restrict sexual expression, because such speech usually carries “an implicit, if not explicit, message in favor of more relaxed sexual mores.”283 Finally, however, Stone concedes that broader subject-matter restrictions generally do not carry the same danger, and so probably should receive more relaxed scrutiny.284

Seth Kreimer has recently advanced an alternative argument for retaining a suspicious attitude towards content-based regulations, which dovetails nicely with Stone’s argument.285 He contrasts the sorts of First Amendment cases typically litigated in the lower courts with those more often seen in the Supreme Court.286 As it turns out, the mix of cases looks very different. In the Supreme Court, most free speech cases involve corporate litigants, large nonprofits, or wealthy individuals challenging campaign finance laws.287 Almost none, and especially almost none where the claimant was successful, involve what Kreimer calls “village tyrants”—i.e., local officials or street level bureaucrats exercising censorial power—or “village Hampdens”—i.e., individuals dissenting against authority.288 In the lower courts, in contrast, two-thirds of the cases in the courts of appeals, and a whopping three-quarters of cases in the district courts, involved village tyrants.289 Similarly, village Hampdens were involved in 79% of lower court cases, and made up 88% of successful claimants (though most claimants, both individuals and businesses, lost).290

The main conclusion Kreimer draws from these statistics is that because of the very different caseloads faced by the Supreme Court versus lower court, doctrinal rules that are imperfect in the Supreme Court may be well-suited to the lower courts.291 In particular, Kreimer concedes that given the Supreme Court’s docket, a more nuanced doctrine that does not evince across-the-board hostility to content regulation may produce better results.292 The lower courts, however, are quite different. The benefit of a flat bar on content discrimination is that it is more likely to protect village Hampdens than a more nuanced approach because it provides clearer signals to village tyrants that they may not suppress speech simply because of disagreement with it.293 In particular, he argues that because of political biases and mental heuristics, a more nuanced doctrine such as proportionality review is unlikely to prevent village tyrants from engaging in inappropriate censorship, especially because most decisions to restrict speech are made in rushed and emotional circumstances, where heuristics are most likely to be in play.294 Hence the need to retain a robust, bright-line rule that flatly bans content discrimination.

It is noteworthy that both the Stone and Kreimer arguments rely at their heart on prophylaxis. Both of them concede that a more nuanced doctrine, permitting some but not all content discrimination, may produce better results. Indeed, Stone is explicitly willing to tolerate broad, but not narrow, subject-matter restrictions.295 Neither, however, trusts either government officials or the courts to fairly administer such a doctrine, and so fall back on the need retain a broad (in Kreimer’s case universal) suspicion of content discrimination.296

To describe their arguments as prophylactic is not, of course, to refute them. However, recognition of the nature of the objections points at least to a partial response. In general, I fully share both Stone’s and Kreimer’s concerns. Clearly, subject-matter restraints can mask viewpoint discrimination, and clearly simple, bright-line rules are more effective at signaling proper behavior than complex standards, especially when the audience is not legally sophisticated. That is why I do not advocate a wholesale abandonment of the content-neutrality doctrine. It is also why I cannot endorse the arguments of the concurring Justices in Reed. Both Justices Breyer and Kagan filed separate opinions in Reed concurring in the result, but disagreeing with the majority’s conclusion that all content-based laws should automatically receive strict scrutiny.297 The difficulty is that this is essentially all the guidance they offer. Justice Breyer merely suggests “the category ‘content discrimination’ is better considered in many contexts . . . as a rule of thumb, rather than as an automatic ‘strict scrutiny’ trigger.”298 And Justice Kagan’s advice is that we “administer our content-regulation doctrine with a dose of common sense.”299 These vague standards are problematic for precisely the reasons asserted by Stone and Kreimer, and must be rejected.300 There are, however, three responses to the objections from prophylaxis, which in combination might overcome those objections, if a different approach is taken.

The first is simply that what this article calls for is neither a broad-based move towards proportionality and balancing, nor a move from rules to standards. Instead, it calls for judicial identification of a few, clearly and objectively defined, circumstances when courts should be open to some distinctions based on content. To claim the benefit of a circumstance such as, for example scarcity, the government must be able to demonstrate that the circumstance exists, that it was the motivating factor behind the content distinction drawn by the government, and that distinction drawn is a logical response to the particular problem posed, whether it be scarcity, secondary effects, or something else.

Second, what I propose is toleration of some content distinctions, but by no means all. To begin with, only content distinctions that are consistent with First Amendment values are permitted, meaning that, generally, speech higher on the First Amendment hierarchy must be preferred to that which is lower. Thus, as noted earlier, a preference for political over commercial speech may be sometimes permissible, if supported by persuasive argument—but the opposite almost never should be. The corollary of this is that viewpoint preferentialism should be flatly impermissible, since such distortion of political or cultural debate is antithetical to the purpose of the First Amendment. Thus, the existence of scarcity might justify favoring political over other speech, but it obviously cannot justify preferring patriotic over other speech, or speech expounding the views of one religious sect over that of another sect.301

There is one possible caveat to the general rule barring viewpoint preferences—the problem of emotional harm in highly sensitive situations such as funerals or health-care clinics. The reality is that in those situations, emotional harm sometimes is closely related to viewpoints. That is, after all, why the Court reversed the IIED award against the Westboro Baptist Church in Snyder v. Phelps.302 Furthermore, it is only the harmful viewpoints that are likely to be expressed in those places. No one protests military funerals except the Westboro Baptist Church, and no one pickets abortion clinics except antiabortion protestors. Protecting individuals in those circumstances, therefore, requires enacting laws which, even if content-neutral on their face, have obvious, predictable, and desired viewpoint-differential impacts. They are, for all practical purposes, viewpoint-based laws. A similar problem was posed in Virginia v. Black, where the Court upheld a Virginia statute that barred burning a cross in order to intimidate someone.303 The majority was no doubt correct in that case that threats are unprotected speech, and that burning a cross is a particularly virulent kind of threat.304 The difficulty, as the dissent pointed out, is that the virulence of cross burning cannot be disentangled from its association with one particular viewpoint—the white, Protestant supremacy advocated by the Ku Klux Klan.305 These cases thus pose the very difficult question of whether the First Amendment ever permits laws with such intended, viewpoint discriminatory effects, even if the goal is to prevent emotional harm rather than distort debate. The general answer is, of course, clearly not; but the answer is harder in specific situations such as where the harmed individual is in a particularly vulnerable circumstance, or where the underlying speech (as in Black) is harmful and unprotected. The only possible conclusion is that this is a very close call, on which reasonable minds will surely differ.

Third, and perhaps most critically, I envision continuing stringent judicial review to enforce the first two limitations. Thus, courts must maintain a tight rein to ensure that any recognized exceptions to the ban on content discrimination remain narrow, and are actually present in a specific case. They must also, over time, identify and define a First Amendment hierarchy of subject matter to give government officials clear guidance about what sorts of preferences are, and are not, permissible. Finally, courts must take particular care to ensure that a subject-matter preference is not acting as a mask for viewpoint discrimination. Thus, as Stone suggests, they should be especially suspicious of narrow, as opposed to broad, subject-matter restrictions. They must also be willing to explore the background circumstances generating legislation, as well as legislative intent, to ferret out such discrimination. Indeed, I would argue that proof of significant disparate impact on a particular viewpoint should be sufficient to create at least a presumption of unconstitutionality, given the well-known overlap between subject-matter-based and viewpoint-based distinctions.

Whether or not the three restrictions I propose, in combination, are sufficient to overcome the objections to loosening of the bar on content discrimination is, of course, in the eyes of the beholder. What this Article should have made clear, however, is that adhering to a strict requirement of content neutrality also has its costs, including requiring society to suffer unnecessary harms and forbidding government officials from enhancing political debate at the heart of the First Amendment. The time has come, I would argue, to take those costs seriously.

VII.     Conclusion

The proposal I have defended here—that we should reformulate free expression doctrine to permit some forms of content discrimination—is both a radical one and a relatively modest one at the same time. It is radical because it attacks the central premise of First Amendment doctrine over the past 40 years, that laws regulating protected speech based on content are always dangerous. It is modest because it does not propose to entirely or even primarily abandon strict scrutiny of content discrimination; it only proposes developing some narrow exceptions to that rule, and permitting content preferences only when consistent with First Amendment values.

There is undoubtedly some risk associated with abandonment of a strict hostility to content discrimination, that permitted content discrimination will become a tool to suppress unpopular messages. And while that risk can be mitigated, it cannot be eliminated. On the other hand, the discussion of case law in Part I quite clearly demonstrates that the current rule is simply not working. It produces irrational results in several important situations, and as a result is regularly evaded by judges. To frankly acknowledge what is already going on, then, has the advantage of honesty and consistency, and it will hopefully result in a more clearly articulated jurisprudence defining the narrow circumstances in which we should abandon strict content neutrality.

Underlying this debate is a deeper conundrum—should we prefer a few simple, clear, but clearly overbroad rules of free speech doctrine, or would a more complex set of rules produce better results? My general preference is for simple rules. They are easier to enforce, they provide better notice to citizens and public officials, and they tend to be more stable over time. The reality is, though, that First Amendment doctrine has moved too far in the direction of simplicity. Recent Supreme Court decisions purport to establish (or reaffirm) the principles that almost all speech is fully protected (except for a handful of historically defined exceptions), and that within that fully-protected realm, content-based regulations are almost never allowed. The results mandated by this approach, however, are clearly intolerable to society and much of the judiciary, which is the cause of the chaos recounted in Part I. It is time, I would argue, to move a little bit, but not too far, in the direction of complexity, in an effort to produce results that we can actually live with.


 

  1. [1]. Police Dep’t of Chi. v. Mosley, 408 U.S. 92, 95 (1972).

  2. [2]. Cf. Williams–Yulee v. Fla. Bar, 135 S. Ct. 1656, 1664–73 (2015) (majority opinion upholding content-based regulation of speech during judicial elections after applying strict scrutiny, but in which only four justices joined the portion of Chief Justice Roberts’s opinion which held that strict scrutiny should apply to the challenged law); Burson v. Freeman, 504 U.S. 191, 211 (1992) (upholding a content-based law after applying strict scrutiny in a plurality opinion).

  3. [3]. See Holder v. Humanitarian Law Project, 561 U.S. 1, 28–39 (2010) (upholding a statute prohibiting material support to foreign terrorist organizations after concluding that the law is, in some applications, a content-based regulation of speech).

  4. [4]. Barry P. McDonald, Speech and Distrust: Rethinking the Content Approach to Protecting the Freedom of Expression, 81 Notre Dame L. Rev. 1347, 1351–52 (2006).

  5. [5]. See generally Brown v. Entm’t Merchs. Ass’n, 564 U.S. 786 (2011) (applying strict scrutiny to a statute forbidding the sale of violent video games to minors); United States v. Stevens, 559 U.S. 460 (2010) (providing constitutional protection to depictions of animal cruelty); United States v. Playboy Entm’t Grp., Inc., 529 U.S. 803 (2000) (applying strict scrutiny to a statute requiring full scrambling of cable television channels primarily dedicated to sexually-oriented programming).

  6. [6]. See infra Part II.A.

  7. [7]. See infra Part II.B.

  8. [8]. Richard Re names this phenomenon “narrowing precedent from below.” See Richard M. Re, Narrowing Supreme Court Precedent from Below, 104 Geo. L.J. 921, 927–35 (2016).

  9. [9]. Reed v. Town of Gilbert, 135 S. Ct. 2218 (2015).

  10. [10]. Id. at 2225.

  11. [11]. Id. at 2224.

  12. [12]. Id. at 2224–25 (alterations in original).

  13. [13]. Id. at 2224.

  14. [14]. Id. (quoting Town of Gilbert: Land Development Code, Glossary of General Terms 24 (2016), https://www.gilbertaz.gov/home/showdocument?id=8485).

  15. [15]. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2225 (2015).

  16. [16]. Id. at 2224–25.

  17. [17]. Id. at 2225.

  18. [18]. Id. at 2224–25.

  19. [19]. Id.

  20. [20]. Id. at 2225.

  21. [21]. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2225 (2015).

  22. [22]. See id.; Reed v. Town of Gilbert (Reed II), 707 F.3d 1057, 1060 (9th Cir. 2013), rev’d, 135 S. Ct. 2218 (2015); Reed v. Town of Gilbert (Reed I), 587 F.3d 966, 971 (9th Cir. 2009).

  23. [23]. Reed II, 707 F.3d at 1079 (Watford, J., dissenting).

  24. [24]. Reed I, 587 F.3d at 972.

  25. [25]. Judge Watford on the Ninth Circuit dissented from the second panel’s opinion. Reed II, 707 F.3d at 1077–81.

  26. [26]. Reed, 135 S. Ct. at 2226. However, the unanimity extended only to the judgment.

  27. [27]. Reed II, 707 F.3d at 1080 (Watford, J., dissenting).

  28. [28]. Id.

  29. [29]. Id. at 1071 (citing Hill v. Colorado, 530 U.S. 703, 719 (2000)); Reed I, 587 U.S. at 974.

  30. [30]. Reed v. Town of Gilbert (Reed II), 707 F.3d 1057, 1071 (9th Cir. 2013), rev’d, 135 S. Ct. 2218 (2015).

  31. [31]. Reed I, 587 F.3d at 975 (“Nothing in the regulation suggests any intention by Gilbert to suppress certain ideas through the Sign Code, nor does Good News claim that Gilbert had any illicit motive in adopting the ordinance.”).

  32. [32]. Id. at 977.

  33. [33]. Reed II, 707 F.3d at 1069.

  34. [34]. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2227 (2015) (quoting Sorrell v. IMS Health Inc., 564 U.S. 552, 567 (2011)).

  35. [35]. Id. (second alteration in original) (quoting Ward v. Rock Against Racism, 491 U.S. 781, 791 (1989)).

  36. [36]. Id.

  37. [37]. Id.

  38. [38]. Id. at 2231–32.

  39. [39]. Id. at 2227–29.

  40. [40]. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2227–29 (2015); see also Seth F. Kreimer, Good Enough for Government Work: Two Cheers for Content Neutrality, 16 U. Pa. J. Const. L. 1261, 1268–69 & nn.20–21, 1274–75 & n.47 (2014); McDonald, supra note 4, at 1377–78 & nn.127–33 (listing cases).

  41. [41]. See, e.g., Simon & Schuster, Inc. v. Members of the N.Y. State Crime Victims Bd., 502 U.S. 105, 117 (1991) (“The Board next argues that [a law imposing] discriminatory financial treatment is suspect under the First Amendment only when the legislature intends to suppress certain ideas. This assertion is incorrect; our cases have consistently held that ‘[i]llicit legislative intent is not the sine qua non of a violation of the First Amendment.’” (second alteration in original)(quoting Minneapolis Star & Tribune Co. v. Minn. Comm’r of Revenue, 460 U.S. 575, 592 (1983))); Ark. Writers’ Project, Inc. v. Ragland, 481 U.S. 221, 228–30 (1987) (finding a selective tax on publications to be content based despite “no evidence of an improper censorial motive”); cf. Burson v. Freeman, 504 U.S. 191, 195, 211 (1992) (finding a law banning campaign speech near polling places to be content-based, even though the law did not distinguish among ideas); Police Dep’t of Chi. v. Mosley, 408 U.S. 92, 94 (1972) (striking down a law banning picketing outside of schools, but permitting labor picketing despite the lack of evidence of hostility to particular ideas); Kreimer, supra note 40, at 1274 & n.45.

  42. [42]. Reed, 135 S. Ct. at 2229–30.

  43. [43]. Id. at 2230–31.

  44. [44]. Id.

  45. [45]. Id. at 2234 (Breyer, J., concurring in the judgment); id. at 2238 (Kagan, J., concurring in the judgment).

  46. [46]. Id. at 2234 (Breyer, J., concurring in the judgment).

  47. [47]. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2238 (2015) (Kagan, J., concurring in the judgment).

  48. [48]. Petition for Writ of Certiorari at 19–27, Reed v. Town of Gilbert, 135 S. Ct. 2218 (2015) (No. 13-502). For a recent, but pre-Reed description of this circuit split, see Marc Rohr, De Minimus Content Discrimination: The Vexing Matter of Sign-Ordinance Exemptions, 7 Elon L. Rev. 327, 356–72 (2015).

  49. [49]. Petition for Writ of Certiorari, supra note 48, at 22–24. A good example of an unambiguous adoption of a pure, purpose-based approach is Brown v. Town of Cary, in which the Fourth Circuit upheld a sign ordinance, even though it exempted holiday displays and art from its restrictions. Brown v. Town of Cary, 706 F.3d 294, 297–98 (4th Cir. 2013). The court rejected the plaintiff’s argument that “a regulation that depends on content distinctions is necessarily content based,” concluding instead that the key question is “why—not whether—the Town has distinguished content in its regulation.” Id. at 301. It also explicitly acknowledged a circuit split on this issue, contrasting its approach to the “absolutist reading of content neutrality” adopted by other circuits. Id. at 302. In Century Radio Co. v. City of Norfolk, the Fourth Circuit originally reaffirmed Brown’s approach to content regulation and, therefore, upheld as content-neutral a sign ordinance that exempted certain types of displays based on their content. Cent. Radio Co. v. City of Norfolk, 776 F.3d 229, 235–36 (4th Cir. 2015). On remand from the Supreme Court after Reed, however, the court changed course, abandoned its focus on intent, and held the sign ordinance to be content-based (and invalidated it). Cent. Radio Co. v. City of Norfolk, 811 F.3d 625, 632–34 (4th Cir. 2016).

  50. [50]. Petition for Writ of Certiorari, supra note 48, at 21–22 (Neighborhood Enters., Inc. v. City of St. Louis, 644 F.3d 728 (8th Cir. 2011); Solantic, LLC v. City of Neptune Beach, 410 F.3d 1250 (11th Cir. 2005); Nat’l Advert. Co. v. Town of Babylon, 900 F.2d 551, 557 (2d Cir. 1990); Matthews v. Town of Needham, 764 F.2d 58, 60 (1st Cir. 1985)).

  51. [51]. Petition for Writ of Certiorari, supra note 48, at 25–26.

  52. [52]. Kreimer, supra note 40, at 1270 n.24.

  53. [53]. Ward v. Rock Against Racism, 491 U.S. 781, 791 (1989).

  54. [54]. Hill v. Colorado, 530 U.S. 703, 719 (2000).

  55. [55]. Id. at 720–21 (quoting Colo. Rev. Stat. § 18-9-122 (1999)).

  56. [56]. Id. at 721 (“Whether a particular statement constitutes a threat, blackmail, an agreement to fix prices, a copyright violation, a public offering of securities, or an offer to sell goods often depends on the precise content of the statement.”).

  57. [57]. Id. at 746–47 (Scalia, J., dissenting).

  58. [58]. See generally Berger v. City of Seattle, 569 F.3d 1029 (9th Cir. 2009) (en banc).

  59. [59]. Norton v. City of Springfield, 768 F.3d 713, 717–18 (7th Cir. 2014).

  60. [60]. Id. at 714–15.

  61. [61]. Id. at 717.

  62. [62]. Id. at 716.

  63. [63]. Id. at 716–17.

  64. [64]. Id. at 718–23 (Manion, J., dissenting).

  65. [65]. Norton v. City of Springfield, 768 F.3d 713, 721–22 (7th Cir. 2014).

  66. [66]. Norton v. City of Springfield, 806 F.3d 411, 412–13 (7th Cir. 2015). Springfield did not put forward any argument that the law satisfied strict scrutiny. Id. at 413.

  67. [67]. Berger v. City of Seattle, 569 F.3d 1029(9th Cir. 2009) (en banc).

  68. [68]. Id. at 1034.

  69. [69]. Id. at 1051.

  70. [70]. Id. at 1050.

  71. [71]. Id. at 1051 (citation omitted). Note that this approach is in direct conflict with the Ninth Circuit’s later analysis in the Reed cases, demonstrating the extent of the confusion. See supra notes 10–33 and accompanying text.

  72. [72]. Berger, 569 F.3d at 1051.

  73. [73]. Berger v. City of Seattle, 569 F.3d 1029, 1078–79 (9th Cir. 2009) (Gould, J., dissenting).

  74. [74]. Id. at 1092 (N.R. Smith, J., concurring in part, dissenting in part).

  75. [75]. ACLU of Nevada v. City of Las Vegas, 466 F.3d 784, 796–97(9th Cir. 2006).

  76. [76]. Id. at 793.

  77. [77]. Id. at 794–95.

  78. [78]. Id. at 796.

  79. [79]. Clatterbuck v. City of Charlottesville, 708 F.3d 549(4th Cir. 2013).

  80. [80]. Norton v. City of Springfield, 768 F.3d 713, 714 (7th Cir. 2014) (citing Clatterbuck, 707 F.3d at 549); id. at 719 (Manion, J., dissenting) (citing Clatterbuck, 708 F.3d at 556).

  81. [81]. Clatterbuck, 708 F.3d at 556.

  82. [82]. Id.

  83. [83]. Id. In the recent decision in Central Radio Co. v. City of Norfolk, the Fourth Circuit suggested that it has abandoned its approach in Clatterbuck. See Cent. Radio Co. v. City of Norfolk, 811 F.3d 625, 632 (4th Cir. 2016).

  84. [84]. See Clatterbuck, 708 F.3d at 556.

  85. [85]. Clatterbuck v. City of Charlottesville, 92 F. Supp. 3d 478, 487–89 (W.D. Va. 2015).

  86. [86]. See infra Part V.B.

  87. [87]. Hill v. Colorado, 530 U.S. 703 (2000).

  88. [88]. Id. at 707 (quoting Colo. Rev. Stat. § 18-9-122(3) (1999)).

  89. [89]. Id. at 721–23.

  90. [90]. Id. at 724.

  91. [91]. Id. at 716, 725–26.

  92. [92]. Id. at 747–48 (Scalia, J., dissenting) (citing Boos v. Barry, 485 U.S. 312, 321 (1988)).

  93. [93]. Hill v. Colorado, 530 U.S. 703, 745–46 (2000) (Scalia, J., dissenting); see also id. at 766–67 (Kennedy, J., dissenting).

  94. [94]. Id. at 768–69 (Kennedy, J., dissenting).

  95. [95]. McCullen v. Coakley, 134 S. Ct. 2518 (2014).

  96. [96]. Id. at 2522 (quoting Mass. Gen. Laws, ch. 266, § 120E1/2(a)–(b) (2007)). The statute exempted several classes of individuals, including people entering or exiting the facility, people walking through the buffer zone, public employees, and employees of the clinic. Id. at 2526.

  97. [97]. The majority nonetheless invalidated the statute on the grounds that the law failed intermediate scrutiny because it silenced substantially more speech than necessary. Id. at 2541.

  98. [98]. Id. at 2530–31.

  99. [99]. Id. at 2531 (quoting City of Renton v. Playtime Theatres, Inc., 475 U.S. 41, 48 (1986)).

  100. [100]. McCullen v. Coakley, 134 S. Ct. 2518, 2531–32 (2014) (quoting Boos v. Barry, 485 U.S. 312, 321 (1988)).

  101. [101]. Id. at 2546 (Scalia, J., concurring in the judgment).

  102. [102]. Id. at 2544.

  103. [103]. Id. at 2546–48.

  104. [104]. See generally Schenck v. Pro-Choice Network of W. N.Y., 519 U.S. 357 (1997); Madsen v. Women’s Health Ctr., Inc., 512 U.S. 753 (1994).

  105. [105]. See Elena Kagan, Private Speech, Public Purpose: The Role of Governmental Motive in First Amendment Doctrine, 63 U. Chi. L. Rev. 413, 483–91 (1996). See generally David L. Hudson, Jr., The Secondary Effects Doctrine: “The Evisceration of First Amendment Freedoms, 37 Washburn L.J. 55 (1997); Mark Rienzi & Stuart Buck, Neutral No More: Secondary Effects Analysis and the Quiet Demise of the Content-Neutrality Test, 82 Fordham L. Rev. 1187 (2013).

  106. [106]. See, e.g., City of Los Angeles v. Alameda Books, Inc., 535 U.S. 425, 448 (2002) (Kennedy, J., concurring) (describing secondary effects doctrine as “something of a fiction”); id. at 457 (Souter, J., dissenting) (noting that zoning regulations targeting adult businesses are “content correlated”); City of Renton v. Playtime Theatres, Inc., 475 U.S. 41, 56 (1986) (Brennan, J., dissenting) (describing secondary effects analysis as “misguided”).

  107. [107]. Young v. Am. Mini Theatres, Inc., 427 U.S. 50 (1976).

  108. [108]. Id. at 72–73.

  109. [109]. Id. at 70.

  110. [110]. Id. at 71.

  111. [111]. Id. at 84–96 (Stewart, J., dissenting).

  112. [112]. City of Renton v. Playtime Theatres, Inc., 475 U.S. 41 (1986).

  113. [113]. Id. at 43.

  114. [114]. Id. at 47–48.

  115. [115]. Id. at 47.

  116. [116]. Id. at 50, 52.

  117. [117]. Id. at 55–65 (Brennan, J., dissenting).

  118. [118]. City of Los Angeles v. Alameda Books, Inc., 535 U.S. 425 (2002).

  119. [119]. Id. at 429–31.

  120. [120]. Id. at 430.

  121. [121]. Id. at 433–39.

  122. [122]. Id. at 448–49 (Kennedy, J., concurring in the judgment).

  123. [123]. Id. at 457, 462 (Souter, J., dissenting).

  124. [124]. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2227 (2015).

  125. [125]. Rienzi & Buck, supra note 105, at 1204–05.

  126. [126]. Id. at 1205–06.

  127. [127]. See Mark Hachman, The Price of Free: How Apple, Facebook, Microsoft and Google Sell You to Advertisers, PCWorld (Oct. 1, 2015, 3:00 AM), www.pcworld.com/article/2986988/privacy/the-price-of-free-how-apple-facebook-microsoft-and-google-sell-you-to-advertisers.html (detailing what personal information big companies collect and how they use that information).

  128. [128]. See generally Cass R. Sunstein, Beyond Cheneyism and Snowdenism, 83 U. Chi. L. Rev. 271 (2016).

  129. [129]. See European Comm’n, Factsheet on the “Right to Be Forgotten” Ruling (C-131/12), http://ec.europa.eu/justice/data-protection/files/factsheets/factsheet_data_protection_en.pdf (last visited Mar. 18, 2017).

  130. [130]. See generally Neil Richards, Intellectual Privacy: Rethinking Civil Liberties in the Digital Age (2015); Neil M. Richards, Reconciling Data Privacy and the First Amendment, 52 UCLA L. Rev. 1149 (2005).

  131. [131]. See generally Jane Bambauer, Is Data Speech?, 66 Stan. L. Rev. 57 (2014); Ashutosh Bhagwat, Sorrell v. IMS Health: Details, Detailing, and the Death of Privacy, 36 Vt. L. Rev. 855 (2012).

  132. [132]. See Sorrell v. IMS Health Inc., 564 U.S. 552, 570 (2011).

  133. [133]. Id. at 557–58.

  134. [134]. Id. at 570 (citations omitted).

  135. [135]. Id. at 570–73.

  136. [136]. See Bhagwat, supra note 131, at 862–63.

  137. [137]. See generally Video Privacy Protection Act of 1988, 18 U.S.C. § 2710 (2012) (restricting disclosure of video rental information).

  138. [138]. Holder v. Humanitarian Law Project, 561 U.S. 1, 28–39 (2010) (upholding a statute prohibiting material support to foreign terrorist organizations after concluding that the law is, in some applications, a content-based regulation of speech).

  139. [139]. See generally Dahlstrom v. Sun-Times Media, LLC, 777 F.3d 937 (7thCir. 2015).

  140. [140]. Id. at 941–42.

  141. [141]. Id. at 949–54.

  142. [142]. Sun-Times Media, LLC v. Dahlstrom, 136 S. Ct. 689 (2015).

  143. [143]. Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n of N.Y., 447 U.S. 557, 561–66 (1980); Va. State Bd. of Pharmacy v. Va. Citizens Consumer Council, Inc., 425 U.S. 748, 761–73 (1976).

  144. [144]. Gertz v. Robert Welch, Inc., 418 U.S. 323, 340 (1974) (stating that “there is no constitutional value in false statements of fact”).

  145. [145]. Miller v. California, 413 U.S. 15, 23–25 (1973).

  146. [146]. See, e.g., City of Los Angeles v. Alameda Books, Inc., 535 U.S. 425, 433–39 (2002); id. at 448–49 (Kennedy, J., concurring in the judgment); City of Renton v. Playtime Theatres, Inc., 475 U.S. 41, 47 (1986); Young v. Am. Mini-Theatres, Inc., 427 U.S. 50, 70 (1976).

  147. [147]. See FCC v. Pacifica Found., 438 U.S. 726, 743, 748–50 (1978).

  148. [148]. New York v. Ferber, 458 U.S. 747, 763–64 (1982)

  149. [149]. See Genevieve Lakier, The Invention of Low-Value Speech, 128 Harv. L. Rev. 2166, 2170–79 (2015).

  150. [150]. See supra Part II.D.

  151. [151]. United States v. Playboy Entm’t Grp., Inc., 529 U.S. 803, 806–07 (2000) (quoting 47 U.S.C. § 561(a) (1994)).

  152. [152]. Ashcroft v. ACLU, 542 U.S. 656, 659–61 (2004).

  153. [153]. Ashcroft v. Free Speech Coal., 535 U.S. 234, 241, 258 (2002).

  154. [154]. United States v. Stevens, 559 U.S. 460 (2010).

  155. [155]. Id. at 481–82.

  156. [156]. See id. at 472 (“Maybe there are some categories of speech that have been historically unprotected, but have not yet been specifically identified or discussed as such in our case law. But if so, there is no evidence that ‘depictions of animal cruelty’ is among them.”).

  157. [157]. Id. at 470.

  158. [158]. Brown v. Entm’t Merchs. Ass’n, 564 U.S. 786, 805 (2011).

  159. [159]. Id. at 792–94; Stevens, 559 U.S. at 472.

  160. [160]. Brown, 564 U.S. at 794–99.

  161. [161]. Id. at 798–99. Indeed, the Court extended full First Amendment protection to video games. Id. at 790.

  162. [162]. Compare id. at 794 (“[M]inors are entitled to a significant measure of First Amendment protection, and only in relatively narrow and well-defined circumstances may government bar public dissemination of protected materials to them.” (alteration in original) (quoting Erznoznik v. Jacksonville, 422 U.S. 205, 212–13 (1975))), with Reno v. ACLU, 521 U.S. 844, 875 (1997) (“[W]e have repeatedly recognized the governmental interest in protecting children from harmful materials.”), and Sable Commc’ns of Cal., Inc. v. FCC, 492 U.S. 115, 126 (1989) (“We have recognized that there is a compelling interest in protecting the physical and psychological well-being of minors. This interest extends to shielding minors from the influence of literature that is not obscene by adult standards.”).

  163. [163]. See supra note 144 and accompanying text (discussing Gertz v. Robert Welch, Inc., 418 U.S. 323 (1974)).

  164. [164]. United States v. Alvarez, 132 S. Ct. 2537, 2546–47 (2012) (plurality opinion).

  165. [165]. Id. at 2542–43.

  166. [166]. Id. at 2543.

  167. [167]. Id. at 2548–49.

  168. [168]. Id. at 2552 (Breyer, J., concurring in judgment).

  169. [169]. See generally Bd. of Trs. of the State Univ. of N.Y. v. Fox, 492 U.S. 469 (1989); Cent. Hudson Gas & Elec. Corp. v. Pub. Serv. Comm’n of N.Y., 447 U.S. 557 (1980).

  170. [170]. See, e.g., Thompson v. W. States Med. Ctr., 535 U.S. 357, 360, 375–76 (2002); Lorillard Tobacco Co. v. Reilly, 533 U.S. 525, 561–66 (2001).

  171. [171]. Sorrell v. IMS Health Inc., 564 U.S. 552, 565–66 (2011) (citing to cases that applied strict scrutiny).

  172. [172]. Id. at 571–72.

  173. [173]. See Metromedia, Inc. v. City of San Diego, 453 U.S. 490, 512–21 (1981) (upholding ban on billboards as applied to commercial billboards, but striking it down as applied to noncommercial billboards); cf. City of Cincinnati v. Discovery Network, Inc., 507 U.S. 410, 412, 428 (1993) (striking down ordinance prohibiting news racks distributing commercial handbills on government property, but permitting news racks containing newspapers).

  174. [174]. John Stuart Mill, The Basic Writings of John Stuart Mill: On Liberty, the Subjection of Women & Utilitarianism 58–59 (2009).

  175. [175]. Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting) (“[T]he best test of truth is the power of the thought to get itself accepted in the competition of the market....”).

  176. [176]. See C. Edwin Baker, Human Liberty and Freedom of Speech 47–69 (1992); Thomas I. Emerson, Toward a General Theory of the First Amendment, 72 Yale L.J. 877, 879 (1963); Seana Valentine Shiffrin, Speech, Death, and Double Effect, 78 N.Y.U. L. Rev. 1135, 1158–85 (2003).

  177. [177]. Whitney v. California, 274 U.S. 357, 372–80 (1927) (Brandeis, J., concurring), overruled in part by Brandenburg v. Ohio, 395 U.S. 444 (1969).

  178. [178]. See generally Alexander Meiklejohn, Free Speech and its Relation to Self-Government (1948); Alexander Meiklejohn, The First Amendment is an Absolute, 1961 Sup. Ct. Rev. 245.

  179. [179]. See Cass R. Sunstein, Democracy and the Problem of Free Speech 121–65 (1993); Robert H. Bork, Neutral Principles and Some First Amendment Problems, 47 Ind. L.J. 1, 20–21 (1971).

  180. [180]. See Robert Post, Participatory Democracy and Free Speech, 97 Va. L. Rev. 477, 482–83 (2011); James Weinstein, Participatory Democracy as the Central Value of American Free Speech Doctrine, 97 Va. L. Rev. 491, 502–03 (2011).

  181. [181]. See Ashutosh Bhagwat, Associational Speech, 120 Yale L.J. 978, 991–94 (2011).

  182. [182]. Citizens United v. FEC, 558 U.S. 310, 339 (2010).

  183. [183]. Snyder v. Phelps, 562 U.S. 443, 451–52 (2011) (quoting Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 758–59 (1985)).

  184. [184]. N.Y. Times Co. v. Sullivan, 376 U.S. 254, 269–71 (1964).

  185. [185]. See Ashutosh Bhagwat, Details: Specific Facts and the First Amendment, 86 S. Cal. L. Rev. 1, 33–35 (2012) (collecting statements in Supreme Court opinions over the course of decades endorsing this position).

  186. [186]. Snyder, 562 U.S. at 450.

  187. [187]. Id. at 458.

  188. [188]. Id.

  189. [189]. See id. at 451–58; see also Hustler Magazine, Inc. v. Falwell, 485 U.S. 46, 50–52 (1988).

  190. [190]. Cox Broad. Corp. v. Cohn, 420 U.S. 469 (1975).

  191. [191]. Fla. Star v. B.J.F., 491 U.S. 524, 527 (1989); Cox Broad. Corp., 420 U.S. at 472–74.

  192. [192]. Smith v. Daily Mail Publ’g Co., 443 U.S. 97, 99 (1979); Okla. Publ’g Co. v. Dist. Court, 430 U.S. 308, 309 (1977).

  193. [193]. Fla. Star, 491 U.S. at 541; Daily Mail Publ’g Co., 443 U.S. at 105–06; Okla. Publ’g Co., 430 U.S. at 311–12; Cox Broad. Corp., 420 U.S. at 496–97.

  194. [194]. Fla. Star, 491 U.S. at 534–35; Daily Mail Publ’g Co., 443 U.S. at 103; Okla. Publ’g Co., 430 U.S. at 311; Cox Broad. Corp., 420 U.S. at 495.

  195. [195]. See Curtis Publ’g Co. v. Butts, 388 U.S. 130, 154–55 (1967); N.Y. Times Co. v. Sullivan, 376 U.S. 254, 267–83 (1964).

  196. [196]. Gertz v. Robert Welch, Inc., 418 U.S. 323, 345–48 (1974).

  197. [197]. See Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 763 (1985).

  198. [198]. Pickering v. Bd. of Educ., 391 U.S. 563, 574 (1968).

  199. [199]. Id. at 568.

  200. [200]. Connick v. Myers, 461 U.S. 138, 147 (1983).

  201. [201]. Id.

  202. [202]. In using the word “private” here, I do not mean to suggest that only public discourse is of high First Amendment value. Private conversations about issues relevant to democracy are surely of very high First Amendment value as well. Here, “private” is meant simply as a reference to speech not of public concern.

  203. [203]. Reed v. Town of Gilbert, 707 F.3d 1057, 1080 (9th Cir. 2013) (Watford, J., dissenting), rev’d, 135 S. Ct. 2218 (2015).

  204. [204]. Lehman v. City of Shaker Heights, 418 U.S. 298, 319–20 (1974).

  205. [205]. Owen M. Fiss, Free Speech and Social Structure, 71 Iowa L. Rev. 1405, 1408 (1986).

  206. [206]. See generally Brandenburg v. Ohio, 395 U.S. 444 (1969); Dennis v. United States, 341 U.S. 494 (1951); Whitney v. California, 274 U.S. 357 (1927); Abrams v. United States, 250 U.S. 616 (1919).

  207. [207]. See generally NAACP v. Claiborne Hardware Co., 458 U.S. 886 (1982); N.Y. Times Co. v. Sullivan, 376 U.S. 254 (1964).

  208. [208]. See supra notes 198–201 and accompanying text.

  209. [209]. See generally Int’l Soc’y for Krishna Consciousness, Inc. v. Lee, 505 U.S. 672 (1992); Hague v. Comm. for Indus. Org., 307 U.S. 496 (1939).

  210. [210]. McCullen v. Coakley, 134 S. Ct. 2518, 2529 (2014).

  211. [211]. The rise of the Internet does not end the problem of scarcity. Certainly, speech opportunities on the Internet are not scarce, but certain forms of non-digital speech opportunities—notably the ability to place signs on public property visible to cars—remain uniquely valuable to some kinds of speakers, including notably insurgent political candidates.

  212. [212]. See Clark v. Cmty. for Creative Non-Violence, 468 U.S. 288, 293 (1984).

  213. [213]. See generally Thomas v. Chi. Park Dist., 534 U.S. 316 (2002).

  214. [214]. Sidewalks have long been classified by the Supreme Court as traditional public forums. See United States v. Grace, 461 U.S. 171, 177 (1983) (discussing how the Court considers sidewalks public forums and how “the government’s ability to permissibily restrict expressive conduct is very limited” in these areas). The classification of medians appears to be less settled, but decisions suggest that they too are public forums. See Satawa v. Macomb Cty. Rd. Comm’n, 689 F.3d 506, 520 (6th Cir. 2012) (“The Mound Road median . . . has features that invite the public to spend time there. It is more like a public park and therefore more likely . . . to qualify as a traditional public forum.”); Warren v. Fairfax Cty., 196 F.3d 186, 196–97 (4th Cir. 1999) (en banc). If a particular median were to be classified as only a limited or nonpublic forum, then state authority to draw subject matter (but not viewpoint) distinctions would certainly be enhanced, but even then it must justify such restrictions as reasonable. See Christian Legal Soc’y v. Martinez, 561 U.S. 661, 679 & n.11 (2010)(discussing how the Court allows restrictions on access to limited public forums so long as the access barrier is “reasonable and viewpoint neutral”). What remains unresolved is whether substantive preferences for political over say commercial speech qualify as reasonable under these rules.

  215. [215]. For a similar argument in a very different factual context, see Ashutosh Bhagwat, Parking at BART, or Economics & Its Discontents, 4 Green Bag 7, 10 & n.20 (2000).

  216. [216]. Cf. Citizens United v. FEC, 558 U.S. 310 (2010).

  217. [217]. Cf. Lochner v. New York, 198 U.S. 45, 75 (1905) (Holmes, J., dissenting).

  218. [218]. See generally Robert Post & Amanda Shanor, Adam Smith’s First Amendment, 128 Harv. L. Rev. F. 165 (2015).

  219. [219]. See supra notes 11–19 and accompanying text.

  220. [220]. City of Cincinnati v. Discovery Network, Inc., 507 U.S. 410 (1993).

  221. [221]. Id. at 413–14.

  222. [222]. Id. at 418.

  223. [223]. Id. at 425–28.

  224. [224]. See supra Part IV.

  225. [225]. Compare Christian Legal Soc’y v. Martinez, 561 U.S. 661, 679 & n.11 (2010) (discussing “limited public forums”), with Davenport v. Wash. Educ. Ass’n, 551 U.S. 177, 189 (2007) (discussing “nonpublic forum”).

  226. [226]. See Christian Legal Soc’y, 561 U.S. at 679 & n.11.

  227. [227]. Rosenberger v. Rector & Visitors of the Univ. of Va., 515 U.S. 819 (1995).

  228. [228]. Id. at 828–31, 845–46; see also Christian Legal Soc’y , 561 U.S. at 681 (stating that schools may “limit official student-group recognition to organizations comprising only students”).

  229. [229]. Perry Educ. Ass’n v. Perry Local Educators’ Ass’n, 460 U.S. 37, 50–55 (1983).

  230. [230]. Id. at 45–49.

  231. [231]. Compare Bd. of Educ. v. Pico, 457 U.S. 853 (1982) (striking down the decision by a school board to remove books from school libraries), with United States v. Am. Library Ass’n, 539 U.S. 194 (2003) (upholding statute requiring public libraries who receive federal funds to install software block materials harmful to minors). In both these cases, the Court notably failed to put together a majority opinion.

  232. [232]. The scarcity problem, of course, only applies to physical books and other materials. It does not require or justify content-based preferences regarding online materials, except subscription databases, where limited funds force choices.

  233. [233]. See generally Turner Broad. Sys., Inc. v. FCC (Turner I), 512 U.S. 622 (1994); Turner Broad. Sys., Inc. v. FCC (Turner II), 520 U.S. 180 (1997).

  234. [234]. See Turner II, 520 U.S. at 185; Turner I, 512 U.S. at 643–52.

  235. [235]. Turner I, 512 U.S. at 630–31.

  236. [236]. Id. at 645–46.

  237. [237]. Turner II, 520 U.S. at 189–225.

  238. [238]. Turner I, 512 U.S. at 674.

  239. [239]. Id. at 674–85 (O’Connor, J., dissenting).

  240. [240]. Turner Broad. Sys., Inc. v. FCC (Turner I), 512 U.S. 622, 676–78 (1994).

  241. [241]. It would of course violate the compelled speech doctrine for government to require private property owners to post political signs. See generally Wooley v. Maynard, 430 U.S. 705 (1977) (holding that New Hampshire could not require noncommercial license plates to read “Live Free or Die”). Prohibiting commercial signs, however, while permitting political ones (at the owner’s discretion) does not seem to raise any such concerns.

  242. [242]. Cantwell v. Connecticut, 310 U.S. 296, 311 (1940).

  243. [243]. Terminiello v. City of Chicago, 337 U.S. 1, 6 (1949); id. at 26 (Jackson, J., dissenting).

  244. [244]. Cohen v. California, 403 U.S. 15, 26 (1971).

  245. [245]. See supra Part II.B.

  246. [246]. Int’l Soc’y for Krishna Consciousness, Inc. v. Lee, 505 U.S. 672, 704–05 (1992) (Kennedy, J., concurring).

  247. [247]. Id. at 706–07.

  248. [248]. Berger v. City of Seattle, 569 F.3d 1029, 1050 (9th Cir. 2009).

  249. [249]. See id. at 1035 (“The Seattle Center.... attracts over ten million visitors annually.”).

  250. [250]. Norton v. City of Springfield, 768 F.3d 713, 714 (7th Cir. 2014).

  251. [251]. ACLU of Nevada v. City of Las Vegas, 466 F.3d 784, 787–88 (9th Cir. 2006).

  252. [252]. See Clatterbuck v. City of Charlottesville, 708 F.3d 549, 552 (4th Cir. 2013).

  253. [253]. See id. at 552.

  254. [254]. See supra Part II.B.

  255. [255]. See supra notes 243–54 and accompanying text.

  256. [256]. See supra text accompanying note 204.

  257. [257]. McCullen v. Coakley, 134 S. Ct. 2518, 2545–46 (2014) (Scalia, J., concurring in the judgment).

  258. [258]. Snyder v. Phelps, 562 U.S. 443, 456–59 (2011).

  259. [259]. I thank my colleague Alan Brownstein for this insight.

  260. [260]. My colleagues Vik Amar and Alan Brownstein have made this argument in greater detail in Alan Brownstein & Vikram David Amar, Death, Grief, and Freedom of Speech: Does the First Amendment Permit Protection Against the Harassment and Commandeering of Funeral Mourners?, 2010 Cardozo L. Rev. de novo 368, 370–79. I am grateful to them for the insight.

  261. [261]. I say “almost,” of course, because of the possibility I have just raised that we should perhaps effectuate such objections in highly vulnerable settings.

  262. [262]. See supra Part II.D.

  263. [263]. See supra Part II.D.

  264. [264]. See City of Renton v. Playtime Theatres, Inc., 475 U.S. 41, 51 (1986) (“[T]he location of adult theatres has a harmful effect on the area and [they] contribute to neighborhood blight....” (quoting Northend Cinema, Inc. v. City of Seattle, 585 P.2d 1153, 1156 (Wash. 1978) (en banc))).

  265. [265]. See generally Rienzi & Buck, supra note 105.

  266. [266]. See text accompanying notes 125–26.

  267. [267]. See supra notes 154–58 and accompanying text.

  268. [268]. See supra notes 159–62 and accompanying text.

  269. [269]. See generally Geoffrey R. Stone, Sex, Violence, and the First Amendment, 74 U. Chi. L. Rev. 1857 (2007).

  270. [270]. See City of Los Angeles v. Alameda Books, Inc., 535 U.S. 425, 445–46 (2002) (Kennedy, J., concurring in judgment).

  271. [271]. Bhagwat, supra note 215, at 8–9.

  272. [272]. See Bambauer, supra note 131, at 110.

  273. [273]. See generally Bhagwat, supra note 185 (arguing that specific facts should, under certain circumstances, be accorded a lower level of First Amendment protection).

  274. [274]. Sorrell v. IMS Health Inc., 564 U.S. 552, 570 (2011).

  275. [275]. See generally Geoffrey R. Stone, Restrictions of Speech Because of Its Content: The Peculiar Case of Subject-Matter Restrictions, 46 U. Chi. L. Rev. 81 (1978).

  276. [276]. Kreimer, supra note 40, at 1279–82.

  277. [277]. See generally Stone, supra note 275.

  278. [278]. Id. at 83.

  279. [279]. Id. at 108.

  280. [280]. Id. at 109–10.

  281. [281]. Id. at 110–11.

  282. [282]. Id. at 111.

  283. [283]. Geoffrey R. Stone, Restrictions of Speech Because of Its Content: The Peculiar Case of Subject-Matter Restrictions, 46 U. Chi. L. Rev. 81, 111–12 (1978).

  284. [284]. Id. at 112–14.

  285. [285]. See generally Kreimer, supra note 40.

  286. [286]. See generally id.

  287. [287]. Id. at 1278.

  288. [288]. Id. at 1279–83.

  289. [289]. Id. at 1278.

  290. [290]. Id. at 1284.

  291. [291]. Seth F. Kreimer, Good Enough for Government Work: Two Cheers for Content Neutrality, 16 U. Pa. J. Const. L. 1261, 1304–05 (2014).

  292. [292]. Id. at 1303–04.

  293. [293]. Id. at 1304–05.

  294. [294]. Id. at 1305–16.

  295. [295]. Stone, supra note 275, at 110–11.

  296. [296]. Id. at 112–14; Kreimer, supra note 40, at 1305.

  297. [297]. Reed v. Town of Gilbert, 135 S. Ct. 2218, 2234–36 (2015) (Breyer, J., concurring); id. at 2236–39 (Kagan, J., concurring).

  298. [298]. Id. at 2234 (Breyer, J., concurring).

  299. [299]. Id. at 2238 (Kagan, J., concurring).

  300. [300]. This is also why I must ultimately disagree with Marc Rohr’s suggestion that we recognize a category of permissible “de minimus content discrimination.” See generally Rohr, supra note 48. Rohr’s proposal is driven by many of the same concerns that drive my argument, but I believe that his proposal is sufficiently imprecise that it poses many of the risks identified by Stone and Kreimer.

  301. [301]. Cf. Pleasant Grove City v. Summum, 555 U.S. 460, 472–81 (2009) (upholding, under the government speech doctrine, a government’s decision to exclude a monument donated by one specific religious group from a public park, while including other religious monuments).

  302. [302]. Snyder v. Phelps, 562 U.S. 443, 458–61 (2011).

  303. [303]. Virginia v. Black, 538 U.S. 343, 367–68 (2003).

  304. [304]. Id. at 363.

  305. [305]. Id. at 381 (Souter, J., dissenting).

*

Martin Luther King, Jr. Professor of Law, University of California, Davis School of Law. B.A., 1986 Yale University. J.D., 1990 The University of Chicago. Contact: aabhagwat@ucdavis.edu. 



Thanks to participants at the 2015 Freedom of Expression Conference held at the Yale Law School for extremely helpful comments, and thanks especially to Lauren Henry for serving as discussant, and providing very valuable feedback (including the title). Thanks also to participants at the University of California, Davis School of Law faculty “Schmooze” for, as always, penetrating insights and perspectives.