The violent Charlottesville protests of August 12, 2017, raise one of the more vexing questions of First Amendment law: at what point does a given expression’s tendency to provoke a violent response in listeners justify government intervention against the speaker? Current doctrine provides no clear answer, and even suggests that such intervention may never be justified. This Essay argues that in some cases it is justified, and proposes a modified Brandenburg v. Ohio incitement standard to define when that is. Setting such a standard would remedy the existing asymmetry between the law’s treatment of speech that incites violence and speech that provokes it. It would protect speakers’ rights while providing a cognizable pre-violence point at which authorities could intervene without fear of violating the First Amendment.

Introduction

The Charlottesville Protests

THE weekend that has rendered “Charlottesville” a national byword for white supremacist agitation was planned in advance.[1] It included Friday, August 11’s surprise tiki-torch march on the University of Virginia’s Central Grounds[2] and the major gathering near the city’s downtown mall on August 12.[3] The focus of this Essay will be on the events of August 12.

Jason Kessler, the principal organizer of the “Unite the Right” rally,[4] applied in May 2017 for a permit to hold his event at the recently renamed Emancipation Park on Saturday, August 12.[5] The City of Charlottesville gave him the permit a few weeks later. The City subsequently granted permits to a number of opposition groups to hold assemblies of their own within a few blocks of the Park.[6] After an attempt by the City of Charlottesville to revoke Kessler’s permit, Judge Glenn Conrad of the Western District of Virginia granted Kessler an injunction, finding that Kessler’s claim that the revocation and modification of the permit was an impermissible content-based restriction on speech would likely succeed on the merits.[7]

Unite the Right was intended to begin at noon, but white nationalist protesters and counterprotesters began gathering near Emancipation Park around eight in the morning.[8] Among the attendees on both sides were people armed and armored with shields, helmets, and sticks. A contingent of self-proclaimed militiamen arrived as well, wearing camouflage and carrying pistols and rifles. They told reporters they were there to keep the peace. Some of the protesters openly carried guns too.[9] Still, at this point early in the morning, one firsthand observer described the atmosphere as “calm,”[10] with a group of twenty or thirty clergy and other counterprotesters linking arms and singing. Then, sometime around 10:30 a.m., skirmishes broke out between the groups. According to one police officer, part of the problem was that the protesters went back on a plan to enter the park from only one side, instead coming in from all directions and clashing directly with counterprotesters.[11] As a result some of the protesters and the police were behind metal barricades in the Park itself, while on Market Street in front of the Park, members of both sides directly confronted each other, reportedly with no police intervention.[12] As the groups neared each other, various members of the crowd exchanged chants and insults. The clergy sang “This Little Light of Mine,” while the white supremacists yelled “Our blood, our soil!”[13] Counterprotesters shouted “Fuck you, Nazis!” while the protesters replied “Fuck you, f——-s!”[14]

Shortly before 11:00 a.m., a large group of white nationalists approached the Park along Market Street. A group of counterprotesters formed a line across the street in order to block their path. The protesters then charged the line, swinging their sticks, punching, and spraying mace and pepper spray. A number of the counterprotesters fought back in the same manner. Participants threw rocks, water bottles, and balloons filled with paint, ink, or urine as “the air was filled with the sounds of fists and sticks against flesh.”[15] At 11:06 a.m., the City of Charlottesville and Albemarle County jointly declared a local emergency.[16] Twenty-two minutes later, Governor Terry McAuliffe declared a state of emergency across the entire Commonwealth.[17]

At the same time, around 11:20 or 11:30 a.m.,[18] the police declared the gathering to be an unlawful assembly,[19] and by about 11:40 a.m. were dispersing the crowds, with the white nationalists migrating toward McIntire Park to continue the demonstration.[20] Scattered fights and verbal exchanges continued in surrounding areas, but the main rally had been shut down before it was even supposed to begin.[21] Tragically, it was after the dispersal, when relative calm had returned to the city, that James Alex Fields, Jr., rammed his car into a crowd of pedestrians, killing Heather Heyer and injuring nineteen others.[22] Later that evening, two state police officers who were monitoring the city, Lieutenant H. Jay Cullen and Trooper Berke M.M. Bates, died when their helicopter crashed.[23]

*  *  *

There are many questions to be asked after a day like August 12. Perhaps the most fundamental, at least from a legal perspective, is the one Charlottesville Mayor Mike Signer asked on August 13: “How do you reconcile public safety and the First Amendment?”[24] The object of this Essay is to provide one incomplete answer to this question in the doctrinal area of First Amendment law known as the “heckler’s veto”[25] or “hostile audience”[26] problem.

The heckler’s veto or hostile audience problem arises when speech is met with an audience that is likely to turn violent on the speaker—in such a scenario, can the government shut down the speech, or must it allow the speaker to continue? This problem is vexing and active. Commentators point to recent events like University of California, Berkeley’s cancellation of a speech by Milo Yiannopoulos[27] and other incidents on college campuses[28] to decry the desecration of the freedom of speech,[29] while government institutions puzzle over how to manage security costs at controversial events consistent with the First Amendment.[30]

In this Essay I argue that in order to better balance public security with freedom of speech, courts should analyze the hostile audience problem in accordance with the template set out in the incitement test of Brandenburg v. Ohio.[31] The Brandenburg standard, discussed in more detail below,[32] holds that the government may not proscribe speech advocating violence or unlawful action unless “such advocacy is directed to inciting or producing imminent lawless action and is likely to incite or produce such action.”[33] The Brandenburg test, which has served as a durable and speech-protective First Amendment principle since it was first articulated, should serve as the model for a clearer standard in the domain of the heckler’s veto, as follows: the government may not proscribe violence-provoking expression unless it is directed to provoking imminent violence and is likely to produce violence.

In order to establish the viability of the proposed approach, the remainder of this Essay proceeds in two Parts. Part I traces three relevant free speech doctrines: the hostile audience problem, “fighting words,”[34] and the incitement test of Brandenburg. Part II argues that because of the current state of the heckler’s veto and fighting words doctrines, there is an asymmetry between the treatment of speech that is directed toward provoking a violent response in others and the treatment of speech that advocates the use of violence. The normative case for revising the present standard centers on the analytical invalidity of this asymmetry as well as practical concerns. Part II also looks to the three most relevant events of the Charlottesville unrest—Judge Conrad’s court order, the insults traded between protesters and counterprotesters on the day of the rally, and the police department’s declaration of unlawful assembly—in order to determine the extent to which the proposed standard might have influenced the course of events. In doing so, I aim to confront directly that most difficult question: how can public safety be reconciled with the First Amendment?

I. Tracing the Doctrines

First Amendment jurisprudence contains a multitude of overlapping yet distinct doctrines.[35] This Essay focuses on three in particular: the hostile audience problem, fighting words, and the incitement test of Brandenburg v. Ohio.

A. Hostile Audience Problem or the Heckler’s Veto

The idea of the heckler’s veto was first conceptualized by Professor Harry Kalven,[36] who described the problem as follows:

A speaker may threaten a breach of peace in two ways, either by inciting to violence or by irritating an audience so that it responds with violence. In the second case, in which the audience is hostile, a difficult issue is posed. Apart from the “fighting words” point of the Chaplinsky case, are there other circumstances in which the police, in order to keep tranquility, are entitled to arrest the speaker rather than the audience? The problem is a genuine puzzle either way it is decided. If the police can silence the speaker, the law in effect acknowledges a veto power in hecklers who can, by being hostile enough, get the law to silence any speaker of whom they do not approve. But the opposing view, that the police must go down with the speaker, has its own obvious difficulties, too.[37]

The heckler’s veto, which may also be classified as the “hostile audience” problem,[38] is indeed difficult, as evidenced by the divergent outcomes of the two canonical cases in the area: Terminiello v. Chicago[39] and Feiner v. New York.[40]

In Terminiello, the United States Supreme Court overturned the disorderly conduct violation[41] of the speaker, a polarizing public figure and recently defrocked Catholic priest.[42] Arthur Terminiello addressed a crowded hall filled with both supporters and detractors. The hall was surrounded on the outside by a much larger group of protesters, some of whom were throwing projectiles through windows and at police officers.[43] As he addressed the audience, Terminiello repeatedly referred to his detractors as “scum” and accused them of plotting to violently overthrow the United States, in addition to making numerous anti-Semitic remarks.[44]

Justice William O. Douglas, writing for a five-Justice majority, found that, although the Illinois state courts had held that Terminiello’s speech could be proscribed as “fighting words” under Chaplinsky v. New Hampshire,[45] because part of the jury instruction indicated that Terminiello could be found guilty if he was engaging in conduct that “stirs the public to anger, invites dispute, brings about a condition of unrest, or creates a disturbance,”[46] his conviction was constitutionally void.[47] Justice Douglas wrote, “[A] function of free speech under our system of government is to invite dispute. It may indeed best serve its high purpose when it induces a condition of unrest, creates dissatisfaction with conditions as they are, or even stirs people to anger.”[48]

Terminiello’s speech could not be censored “unless shown likely to produce a clear and present danger of a serious substantive evil that rises far above public inconvenience, annoyance, or unrest.”[49] Therefore, even if in fact Terminiello’s speech had been producing a clear and present danger, the possibility that he was convicted for merely stirring the public to anger and creating unrest rendered his penalization void.

Contrast this holding with the one in Feiner, decided two years later. Irving Feiner, a progressive college student who was speaking through a loudspeaker on a crowded street corner, made numerous inflammatory political statements.[50] He was surrounded by a restless crowd, among whom were people vocally threatening to attack him.[51] The police asked him to step down and stop speaking. He ignored them and continued talking until finally one officer arrested him.[52] Feiner was convicted of the misdemeanor of disorderly conduct and received a thirty-day jail sentence.[53] In this case a six-Justice majority found that Feiner’s speech was not protected. Chief Justice Fred M. Vinson, for the majority, acknowledged that the government “may not unduly suppress free communication of views . . . under the guise of conserving desirable conditions.”[54] Still, the Court found:

It is one thing to say that the police cannot be used as an instrument for the suppression of unpopular views, and another to say that, when as here the speaker passes the bounds of argument or persuasion and undertakes incitement to riot, they are powerless to prevent a breach of the peace.[55]

Justice Hugo Black, writing in dissent, was appalled by this holding, characterizing it as “a long step toward totalitarian authority.”[56] Black argued that rather than arresting the speaker, the police first should have made “all reasonable efforts to protect him[,]” including, if necessary, arresting hecklers who were threatening violence.[57] Only once such efforts were exhausted would it be acceptable for police to silence the speaker.

Feiner has never been overruled, but in the years since it was decided no case has followed it in upholding a speaker’s conviction.[58] Indeed, a number of post-Feiner cases make it appear as though Justice Black’s dissent has carried the day.[59] The rule for hostile audiences, therefore, appears to be that police and localities must make reasonable efforts to control the hecklers and protect the speaker, only stopping the speaker “if crowd control is impossible and a threat to breach of the peace imminent.”[60] The validity of this framing, however, is open to debate, as no Supreme Court case since Feiner has directly addressed in what circumstances it is acceptable to shut down speech in order to preserve security. Indeed, the direction of the Court has been to require localities to protect speakers regardless of cost.[61] Thus it is at a minimum doubtful whether lower courts and localities may follow Feiner and allow speech to be prevented on the basis of a hostile reaction. Some authorities go further and suggest that Feiner may no longer be good law at all.[62]

B. Fighting Words

In Chaplinsky v. New Hampshire, the Court gave its canonical definition of “fighting words”: words “which by their very utterance inflict injury or tend to incite an immediate breach of the peace.”[63] Such speech, along with “the lewd and obscene, the profane, [and] the libelous” is not protected because, the Court reasoned, it is of such low social value that its benefits are outweighed by the social interest in preserving order and morality.[64] The fighting words doctrine, it would seem, is a good answer to the question of the extent to which speech can be restricted because of the violent reaction it provokes in others: if it is a fighting word, then it is not entitled to constitutional protection, and therefore the government may restrict it in the interest of public safety.[65]

But, similar to what happened to Feiner,[66] since Chaplinsky was decided in 1942 the Court has not once upheld a conviction under the fighting words doctrine.[67] Indeed, the case law demonstrates a steady retreat from Chaplinsky’s self-assured classification of fighting words as an unprotected category.

In Cohen v. California, the Court assessed the scope of the fighting words doctrine as it applied to profane language printed on a jacket.[68] In finding that the message did not constitute fighting words (and more generally, was protected by the First Amendment), the Court put a series of new glosses on Chaplinsky, finding that, in order to be considered fighting words, the message on the jacket would need to be directed at an individual listener, and subject to the reasonable interpretation that the message was a “direct personal insult.”[69] Because the message on Paul Cohen’s jacket was not aimed at anyone in particular, and could not be interpreted as a direct insult, it lay outside the newly formulated fighting words doctrine.[70]

The year after Cohen, the Court decided four more cases that further call into question the continued relevance of the fighting words doctrine.[71] The first of these, Gooding v. Wilson, involved the following words spoken to a police officer at an antiwar rally: “White son of a bitch, I’ll kill you.”; “You son of a bitch, I’ll choke you to death.”; and “You son of a bitch, if you ever put your hands on me again, I’ll cut you all to pieces.”[72] The Court overturned the speaker’s conviction for violating a Georgia law making it a misdemeanor to use “opprobrious” language “tending to cause a breach of the peace,”[73] finding the statute overbroad and therefore facially unconstitutional under the First Amendment.[74] The other three cases were resolved similarly.[75]

In light of these and subsequent decisions, the extent to which the fighting words doctrine carves out any room for governments to regulate violence-provoking speech must be considered gravely in doubt.[76] If the fighting words doctrine is to serve as a limitation on how far speech can go before it is subject to regulation, then there is in effect almost no limit at all.

C. The Incitement Test of Brandenburg v. Ohio

In Brandenburg, a Ku Klux Klan leader was caught on film addressing a group of supporters.[77] In the racially charged speech, he made vague threats about taking “revengeance” on the U.S. government for its efforts to “suppress the white.”[78] As a result he was convicted under an Ohio statute criminalizing advocacy of “the duty, necessity, or propriety of crime, sabotage, violence, or unlawful methods of terrorism as a means of accomplishing industrial or political reform.”[79] The Court struck down the statute for violating the First Amendment, in the process formulating a new standard for advocacy of violence:

[T]he constitutional guarantees of free speech and free press do not permit a State to forbid or proscribe advocacy of the use of force or of law violation except where such advocacy is directed to inciting or producing imminent lawless action and is likely to incite or produce such action.[80]

Thus for violence-advocating speech to be proscribable, there must be intent (“directed to”), imminence, and likelihood of actual violence.[81] This kind of advocacy is to be distinguished from “abstract teaching [of] the moral propriety or even moral necessity” of violence.[82] Because Ohio’s statute did not make this important distinction, the Klansman’s conviction could not be upheld.[83]

Subsequent cases demonstrate how Brandenburg is applied. In Hess v. Indiana, for example, a group of antiwar protesters took to the streets, blocking traffic.[84] Law enforcement arrived and the protesters moved to the sidewalks on either side of the road, but as they did so one protester allegedly shouted, to no one in particular, “We’ll take the fucking street later.”[85] The speaker was arrested and convicted under a disorderly conduct statute for these words.[86] In reversing the conviction, the Court found that the statement “could be taken as counsel for present moderation” or, at worst, “nothing more than advocacy of illegal action at some indefinite future time.”[87]

NAACP v. Claiborne Hardware Co. took the imminence requirement a step further.[88] In that case, a civil rights activist urging other members of the movement to adhere to a boycott of white-owned businesses told a crowd of supporters, in the midst of a lengthy and passionate speech, “If we catch any of you going in any of them racist stores, we’re gonna break your damn neck.”[89] The Court, in applying Brandenburg to the statement, found that because the speaker did not actually authorize, ratify, or directly threaten violence, and his rhetorical appeals did not in fact incite lawless action, his speech was protected.[90]

Hess and Claiborne emphasize the importance of the imminence requirement.[91] In Hess, the speaker’s use of temporally indefinite language (“later”) was enough for the Court to find that imminence was not met,[92] and in Claiborne the lack of actual violence immediately following the speech, especially in the context of an emotional plea to supporters, meant the expression was protected.[93] Thus, there is a high threshold for imminence under Brandenburg, resulting in a very speaker-protective doctrine.[94]

II. Revising the Violence-Provoking Speech Standard and Applying It to the Events in Charlottesville

A. A Better Approach

What separates Brandenburg v. Ohio from whatever remains of Feiner v. New York and Chaplinsky v. New Hampshire is the clarity of the standard enunciated. While the Brandenburg test even protects speakers who believe in violence and advocate for it in an abstract or rhetorical manner, it also clearly allows for restrictions at the extremes.[95] For example, if a radical anarchist instructed a loyal crowd to pick up whatever projectiles they could find and immediately launch them at passersby, the police could certainly intervene to stop the speech; intent, imminence, and likelihood are easily met. Thus Brandenburg serves the dual purposes of ensuring wide latitude for speakers while providing adequate notice of when speech may cross the boundaries of constitutional protection.

As the above discussion in Sections I.A and I.B demonstrates, there is at present no comparably clear outer limit for violence-provoking speech. In theory, after Cohen v. California, a speaker who was deliberately trying to provoke a violent response from a listener could be subject to speech restriction.[96] But in the absence of further elaboration, it is difficult to see how such actual intent could be proven, especially given the unlikelihood that a speaker would truly intend for another person to attack her.

This difficulty suggests that, as with Brandenburg, the appropriate prong to focus on is imminence.[97] Focusing on imminence allows governments to make assessments about public security in a way that will allow for as much speech as possible, because even highly inflammatory speech will be allowed to continue up to the point where violence is imminent. Thus, the proposed hostile audience standard is that violence-provoking speech is to be protected unless it is directed to[98] provoking imminent violence and is likely to precipitate violence.

Imagine, for instance, that a neo-Nazi group entered a predominately Jewish neighborhood whose residents included a high number of Holocaust survivors and descendants of Holocaust victims.[99] The demonstrators wear Nazi regalia, carry symbols of Nazism, and direct deeply wounding and insulting anti-Semitic language at residents. For the purposes of this hypothetical, let us imagine that these insults are as heinous as possible, referencing real, brutal crimes that the listeners themselves experienced, or that directly affected the listeners’ family members. It is possible, in such a scenario, that even listeners with peaceful intentions and dispositions might temporarily be unable to control themselves, and react with hostility. Under the proposed standard, the police could intervene and shut down the Nazis’ speech. This is so because the expression is directed to[100] provoking imminent violence and such violence is very likely to occur. Imminence and likelihood are present given the directness of the communication as well as the broader context of extreme antipathy between speaker and listener. Although the police in such circumstances would of course be justified in restraining the hostile audience from attacking the demonstrators, they would also be allowed to prevent the demonstrators from continuing their violence-provoking speech.

B. Would the New Standard Have Made a Difference?

1. Judge Conrad’s Preliminary Injunction

If the new violence-provoking speech standard were correctly applied, it would have no effect on the outcome of enjoining the revocation of the protest permit in Kessler v. City of Charlottesville.[101] As the court in that case noted, the city revoked Kessler’s permit, but not the counterprotesters’, which provided evidence that the “eleventh-hour” revocation was a based on the content of Kessler’s message.[102] City Councilmembers’ anti-Kessler social media posts provided further support for this conclusion.[103] As a content-based restriction, the permit modification would be subject to strict scrutiny.[104] Based on the factual circumstances, the City had very little hope of showing that its action was narrowly tailored to further a compelling state interest.

With the benefit of hindsight, Judge Conrad’s finding that “there is no evidence to support the notion that many thousands of individuals are likely to attend the demonstration”[105] seems to ignore certain realities, but his legal conclusions were surely sound. The proposed hostile audience standard does not change that in the slightest. Courts should still scrutinize government actions that appear content-based, and the high imminence threshold of the standard means that ex ante suppression through a permit change would remain a virtual nonstarter.[106]

2. Verbal Exchanges on August 12

Among the words the protesters were reported to have spoken at the rally were: “Our blood, our soil!”[107] “Fuck you, f——-s!”; “Go the fuck back to Africa!”; “Fuck you, n——-!”; and “Dylann Roof was a hero!”[108] The counterprotesters, in turn, are recorded as saying: “Fuck you, Nazis!” and “Go the fuck home!”[109] Any of these statements would almost certainly have been considered fighting words by the Chaplinsky-era Court, which upheld a conviction for someone who said “damned Fascist.”[110] Indeed, the white supremacists’ statements in particular seem the very definition of violence-provoking speech. Nonetheless, any statute that attempts to proscribe this speech would have to be very narrowly tailored so that (a) it could be applied only to truly violence-provoking (under the proposed intent-imminence-likelihood standard)—and therefore unprotected—speech[111] and (b) it did not discriminate based on content within the unprotected category.[112] A revitalized violence-provoking speech standard makes a statute that could bar and permit state intervention in these circumstances theoretically possible, while under current doctrine this possibility is doubtful.

On the other hand, although arrest under a statute would likely raise problems under numerous areas of free speech law, police intervention through removal or detention of speakers without any arrests could be possible upon the utterance of some of these insults.[113] The fact that in Charlottesville many of these verbal attacks were immediately followed by physical clashes suggests that violence was in fact imminent and likely. Thus, upon hearing these exchanges and observing the scene, under the proposed standard, police might have intervened sooner and prevented physical confrontation.

3. Declaration of Unlawful Assembly

On August 12, after street violence had been under way for about an hour, the police declared that the gathering at Emancipation Park and the surrounding streets was unlawful under Section 18.2-406 of the Virginia Code.[114] This section reads, in full:

Whenever three or more persons assembled share the common intent to advance some lawful or unlawful purpose by the commission of an act or acts of unlawful force or violence likely to jeopardize seriously public safety, peace or order, and the assembly actually tends to inspire persons of ordinary courage with well-grounded fear of serious and immediate breaches of public safety, peace or order, then such assembly is an unlawful assembly. Every person who participates in any unlawful assembly shall be guilty of a Class 1 misdemeanor. If any such person carried, at the time of his participation in an unlawful assembly, any firearm or other deadly or dangerous weapon, he shall be guilty of a Class 5 felony.[115]

Interestingly enough, the current version of this statute was amended after a 1971 Virginia Supreme Court case held the previous version unconstitutionally overbroad because it included “assemblies that pose no clear and present danger.”[116] Thus the current version, enacted in 1975, added a likelihood requirement (“likely to jeopardize seriously[117] public safety”) and, with the language “tends to inspire persons of ordinary courage with well-grounded fear of serious and immediate breaches of public safety,” an imminence requirement as well. As such, Virginia’s unlawful assembly statute appears to conform with the constitutional requirements of Brandenburg, and indeed the Eastern District of Virginia later found that “[o]n its face, Section 18.2-406 does not impermissibly infringe upon . . . first amendment rights.”[118]

Under this Essay’s proposed approach, this statute might have been invoked earlier, since the police declared unlawful assembly after violence had already broken out. The current wording of the statute does not require the police to wait for the commission of illegal acts for an assembly to be declared unlawful, but the violence-provoking standard would allow the police to consider both the actions and the words spoken by those assembled, in the context of the imminence and likelihood of violence, to determine when to break up the gathering.

On August 12, this might have happened as soon as protesters and counterprotesters faced each other directly on Market Street and exchanged insults. The words themselves would not have been enough to show imminence or likelihood of violence, but the physical nearing of the groups and the possession of weapons, combined with the virulence of the epithets, could have provided a basis for declaration of unlawful assembly. Setting out a violence-provoking standard would give police the confidence that in such extreme cases, intervention could happen sooner rather than later. It would also provide notice to speakers that both actions and language of escalation could lead to a premature end to their expressive activity.

4. A Road Not Taken: Disorderly Conduct

Virginia’s disorderly conduct statute, Virginia Code Section 18.2-415,[119] presents another avenue that Charlottesville police might have taken to prevent violence.[120] Some of its language tracks the proposed violence-provoking speech test.[121] Although it prohibits “conduct having a direct tendency to cause acts of violence” by those at whom the conduct is directed, the statute also makes clear that this conduct “shall not be deemed to include the utterance or display of any words.”[122] Virginia courts have confirmed that any application of Section 18.2-415 to speech would be unconstitutional.[123] Under the proposed violence-provoking speech standard, a disorderly conduct statute that included the intent, imminence, and likelihood requirements could constitutionally be applied to words in addition to actions.

Conclusion

Revitalizing a test whereby violence-provoking speech could, under the appropriate circumstances, be unprotected would not likely effect a revolution in the operation of the First Amendment. But it would provide law enforcement and speakers with a workable set of boundaries allowing for prevention of needless violence.

When racist agitation turns to civil unrest and ultimately to tragedy, it is rational to ask what can be done to prevent it from happening again. The answer outlined here is far from complete, but it is at least a first step toward reconciling public safety with the freedom of speech.