Concept
Fighting Words
Regulation of unprotected expression
Fighting Words
Fighting words differ from incitement in the direction of the danger. In the incitement cases, the worry is that an audience will follow the speaker into lawlessness. Here the worry runs the other way: that the person addressed will answer the speaker with violence.1 The Court has built two doctrines on that concern. One holds that “fighting words,” meaning speech directed at another person and likely to provoke a violent response, fall outside the First Amendment. The other, a series of hostile-audience cases taken up below, asks when a speaker may be punished because of the way a crowd reacts.2
Chaplinsky v. New Hampshire placed insulting or “fighting” words among the narrow classes of speech whose punishment raises no constitutional problem. The Court reasoned that such utterances are no essential part of any exposition of ideas and have so slight a social value as a step to truth that any benefit from them is clearly outweighed by the social interest in order and morality. “Resort to epithets or personal abuse,” it concluded, “is not in any proper sense communication of information or opinion safeguarded by the Constitution.”3
The category has survived, but it rarely decides a case for the government. Labeling speech as fighting words has not been the dispositive ground of decision in many Supreme Court cases since Chaplinsky, and the Court generally has found breach-of-the-peace and disorderly-conduct statutes defended on that rationale to be facially overbroad.4 In overturning fighting-words convictions, the Court has used three techniques. It has narrowed the category to speech directed at another person and likely to produce a violent response. It has found laws prohibiting fighting words unconstitutionally vague or overbroad. And it has treated laws that prohibit only some fighting words, such as expressions of hate based on race or gender, as impermissible content-based restrictions.5 The first technique shapes the definition, taken up in The Chaplinsky Rule. The second explains why drafting decides these cases, the subject of Narrow Statutes Required. The third returns in the hate-speech discussion later in the chapter.
Together the techniques squeeze legislatures from both sides. A narrow law that outlaws some fighting words but not others invites invalidation as a content-based restriction, and a broad law invites invalidation for vagueness or overbreadth.6 The Court has paid close attention to the line between content-based and content-neutral laws, so a statute that singles out some fighting words for punishment receives that scrutiny.7 A fighting-words problem therefore asks two questions: do the words fit the narrow definition, and is the statute drawn tightly enough to survive?
The Chaplinsky Rule
Chaplinsky begins from the premise that free speech is not absolute “at all times and under all circumstances.” Certain “well-defined and narrowly limited classes of speech” may be prevented and punished without raising any constitutional problem, and the Court listed among them “the insulting or ‘fighting’ words — those which by their very utterance inflict injury or tend to incite an immediate breach of the peace.”8
The facts supplied the test case. Chaplinsky, a Jehovah’s Witness, was convicted in Rochester, New Hampshire, under a statute forbidding any person to “address any offensive, derisive or annoying word to any other person who is lawfully in any street or other public place,” or to call that person “by any offensive or derisive name.”9 After a disturbance at a busy intersection, a traffic officer started walking him to the police station. On the way they met Marshal Bowering, who had been told that a riot was under way and who repeated an earlier warning to Chaplinsky. Chaplinsky then addressed to Bowering the words that appear in the complaint: “damned racketeer” and “damned Fascist.”10
Read literally, the statute could reach any annoying remark. The New Hampshire Supreme Court confined it. Its purpose, the state court said, was to preserve the public peace, so that no words were “forbidden except such as have a direct tendency to cause acts of violence by the person to whom, individually, the remark is addressed.”11 Derisive and annoying words fell within the statute “only when they have this characteristic of plainly tending to excite the addressee to a breach of the peace.” As construed, the statute prohibited “the face-to-face words plainly likely to cause a breach of the peace by the addressee,” and the test was objective: “what men of common intelligence would understand would be words likely to cause an average addressee to fight.”12 Even without that gloss, an earlier decision had construed the statute as aimed at conduct whose direct tendency was to provoke the person addressed to violence.13
The Supreme Court accepted that reading and affirmed. Argument was unnecessary, it said, to show that “damned racketeer” and “damned Fascist” are “epithets likely to provoke the average person to retaliation, and thereby cause a breach of the peace.” The state court’s refusal to admit evidence of provocation, or evidence bearing on the truth or falsity of the remarks, raised no constitutional objection, and whether such facts supplied a defense or mitigation was for the state court to decide. The statute, “on its face and as applied,” did not contravene the Fourteenth Amendment.14
The narrowing construction did the constitutional work, as the contrast with Gooding v. Wilson shows. There the Court voided a Georgia statute against “opprobrious words or abusive language, tending to cause a breach of the peace” because the Georgia courts had not limited it to fighting words as Chaplinsky defined them.15 The Narrow Statutes Required subsection returns to the case.
Chaplinsky framed the test objectively, but the cases applying it have focused on the circumstances in which the words were used. Instead of asking only whether the words would cause an average addressee to fight, courts ask whether the language was used under circumstances that were personally provocative and created an actual threat of imminent disorder.16 Context therefore carries most of the weight. No words automatically fall into the category, and the same words may qualify in a volatile confrontation and not in an ordinary conversation.17 Under the objective formulation, the measure is the average addressee, so one listener’s unusual temper neither expands nor shrinks the category. Objective measures of this kind recur in other unprotected categories. In Counterman, the trial court convicted a defendant under a Colorado statute that made it unlawful to “repeatedly communicate with another person” in order to cause “a reasonable person to suffer emotional distress,” and it applied an objective standard under which a reasonable person would find the messages threatening.18 The True Threats section takes up that case.
The doctrine concerns words directed at a person, and that limit has teeth. In Texas v. Johnson, the defendant was convicted under a statute that prohibited desecrating a flag in a manner he “knows will seriously offend one or more persons likely to observe or discover [such] action,” and the government argued that his conduct fell within the fighting-words exception.19 Justice Brennan answered for the Court: “No reasonable onlooker would have regarded Johnson’s generalized expression of dissatisfaction with the policies of the Federal Government as a direct personal insult or an invitation to exchange fisticuffs.”20 The State’s interest in maintaining order, the Court concluded, was not implicated on those facts.21 Critics ask whether an exception narrowed to invitations to a breach of the peace gives too much protection to listeners who threaten to answer provocation with their fists, and whether it rests on an outdated and gendered honor culture.22
Chaplinsky’s definition has two parts, and only the second still does much work. Words that “by their very utterance inflict injury” would let a state punish speech because it wounds the listener. Words that “tend to incite an immediate breach of the peace” address the risk of a fight. Cohen v. California shows why the first part could not carry independent weight. Cohen was convicted of disturbing the peace by “offensive conduct” for wearing a jacket bearing a vulgar anti-draft slogan in a Los Angeles courthouse. Justice Harlan reversed, refusing to let government excise scurrilous expression from public discourse. “[O]ne man’s vulgarity is another’s lyric,” he reasoned, speech often serves an emotive value, and banning a manner of expression risks suppressing the ideas themselves.23 The slogan was directed at no particular person, so it was neither a personal epithet nor an invitation to fight.24 Later cases accordingly rest the category on the breach-of-the-peace rationale, as Johnson’s focus on “an invitation to exchange fisticuffs” and on the State’s interest in maintaining order illustrates.25 A student who asks only whether the words were vulgar or hurtful is applying the discarded branch. The question is whether they invited a fight.
Hostile Audiences
Fighting words and incitement both locate the danger in the speech itself. A third pattern locates it in the audience: the speaker’s message may be lawful, and the threat of violence comes from listeners who dislike it. This is the problem of the hostile audience, or the heckler’s veto. If a breach of the peace is threatened, whom should the police arrest, the speaker or the crowd?30 The government’s usual route to the speaker runs through the fighting-words label, and that route requires words directed at a particular person. The argument failed in Johnson for that reason: the claim that flag burning was likely to provoke a violent response from onlookers could not turn speech aimed at no one into fighting words.31
The cornerstone case is Terminiello v. City of Chicago. After a disturbance, Terminiello was convicted under a breach-of-the-peace statute that the trial judge construed to include speech that “stirs the public to anger, invites dispute, brings about a condition of unrest, or creates a disturbance.”32 By a vote of five to four, the Court reversed. A conviction resting on that ground was defective, because speech is protected “unless shown likely to produce a clear and present danger of a serious substantive evil that rises far above public inconvenience, annoyance or unrest.” Justice Douglas explained that “[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.”33 The defect lay in the jury charge, which described a great deal of effective political speech, and it is the same overbreadth defect that returns in Narrow Statutes Required. Because the reversal rested on the charge, Terminiello left open what the police may do when a crowd’s violence is real, and later cases supplied the answer.
Cantwell v. Connecticut had pointed the same way. The Court found “no assault or threatening of bodily harm, no truculent bearing, no intentional discourtesy, no personal abuse,” only “an effort to persuade a willing listener” to buy a book or contribute money in the interest of what Cantwell “conceived to be true religion.”34 Together the cases establish that a speaker cannot be silenced merely because of the reactions of a hostile audience.35
Feiner v. New York tested that principle. A street speaker ignored police requests to stop while the crowd grew restless and one onlooker threatened to remove him personally. The Court affirmed his conviction, deeming him to have been inciting a riot.36
Later decisions refused to extend Feiner to peaceful demonstrators. In Edwards v. South Carolina, the Court reversed breach-of-the-peace convictions of students who marched to the State House grounds carrying protest signs. Onlookers gathered, but no one turned violent, and police protection was sufficient to meet any foreseeable disorder.37 In Cox v. Louisiana, the fear of violence rested on onlookers across the street, none shown to have threatened anyone.38 In Gregory v. City of Chicago, marchers walked to the mayor’s home while angry onlookers threatened them, and the police ordered the marchers to leave. The Court unanimously reversed the disorderly-conduct convictions, because the marchers were engaged in lawful political protest.39
What separates Feiner from these cases is the finding about the speaker. In Feiner, the Court concluded that the speaker had deliberately incited the crowd and then defied the police. The marchers in Edwards, Cox, and Gregory were lawful demonstrators who threatened no one, and the danger came from onlookers whom the police were positioned to control. Generalized fears of disorder cannot justify breaking up a lawful demonstration, and the police must handle the crowd.40
Nevertheless, Feiner has increasingly been read narrowly. When a speaker intentionally provokes a hostile reaction and imminent disorder is probable, the speech is not protected and the police may move against the speaker. A speaker who is only engaged in protected expression keeps the protection, and the police must move against the crowd.41 One question stays open. When the police cannot control the crowd and violence is imminent, the doctrine is unclear, and in that emergency the speaker usually will be stopped, whether or not he should be.42
Narrow Statutes Required
A statute that punishes speech as fighting words survives a facial attack only if its text, or an authoritative state-court construction, confines it to the unprotected category. Gooding v. Wilson supplies the example. The Georgia statute punished anyone who, “without provocation,” used toward another, in his presence, “opprobrious words or abusive language, tending to cause a breach of the peace.” The defendant had told police officers “I’ll kill you” and “I’ll choke you to death.”48 The Court nonetheless set aside the defendant’s words and examined the statute as Georgia’s courts had construed it. They had used “breach of the peace” to cover merely offensive language, so the statute swept too broadly.49 Because the Georgia courts had not limited it to fighting words as Chaplinsky defined them, the statute made it a “breach of the peace” merely to speak words offensive to some who hear them.50
The doctrine at work is overbreadth. A law is overbroad if it regulates substantially more speech than the Constitution allows government to regulate.51 Overbreadth analysis does not ask whether the challenger’s own speech is protected. It strikes the statute down entirely because the statute might be applied to others, not before the Court, whose activities are protected, and the law stays unenforceable until a legislature rewrites it or a properly authorized court narrows it.52 The doctrine thus creates a special First Amendment exception to the usual rule that a litigant may not assert the constitutional rights of third parties, so an individual whose own conduct could constitutionally be punished may challenge the law on its face.53 That is why Gooding’s threats did not defeat his claim.
Chaplinsky shows the other side. The New Hampshire statute also spoke broadly, yet the state supreme court had confined it to face-to-face words with a direct tendency to cause violence by the person addressed, and the conviction stood on that reading.54 Both statutes reached beyond fighting words on their face. What differed was whether the state’s courts had already told the statute where to stop.
A narrowing construction rescues a statute only if it actually removes protected speech from its reach. The Court has struck legislation when the state court’s construction failed to cure the overbreadth problem, as it did in Lewis v. New Orleans.55 Gooding’s technique of invalidating a statute for vagueness and overbreadth, and so avoiding the need to define fighting words with particularity, has been employed frequently since.56 Convictions in Rosenfeld v. New Jersey and Brown v. Oklahoma were vacated in light of Gooding. The same defect doomed the ordinance in City of Houston v. Hill, which made it a crime to “oppose, molest, abuse or interrupt any policeman in the execution of his duty.” The Court held it substantially overbroad.
Vagueness is a different objection from overbreadth. All laws regulating conduct can be challenged under the due process vagueness doctrine, but courts are particularly troubled by vague laws that restrict speech, out of concern that they will chill constitutionally protected speech, which is “delicate and vulnerable, as well as supremely precious in our society.”57 Overbreadth asks how much protected speech a statute reaches, however clear its terms. Courts often reach both objections against the same fighting-words statute, and a broad law will probably fall on one ground or the other.58
R.A.V. v. City of St. Paul displays both problems in a single case. The St. Paul ordinance, as interpreted by the state court, was limited to fighting words. Four Justices concurring in the judgment would nonetheless have held it overbroad, because it also made criminal expressive conduct that causes only hurt feelings, offense, or resentment, which the First Amendment protects.59 Justice Scalia, writing for five Justices, instead held the ordinance facially unconstitutional because it discriminated on the basis of speech content and was underinclusive.60 R.A.V. does not forbid every distinction within a category. When the basis for the distinction consists entirely of the very reason the whole class of speech is proscribable, no significant danger of idea or viewpoint discrimination exists, and a state might, for example, prohibit only the obscenity that is most patently offensive in its prurience. It may not prohibit only the obscenity that includes offensive political messages.61 A statute drawn around hostile messages falls on the wrong side of that line, and the hate-speech section below takes up the rule.
When a court does ask whether particular words were fighting words, it examines the actual circumstances. In Lewis, Justice Powell, concurring, observed that it was highly improbable that the words used would provoke a physical encounter “between the middle aged woman who spoke them and the police officer in whose presence they were uttered.”62 The more the inquiry turns on such facts, the closer fighting words comes to the clear-and-present-danger approach traced earlier in the chapter.
Because the Court so often decides these cases on the breadth of the statute, it has had little occasion to say what fighting words are. Overbreadth has, arguably, eclipsed the doctrine, which survives as a recognized category but rarely supplies the ground of decision.63 The link between offense and overbreadth carries into the next section, which takes up profane, hurtful, and disparaging speech that stays protected.