Concept
Unprotected Categories and R.A.V.
Content-based regulation of protected expression
Unprotected Categories and R.A.V.
Speech in a historic unprotected category may be regulated because of its content, but the category does not license every content line a legislature might draw inside it. R.A.V. v. City of St. Paul settled that point. St. Paul punished the display of a symbol that arouses anger, alarm, or resentment in others on the basis of race, color, creed, religion, or gender, and the state courts confined the ordinance to fighting words. Chaplinsky v. New Hampshire defined those words as ones “which by their very utterance inflict injury or tend to incite an immediate breach of the peace,” and presumed that they play little or no part in the exposition of ideas and so fall outside the First Amendment umbrella.1 The Supreme Court assumed that everything the ordinance reached was proscribable and still held it facially invalid, because the ordinance prohibited otherwise permitted speech solely on the basis of the subjects the speech addresses.2
Incitement, fighting words, obscenity, defamation, and the other historic categories are defined by what the speech says or does. A law reaching them is therefore content based, yet it escapes the strict scrutiny that governs content-based laws elsewhere in this chapter.3 The Court justifies the exception on the ground that such speech offers so little as a step toward truth that any benefit is clearly outweighed by the social interest in order and morality.4 United States v. Alvarez restates the list: advocacy intended and likely to incite imminent lawless action, obscenity, defamation, speech integral to criminal conduct, fighting words, child pornography, fraud, true threats, and speech presenting a grave and imminent threat that government has the power to prevent.5
The list does not grow through ad hoc balancing. United States v. Stevens rejected the argument that a category loses protection whenever its social costs outweigh its value, and the Court declined to add depictions of animal cruelty.6 Hate speech is likewise absent from the list, so a law aimed at it is content based like any other and must satisfy rigorous First Amendment standards.7
Before R.A.V., the conventional view was that government had broad latitude inside these categories and that a law there needed only to satisfy rational basis review.8 That view left a gap. If a category’s low value is the whole reason government may regulate it, a legislature can still use the category as a vehicle for hostility toward an idea, by punishing only the members that carry a disfavored message. R.A.V. gave libel as its example: government may proscribe libel, but it may not proscribe only libel critical of the government.9 The Court did not go on to treat every content line inside a category as invalid. It recognized exceptions, and Virginia v. Black later applied one of them to a cross-burning statute.10 The subsections below state the rule R.A.V. announced and then mark where it stops.
The R.A.V. Rule
Government may proscribe a category such as fighting words, but it may not make a further content distinction within the category that is unrelated to the reason the category is proscribable.11 R.A.V. announced that rule when it struck down St. Paul’s ordinance. This section follows the case in the order the Court decided it: the burning cross and the ordinance, the state court’s narrowing construction, the rule and its reasons, its application to the ordinance’s text, and the objections of the four Justices who joined the judgment but rejected the majority’s reasoning.
In the predawn hours of June 21, 1990, several teenagers allegedly taped broken chair legs into a crude cross and burned it inside the fenced yard of a black family who lived across the street from the house where the petitioner was staying.12 Many laws could have punished that conduct, but one provision under which St. Paul charged the petitioner, then a juvenile, was its Bias-Motivated Crime Ordinance. The ordinance made it disorderly conduct to place on public or private property a symbol or object, including a burning cross or Nazi swastika, that one knows or has reasonable grounds to know “arouses anger, alarm or resentment in others on the basis of race, color, creed, religion or gender.”13 The petitioner moved to dismiss on the ground that the ordinance was substantially overbroad and impermissibly content based. The trial court granted the motion, and the Minnesota Supreme Court reversed.14
The state court answered the overbreadth charge by reading the phrase “arouses anger, alarm or resentment in others” to reach only conduct amounting to fighting words, meaning conduct that itself inflicts injury or tends to incite immediate violence, so that the ordinance covered only expression the First Amendment does not protect.15 It also held the ordinance narrowly tailored to a compelling interest in protecting the community against bias-motivated threats to public safety and order.16 The Supreme Court considered itself bound by that construction and accepted that the ordinance reaches only fighting words within the meaning of Chaplinsky v. New Hampshire.17 The petitioner urged the Court to narrow the Chaplinsky formulation and strike the ordinance as overbroad. The Court found it unnecessary to decide that question and assumed, for the sake of argument, that all the expression the ordinance reached was proscribable.18
That assumption removes the easy answer and gives the case its structure. If the ordinance reached only unprotected speech, government might seem free to punish that speech however it likes, and the Court had to explain why the First Amendment still mattered. The explanation starts from two points: content-based regulations are presumptively invalid, and a long tradition permits restrictions on a few limited categories, among them obscenity, defamation, and fighting words, which carry so little value as a step to truth that any benefit is clearly outweighed by the social interest in order and morality.19
Statements that such categories lie outside the First Amendment, the Court said, must be taken in context. They mean that these areas of speech can be regulated because of their constitutionally proscribable content. They do not mean that the categories are “entirely invisible to the Constitution,” so that they may be made vehicles for content discrimination unrelated to their distinctively proscribable content.20 The Court’s illustration was libel: government may proscribe libel, but it may not proscribe only libel critical of the government. A city council likewise may not prohibit only those legally obscene works that criticize the city government.21
The reason runs through what makes a category regulable. A category is proscribable because of one feature of the speech, and that feature supplies the government’s justification for reaching it. A city that punishes obscene works only when they criticize the city has stopped acting on that feature and begun acting on a second one, the idea expressed, which the justification for banning obscenity does not touch. The Court offered two parallels: government may regulate a noisy sound truck for its noise but not for the ideas it broadcasts, and a flag burned in violation of an ordinance against outdoor fires may be punished while a flag burned in violation of an ordinance against dishonoring the flag may not.22
The Court chose the name of the limit with care. The First Amendment imposes a “content discrimination” limitation on a state’s prohibition of proscribable speech and no “underinclusiveness” limitation as such. A state may, for example, prohibit obscenity only in certain media or markets, because that prohibition would be underinclusive without discriminating on the basis of content.23 A law offends the rule only when the line it draws inside the category follows what the speech says.
The ordinance’s own terms show subject-matter discrimination. The phrase “arouses anger, alarm or resentment in others” had been limited to fighting words, but the remaining terms confine the ordinance to fighting words that insult or provoke violence “on the basis of race, color, creed, religion or gender.”27 Abusive invective, however vicious, stayed permissible unless it was addressed to one of those disfavored topics. Speakers who used fighting words to express hostility on the basis of political affiliation, union membership, or homosexuality were not covered, so the city had singled out particular topics from the many on which a speaker might use fighting words.28
In practice, the Court added, the ordinance went beyond content discrimination to viewpoint discrimination.29 Odious racial epithets would be forbidden to proponents of every view, but fighting words that do not invoke race, color, creed, religion, or gender, such as aspersions on a person’s mother, could seemingly be used freely on the placards of those arguing in favor of tolerance and equality, while their opponents could not use words that insult on those grounds.30 The sign example shows the asymmetry. A placard declaring all “anti-Catholic bigots” misbegotten is permitted, and one declaring all “papists” misbegotten is not, because only the second insults on the basis of religion.31 A rule keyed to the group an insult targets burdens the speaker who attacks the group and spares the speaker who attacks the group’s critics. The debate is policed unevenly even though no sentence of the ordinance names a side.
Because the ordinance sorted by subject and viewpoint, the question became whether it fit an exception to the bar on content discrimination, and the answer turned on why fighting words are proscribable. They fall outside the First Amendment because their content embodies a particularly intolerable and socially unnecessary mode of expressing whatever idea the speaker wishes to convey.32 A city that wanted its line to track that reason could have selected for prohibition only the fighting words that communicate ideas in a threatening manner as opposed to a merely obnoxious one. St. Paul instead proscribed fighting words of whatever manner that communicate messages of racial, gender, or religious intolerance, and selectivity of that sort creates the possibility that the city is seeking to handicap particular ideas.33 The city’s own statements turned the possibility into a certainty. It conceded that the ordinance applies only to racial, religious, or gender-specific symbols such as a burning cross or Nazi swastika, and it had argued in juvenile court that the content of the cross’s message is what the ordinance attempts to legislate.34 The ordinance therefore fit none of the exceptions the Court had described, and the next subsections take those exceptions up one at a time.
St. Paul’s last defense was that the content line was narrowly tailored to compelling interests, namely ensuring the basic human rights of members of groups historically subjected to discrimination, including their right to live in peace where they wish. The Court did not doubt that these interests are compelling or that the ordinance promotes them.35 The dispositive question was whether content discrimination is reasonably necessary to achieve them, and it is not: an ordinance not limited to the favored topics would have precisely the same beneficial effect.36 An adequate content-neutral substitute also suggests that the content line serves some other purpose. With a general ban available, the only interest the content limitation distinctively serves is the city council’s special hostility toward the particular biases it singled out. The council may express that hostility, but not through unique limitations on speakers who disagree.37
A drafter who tries to stay within the rule meets a second hazard. A narrow fighting-words law is likely to be an impermissible content-based restriction because it outlaws some fighting words but not others, while a broad one is likely to be invalidated as vague or overbroad. That pairing makes it difficult for any fighting-words law to survive.38
Four Justices agreed that the judgment should be reversed and rejected the majority’s reasoning. Justice White, joined by Justices Blackmun and O’Connor and, in part, by Justice Stevens, would have decided the case on the ground the petitioner pressed: the ordinance is fatally overbroad.39 The state court had read the ordinance to reach expression that by its very utterance causes “anger, alarm or resentment,” yet the fighting-words cases hold that such generalized reactions do not strip expression of constitutional protection. An ordinance that also criminalizes expression causing only hurt feelings, offense, or resentment is therefore overbroad on its face.40 That was the question the majority sidestepped when it assumed the ordinance reached only proscribable speech.
White’s deeper objection was to the majority’s rule. The categories of proscribable speech are content based and are unprotected because their expressive content is worthless or of minimal value. It is therefore inconsistent to hold that government may proscribe an entire category because its content is evil but may not treat a subset of the category differently, since the content of the subset is by definition worthless as well.41 He also charged that the majority renounced strict scrutiny, because on its view a narrowly drawn content-based ordinance could never pass if its object could be accomplished by banning a wider category of speech.42 The practical complaint was that the rule invites the continuation of the very conduct the city wants to stop until the city cures the underbreadth with a catchall phrase.43 The majority replied that the concurrences had attacked a position of their own invention, because the rule it announced bars content discrimination within a category and does not bar underinclusiveness as such.44
Justice Blackmun’s separate concurrence added the worry that, by forbidding a state to regulate speech that causes great harm unless it also regulates speech that does not, the Court seemed to abandon the categorical approach and to relax the scrutiny applied to content-based laws. He joined the judgment because he agreed with Justice White that the ordinance reaches beyond fighting words.45 Justice Stevens agreed that the ordinance is overbroad for the reasons in Part II of White’s opinion, but he wrote separately because the allure of absolute principles had, in his view, skewed the analysis of both the majority and Justice White.46 It is too simple, he argued, to declare expression protected or unprotected, or a regulation content based or content neutral.47 Had the ordinance not been overbroad, he would have upheld it as a selective subject-matter regulation of proscribable speech.48
Stevens read the ordinance as regulating speech on the basis of the harm it causes. He answered the majority’s sign example by observing that the ordinance would bar a Muslim from hoisting a sign claiming that all Catholics were misbegotten, just as it would bar a Catholic from hoisting a similar sign about Muslims.49 The ordinance, in his image, bans punches below the belt by either party.50 The majority called the harm reading wordplay: what separates the anger, fear, and sense of dishonor the ordinance addresses from the same reactions to other fighting words is that a distinctive idea, conveyed by a distinctive message, causes them.51 The rule is narrower than the majority’s rhetoric suggests, because the Court recognized exceptions, including one for distinctions based on the very reasons the class of speech is proscribable, and the next subsections take them up in turn.52
Same-Reason Distinctions
The first exception tells a legislature when it may draw a content line inside a proscribable category without creating the danger the R.A.V. rule guards against. The Court stated it in these terms: “When the basis for the content discrimination consists entirely of the very reason the entire class of speech at issue is proscribable, no significant danger of idea or viewpoint discrimination exists.”55 The rule bars lines that are unrelated to the reason the category is proscribable, and this exception covers the related kind: a line that tracks that reason and nothing else. A state therefore need not ban all obscenity or all threats in order to ban some, so long as it selects the portion by the feature that justified regulating the class at all.56
The exception follows from what the general ban on content discrimination protects. Content discrimination matters because it “raises the specter that the Government may effectively drive certain ideas or viewpoints from the marketplace.”57 Inside a proscribable category, government already holds a justification for reaching the whole class, and that justification has nothing to do with any idea the speech carries. The Court drew the consequence directly: “Such a reason, having been adjudged neutral enough to support exclusion of the entire class of speech from First Amendment protection, is also neutral enough to form the basis of distinction within the class.”58 If a reason can justify removing every member of a class, ranking the members by how strongly that reason applies gives the legislature no new motive. A line built from some other content feature has no such backing, because nothing explains why that feature matters except the legislature’s attitude toward the message.
What counts as the reason depends on the category. Fighting words are excluded from First Amendment protection because their content embodies a particularly intolerable and socially unnecessary mode of expressing whatever idea the speaker wishes to convey, so mode is the feature a same-reason line among fighting words has to follow.59 The Court’s illustrations from other categories show the exception working and show where it stops.
Obscenity supplies the first illustration. A state might choose to prohibit only the obscenity that is most patently offensive in its prurience, meaning the most lascivious displays of sexual activity, and that choice is permitted because the line follows the very quality that makes obscenity regulable.60 The same state may not prohibit only the obscenity that includes offensive political messages.61 Both statutes address one category and both sort its members by content, yet they differ in what they sort by. The first ranks works by degree of prurience, which asks nothing about any idea the work expresses. The second ranks them by political message, a feature that plays no part in why obscenity may be regulated, so its only work is to single out a viewpoint.
Threats of violence give the second illustration. Federal law can criminalize only those threats directed against the President, because the reasons threats fall outside the First Amendment (protecting individuals from the fear of violence, from the disruption that fear engenders, and from the possibility that the threatened violence will occur) have special force when applied to the person of the President.62 The subclass is chosen by the intensity of the very harms that define the category. The government may not, however, criminalize only those threats against the President that mention his policy on aid to inner cities.63 The speaker, the category, and the target are unchanged, and the added fact is a reference to policy. That reference increases none of the fear, disruption, or likelihood of violence, so the line now follows a message and the exception no longer covers it.
Commercial speech gives a third illustration: a state may regulate price advertising in one industry but not in others on the strength of the risk of fraud, which is one of the characteristics that justifies depriving commercial speech of full First Amendment protection.64
St. Paul’s ordinance fell on the wrong side of this line, and the contrast shows what the exception demands. The Court said the ordinance “assuredly does not fall within the exception for content discrimination based on the very reasons why the particular class of speech at issue (here, fighting words) is proscribable.”65 St. Paul had not singled out an especially offensive mode of expression. It had not, for example, selected for prohibition only those fighting words that communicate ideas in a threatening as opposed to a merely obnoxious manner, which would have followed the reason the category is proscribable. It had instead proscribed fighting words of whatever manner that communicate messages of racial, gender, or religious intolerance.66 Selectivity of that kind creates the possibility that the city is seeking to handicap particular ideas, and the Court found that possibility enough by itself to make the ordinance presumptively invalid.67 A ban on threatening fighting words would have sorted by mode. The ordinance actually enacted sorted by message.
Justice White pressed the obvious objection: the President example and St. Paul’s ordinance rest on the same logic. Just as Congress may decide that threats against the President entail more severe consequences than other threats, he reasoned, a city council may decide that threats based on the target’s race, religion, or gender cause more severe harm than other threats.68 The answer lies in what measures the line. The President provision picks out the instances of a threat in which the category’s own harms are greatest, while St. Paul’s ordinance picked out fighting words by the topic of the invective, the racial, gender, or religious intolerance of the message. A legislature’s judgment that one subclass is worse qualifies for the exception only when the worseness is measured in the category’s own terms. When the measure is the subject or idea the speech conveys, the judgment is an attitude toward that idea.
The exception also determines the level of review. The Court did not hold that every content-based distinction within a proscribable category receives strict scrutiny. Where there is no threat of viewpoint discrimination, strict scrutiny does not apply.69 A same-reason line therefore faces no presumption of invalidity and needs no compelling-interest showing, while a line that fits no exception is presumptively invalid and goes on to the necessity inquiry. When a law regulates only part of a category, identify the reason the whole category is proscribable, identify the feature the law uses to pick its subset, and ask whether the two are the same. The intent-to-intimidate ban in Virginia v. Black, taken up in its own subsection below, was upheld on the strength of this exception.70
Secondary Effects and Conduct
The second exception applies when a subclass of proscribable speech comes with harms that exist whatever message the speech carries. The Court said that a content-defined subclass may be treated differently when it happens to be associated with particular secondary effects of the speech, so that the regulation is “justified without reference to the content of the . . . speech.”73 Its illustration was a state that permits all obscene live performances except those involving minors.74 That line follows the participation of minors, a feature whose harms the state can describe without saying anything about which ideas a performance expresses, so drawing it gives officials no way to favor or handicap a message. The exception borrows its vocabulary from Renton, which the earlier discussion of content-neutral justifications covered in detail. There the formula decided which tier of review governed a zoning ordinance; here it decides whether a line drawn inside a category the government may already reach is suspect at all.
The same discussion supplies the limit that defeated St. Paul’s argument. St. Paul contended that its ordinance targeted secondary effects, because its purpose was to protect against the victimization of people who are vulnerable by virtue of membership in a historically disfavored group.75 The Court assumed for argument that an ordinance that completely proscribes a category could be directed at secondary effects, and still held that St. Paul’s was not. It quoted Boos v. Barry for the point that “[l]isteners’ reactions to speech are not the type of ‘secondary effects’ we referred to in Renton.”76 A footnote closed the remaining gap. Even if the categories St. Paul chose were the ones most likely to provoke violence, the only reason they would be especially correlated with violence is that they convey a particularly odious message. The chain of causation then runs through the persuasive force of the expression, and the harm is a primary effect of the speech.77
A bar-exam answer on a hate-speech ordinance defended as protection for victims should therefore ask where the harm originates. If it would disappear were the audience unable to understand the message, it is a reaction to the message and falls outside the exception, as with St. Paul’s ordinance. If it attaches to the participants or the setting, as with performances involving minors, the exception can apply.
The Court added a related point about conduct. Words can violate laws directed at conduct and not at speech: a law against treason is violated by telling the enemy the nation’s defense secrets, so a content-based subcategory of a proscribable class can be swept up incidentally in a statute aimed at conduct.78 The Court’s example was Title VII of the Civil Rights Act, whose general prohibition on sex discrimination in employment can be violated by sexually derogatory fighting words.79 The reason follows from what the employer is liable for. Liability rests on the discriminatory act in the workplace, and speech is one means of committing it, so the law selects no speaker because of the idea expressed. The Court put the limit in one sentence: where the government does not target conduct on the basis of its expressive content, acts are not shielded from regulation merely because they express a discriminatory idea or philosophy.80 St. Paul’s ordinance lies on the other side of that sentence, because it punished a display only when the display carried a message of bias.
A conduct label does not settle the matter, because the question is what triggers liability. In Holder v. Humanitarian Law Project the Government argued that the material-support statute was a conduct regulation reviewable under the lenient O’Brien standard. The Court disagreed, because whether the plaintiffs could speak to the designated groups depended on what they said. It pointed to Cohen v. California, where a generally applicable breach-of-the-peace law reached a jacket bearing an epithet only because of the message the jacket communicated, and the Court applied more rigorous scrutiny.81 Wisconsin v. Mitchell shows the other result. The Court unanimously upheld a longer sentence for an aggravated battery committed because of the victim’s race and rejected the argument that R.A.V.’s viewpoint principle forbids aggravating the penalty for bias-motivated conduct. The decision distinguished R.A.V. as a case about a viewpoint-selective law aimed expressly at otherwise unprotected words or symbols.82
The concurring Justices attacked this exception directly, and their critique marks where the doctrine is contested. Justice White observed that there is a simple explanation for the Court’s eagerness to craft an exception to its new rule: under the general rule the Court applies, Title VII hostile-work-environment claims would suddenly be unconstitutional.83 The majority’s answer is that the conduct exception depends on what the government targets, and a law that singles out speech for its message gets no help from it. An answer on this point should state the exception as the majority framed it and treat White’s objection as a challenge to how readily it can be invoked.
No Risk of Suppression
The residual exception appears in the same passage that states the other two. After naming the same-reason and secondary-effects bases for a content line inside a proscribable category, the Court observed that there may be other bases as well. It added that to validate selectivity where totally proscribable speech is at issue, it may not even be necessary to identify any particular neutral basis, so long as the nature of the content discrimination leaves “no realistic possibility that official suppression of ideas is afoot.”86 This third exception changes the question a court asks. The first two exceptions ask what neutral reason supports the line, while this one asks whether the line could be turned against an idea at all. When the answer is plainly no, the legislature owes no further explanation.
The exception follows from why the R.A.V. rule exists. The rule polices lines inside a category because a legislature that may reach the whole category can use a subset to handicap a disfavored idea, and a neutral basis is the evidence that this has not happened. Evidence matters only where there is something to prove. If a line sorts speech by a feature that no official could connect to any message, nobody can use it to favor one side of a debate, and demanding a neutral basis would force the legislature to justify a distinction that threatens nothing. The Court’s own illustration shows how little ground the exception covers: it said it could not “think of any First Amendment interest that would stand in the way of a State’s prohibiting only those obscene motion pictures with blue-eyed actresses.”87
Obscene films fall in a category a state may ban entirely, so the state could prohibit all of them. A performer’s eye color expresses no idea, belongs to no viewpoint, and marks no group of speakers whose position in a public debate the state might want to weaken. The narrower ban therefore gives officials no means of handicapping a message. The illustration is deliberately absurd, and the absurdity carries the lesson: the exception is satisfied only when the feature that selects the subset lies so far from ideas that suppression is not a realistic worry. An ordinary content line, one that follows a topic or a position, will not clear that bar.
St. Paul’s ordinance sits at the opposite pole. It reached only fighting words that insult or provoke violence on the basis of race, color, creed, religion, or gender, and abusive invective, however vicious, remained permissible unless it was addressed to one of those specified topics. Speakers who used fighting words to express hostility on the basis of political affiliation, union membership, or homosexuality were not covered.88 Each of those lines follows subject matter, and a line that follows subject matter lets a city handicap the ideas that travel under the favored or disfavored subjects. The Court held that the ordinance did not fall within any more general exception for selectivity that is beyond the suspicion of official suppression of ideas, and the city’s own statements in the case gave ample basis for that suspicion.89 St. Paul conceded that the ordinance applied only to racial, religious, or gender-specific symbols such as a burning cross or a Nazi swastika, and it had argued below that the content of the cross’s message was what the ordinance attempted to legislate.90
When the exception does apply, the line is tolerated because the category is already proscribable and no idea is at risk. Subject to this limitation, the Court said, regulation of fighting words, like regulation of noisy speech, may address some offensive instances and leave other, equally offensive, instances alone.91 That sentence fits the earlier point that R.A.V. limits content discrimination within a category and sets no general limit on underinclusiveness. An answer on a new statute should therefore ask what feature the law uses to select its subset, and then ask whether anyone could use that feature to handicap an idea. A line keyed to a topic, a viewpoint, or the message a symbol conveys fails the second question, so the exception is unavailable and the law carries the presumption of invalidity.
Virginia v. Black Contrast
Virginia v. Black shows that the R.A.V. rule does not condemn every content line drawn inside a category of proscribable speech. The Court held that a state may ban cross burning carried out with the intent to intimidate, while the provision of the Virginia statute treating any cross burning as prima facie evidence of intent to intimidate rendered the statute unconstitutional in its current form.94 One statute thus produced a line that survived and a line that failed, which makes the case a compact test of the exceptions in the preceding subsections. A student who remembered only R.A.V.’s headline, that government may not sort fighting words by message, would predict that any statute aimed at cross burning must fall. Black explains why that prediction fails for the intent-to-intimidate ban and why the statute still lost on a different feature.
Virginia’s statute made it unlawful “with the intent of intimidating any person or group of persons, to burn, or cause to be burned, a cross on the property of another, a highway or other public place,” and made the offense a Class 6 felony.95 A second sentence provided: “Any such burning of a cross shall be prima facie evidence of an intent to intimidate a person or group of persons.”96 Prima facie evidence suffices to establish a fact unless the opposing party rebuts it, so the second sentence invited the act of burning to stand in for the intent the first sentence required. Three respondents were convicted separately, and the Virginia Supreme Court held the statute facially unconstitutional on two grounds. It found the statute analytically indistinguishable from the St. Paul ordinance because it selectively chose cross burning for its distinctive message, and it found the prima facie provision overbroad because it raised the probability of prosecutions that chill protected speech.97 The Supreme Court treated the two features separately and reached opposite results on them.
The Court began by conceding that cross burning is symbolic expression, since the burner chooses the cross because it carries the message the burner wants to send, and it then held that this fact does not resolve the constitutional question.98 Government may regulate true threats, which are statements by which a speaker means to communicate a serious expression of an intent to commit an act of unlawful violence to a particular individual or group, and intimidation in the constitutionally proscribable sense is a type of true threat.99 The category exists because a prohibition on true threats protects individuals from the fear of violence, from the disruption that fear engenders, and from the possibility that the threatened violence will occur.100 Each of those harms flows from what the recipient reasonably expects to happen, so each exists whatever the recipient thinks of the speaker’s politics. Cross burning done with intent to intimidate therefore sits in a category that government may reach because of its content, and the R.A.V. question becomes whether Virginia could punish this one kind of true threat while leaving others alone.
The Court answered yes, because the choice of cross burning fits the exception for lines drawn on the very reason the whole class is proscribable. R.A.V. had not held that the First Amendment prohibits every form of content discrimination within a proscribable area, and it had approved a ban on only those threats directed at the President and a ban on only the most patently offensive obscenity.101 Virginia could outlaw cross burning done with intent to intimidate because burning a cross is “a particularly virulent form of intimidation,” and instead of prohibiting all intimidating messages the state could regulate this subset “in light of cross burning’s long and pernicious history as a signal of impending violence.”102 A state may likewise prohibit only those forms of intimidation “most likely to inspire fear of bodily harm.”103 The line tracks the category’s own harm, because a signal with a history of preceding violence carries that harm in an unusually high degree. The statute also differs from St. Paul’s ordinance because it does not single out only speech directed toward one of the specified disfavored topics. It reaches a cross burned to intimidate whether the victim was chosen for race, gender, or religion or for political affiliation, union membership, or homosexuality, and the Court noted that cross burners do not in fact direct their intimidating conduct solely at racial or religious minorities.104
The dispute over this application was sharp. Justice Souter, joined by Justices Kennedy and Ginsburg, agreed with the majority that the statute makes a content-based distinction within the category of intimidating expression, but he would have held the virulence exception inapplicable.105 In his account, threats against the President are not identified by any message that accompanies them, so that line picks out special risks and no special message, while a content-based ban on cross burning may be a subtle effort to suppress the message of white supremacy that even nonthreatening cross burning broadcasts.106 Because no R.A.V. exception applied, he would have required the statute to survive strict scrutiny, which it could not, since a content-neutral ban on intimidation would achieve the same object without singling out content.107 The majority answered with history, since a burning cross has long signaled impending violence, and with the intent element, which confines the ban to burnings meant to place the victim in fear of bodily harm.108 The disagreement shows where a challenger to a subcategory ban will aim: at whether the feature that selects the subset is the category’s own harm or the idea that travels with it.
The prima facie provision failed for a reason that grows out of the same logic. The ban is valid because it reaches intimidation, and the provision as construed in the jury instruction “strips away the very reason why a State may ban cross burning with the intent to intimidate.”109 Virginia’s own court had said that the act of burning a cross alone, with no evidence of intent to intimidate, would suffice for arrest and prosecution and would defeat a motion to strike the evidence at the end of the Commonwealth’s case-in-chief.110 A jury could therefore convict in every cross-burning case in which the defendant chose not to put on a defense. Even when a defendant like Black did present one, the provision made a finding of intent to intimidate more likely whatever the particular facts.111
The harm lies in what the act itself can mean. A burning cross may be intimidation, but it may also be a statement of ideology or a symbol of group solidarity. The Court acknowledged that burning a cross at a political rally would almost certainly be protected expression, and cross burnings also appear in films and plays.112 The provision made no effort to separate a cross burned at a public rally from one burned on a neighbor’s lawn, or one directed at an individual from one directed at like-minded believers.113 That blurring risks prosecution, and possibly conviction, of a person engaged only in lawful political speech at the core of the First Amendment, so the provision “would create an unacceptable risk of the suppression of ideas.”114 The anger or hatred that most viewers feel at a burning cross does not justify banning all cross burnings, and the Court concluded: “The First Amendment does not permit such a shortcut.”115
A plurality of four Justices reached that conclusion about the provision, and Justice Souter’s opinion, joined by two others, took a similar position.116 Justice Scalia, joined by Justice Thomas, dissented on this point. He argued that overbreadth turns on whether people engaged in protected speech can actually be convicted, and that the class the plurality imagined (those who burn a cross in public view without intent to intimidate, are charged anyway, and refuse to present a defense) was too small to make the statute substantially overbroad.117 Black’s conviction could not stand because the jury instruction relied on the provision, and the judgment as to Elliott and O’Mara was vacated and remanded, because the Virginia Supreme Court had not yet decided whether the provision was severable from the rest of the statute or what it meant.118
The surrounding principle explains why the intent element carries so much weight. Offensive speech, including symbols of hate, stays protected unless it falls within a proscribable category such as true threats. When the National Socialist Party of America announced a peaceful demonstration in Skokie, Illinois, a town with many survivors of Nazi concentration camps, the village’s attempts to stop it by injunction and by new ordinances failed in court, because government cannot suppress a speaker for the reaction of the audience.119 In the Seventh Circuit’s decision in Collin v. Smith, the village did not contend that violence would follow, and the court rejected the argument that the psychic trauma of residents could justify criminalizing protected expression in advance.120 Black marks the point where a symbol of hate loses protection, which is when it is used to place a victim in fear of bodily harm. Courts may resolve swastika displays the same way, by separating those made with intent to intimidate from those made without it.121 Counterman v. Colorado later described the true-threat test as subjective and held that recklessness is a sufficient mental state, so an intent-to-intimidate element meets the requirement with room to spare.
A new statute that regulates only part of a proscribable category can be tested with the two questions Black answered. The first asks what feature selects the subset and whether that feature is the category’s own harm, here the fear of bodily harm, or the idea the speech carries. The second asks whether the statute still requires the government to prove that feature in each case, or lets the act alone carry the conviction. A line that tracks the category’s harm and requires proof of it, as the intent-to-intimidate ban does, satisfies both questions. A line that sorts by the speaker’s disfavored topic, or a presumption that makes proof of the harm unnecessary, fails one of them.