Concept
Flag Burning Cases
Regulation of expressive conduct
Flag Burning Cases
Burning the American flag as political protest is protected expression, and the flag cases are the Court’s clearest example of a law that fails the gatekeeping question. In United States v. Eichman, the Court struck down a federal statute that prohibited burning or otherwise desecrating the flag, except to dispose of a worn or soiled one, because the statute was aimed at suppressing the communicative impact of expressive conduct. Government may still stop imminent lawless action, and it may still enforce laws whose interests have nothing to do with the message. It may not punish a person for the message a burning flag sends.
The draft-card case came out the other way, and the difference shows why. O’Brien claimed, as the flag burners later would, that his burning was protected because he did it in demonstration against the war and the draft.1 The Court upheld his conviction under the four-part test: a regulation is justified if it lies within the government’s constitutional power, furthers an important or substantial interest, rests on an interest unrelated to the suppression of expression, and restricts expression no more than is essential to that interest.2 The interest was keeping registration certificates available as proof that a registrant had done what the law required, and that interest exists whether or not anyone reads a burning as protest.3 The lower courts had rejected his challenge as well. The District Court held that it was not competent to inquire into Congress’s motives and that the Amendment was a reasonable exercise of the power to raise armies, and the Court of Appeals affirmed the conviction on a different statutory ground.4
A flag statute stands on different ground, and O’Brien said as much. It set apart the case in which the government’s interest in regulating conduct “arises in some measure because the communication allegedly integral to the conduct is itself thought to be harmful.” It pointed to Stromberg, which struck down a ban on displaying any flag as a sign of opposition to organized government, and it explained that a statute aimed at suppressing communication cannot be sustained as a regulation of noncommunicative conduct.5 A flag law belongs with Stromberg. The government values the flag as a symbol, so the harm a flag law targets appears only when someone’s treatment of the flag communicates something. An interest of that kind is related to suppression, and the case moves to the second track of the gatekeeping question. O’Brien no longer applies, and the law must be analyzed under general First Amendment principles, which here means the exacting scrutiny that content-based restrictions receive.6 The same physical act, burning an object, can land on either track. The government’s reason for punishing it decides which.
The subsections below follow the cases in the order they reached the Court: Johnson, then Congress’s reply and Eichman, then the laws that still reach a person who burns a flag, and finally cross burning and bias-motivated crimes, where the line between punishing an act and punishing an idea does more work.
Texas v. Johnson
In Texas v. Johnson, Gregory Lee Johnson burned an American flag in front of Dallas City Hall at the end of a protest march held during the 1984 Republican National Convention. Texas convicted him under a statute forbidding desecration of a venerated object, which it defined to include physically mistreating a flag in a way the actor knows will seriously offend one or more persons likely to observe or discover the act. Liability therefore turned on how onlookers were likely to react. The Texas Court of Criminal Appeals reversed the conviction, and the Supreme Court affirmed, holding the statute unconstitutional as applied to Johnson. The Court addressed only that claim, because a prosecution of someone who had not engaged in expressive conduct would pose a different case and the narrower ground was enough.7
The Court then worked through the gatekeeping questions in order. The first, whether the burning was expressive, took a sentence. Texas conceded that it was, and the Court called the concession prudent: the burning ended a political demonstration held while the Republican Party renominated a President, so its political character was “both intentional and overwhelmingly apparent.” Both parts of the Spence inquiry were met, and the case turned on the second question, whether Texas’s reasons for punishing the burning were related to the suppression of expression. Texas offered two reasons. The Court held the first inapplicable on this record and the second related to suppression.
The peace interest failed on the facts. No disturbance occurred or threatened to occur because of the burning, and Texas admitted that no actual breach took place. The State pointed to disorder earlier in the march, but it brought no charges for that conduct and did not show that disorder was a likely response to Johnson’s own act. Its only trial evidence of audience reaction came from people who were seriously offended, so its position came to this: an audience that takes serious offense is necessarily likely to disturb the peace.8 The Court refused to presume it. If offense alone made disorder likely, a hostile audience would hold a veto over any message, and the more people disliked an idea, the stronger the State’s case for silencing it. The cases instead require a look at the actual circumstances, asking under Brandenburg v. Ohio whether the expression “is directed to inciting or producing imminent lawless action and is likely to incite or produce such action.”9
Johnson’s burning was not “fighting words” either. That narrow class, described in Chaplinsky v. New Hampshire, covers words likely to provoke the average person to retaliation and so cause a breach of the peace, and no reasonable onlooker would have taken Johnson’s generalized dissatisfaction with federal policy as a direct personal insult or an invitation to exchange fisticuffs. The Court added that Texas already had a statute against breaches of the peace, which suggested that the State did not need to punish this burning to keep order, and it stated that nothing in its holding forbids a State to prevent imminent lawless action.10 Once the facts failed to implicate the peace interest, it dropped out of the analysis.
In Johnson the record showed offended onlookers and nothing more. On the rally facts the record shows every element of the incitement standard, which is why the Court took care to leave States free to stop imminent lawless action.
The second interest, preserving the flag as a symbol of nationhood and national unity, was implicated, and it decided the case. Texas feared that burning a flag would lead people to believe either that the flag does not stand for national unity or that such unity does not exist. The Court observed that those concerns arise only when a person’s treatment of the flag communicates some message.11 The structure of the statute confirmed the point. Johnson would not have been convicted had he burned the flag to dispose of a dirty or torn one, a method federal law recognizes as proper and one to which Texas had no objection. The law therefore protected the flag only against mistreatment that would cause serious offense, so whether Johnson violated it depended on the likely communicative impact of his act. Texas argued that the knowledge element looked to the actor’s intent and not to the crowd’s reaction, but the Court found that distinction too precious to matter, since both kinds of statutes aim to protect onlookers from being offended by the ideas a prohibited act expresses.12
Because the unity interest was related to expression, the case fell outside O’Brien. It is the situation O’Brien itself set apart, where the government’s interest in regulating conduct arises because the communication integral to the conduct is thought to be harmful.13 The Court relied on Boos v. Barry, which treated the emotive impact of speech on its audience as part of the speech’s content and not as a secondary effect, to classify the restriction as content-based, and it applied “the most exacting scrutiny.” That scrutiny fell on a prosecution for political expression, since Johnson was punished for voicing dissatisfaction with the policies of his country, which sits at the core of the First Amendment’s values.
Texas argued that the flag is a symbol with a determinate range of meanings, so that conduct tending to cast doubt on nationhood or national unity conveys a harmful message that may be prohibited. The Court answered with the principle on which the opinion rests: “If there is a bedrock principle underlying the First Amendment, it is that the government may not prohibit the expression of an idea simply because society finds the idea itself offensive or disagreeable.”14 The harm Texas identified is a harm of message, so the principle applies directly. The Court had never recognized a flag exception. In Street v. New York it held that a State may not criminally punish a person for words critical of the flag, and in West Virginia v. Barnette it held that government may not compel conduct evincing respect for the flag, because no official may prescribe what is orthodox in politics.
Texas tried to confine those cases to words. Even if its interest could not support punishing speech critical of the flag, it argued, the interest could support forbidding the flag’s outright destruction. The Court replied that the verbal or nonverbal form of expression is of no moment when the conduct is expressive and the regulation is related to expression, and that Barnette and Spence themselves protected expressive conduct. A second reply exposed what Texas’s position would allow. A State that could forbid burning wherever it endangers the flag’s symbolic role, yet permit it wherever burning promotes that role, as when a person ceremoniously burns a dirty flag, would let the flag serve as a symbol in only one direction. The Court had never held that government may ensure that a symbol expresses only one view of itself, and in Schacht v. United States it struck down a federal proviso that let an actor wear a military uniform only if the portrayal did not tend to discredit the armed force.15 A rule protecting only some symbols also leads into territory with no discernible boundaries. State flags, copies of the Presidential seal, and the Constitution might all claim the same protection, and a court could decide among them only by consulting its own political preferences. The Court therefore declined to create a separate category for the flag.16
The Court did not deny the flag a special place. Texas’s ends were not the objection, only its means: government has a legitimate interest in preserving the flag as an unalloyed symbol of the country, and it serves that interest through precatory regulations describing proper treatment. Encouraging proper treatment differs from criminally punishing a person for burning a flag as political protest, because national unity is something officials may foster by persuasion and example, and the question is whether compulsion is a permissible means.17 The answer to a flag burning is more speech: waving one’s own flag, saluting the one that burns, or giving its remains a respectful burial. “We do not consecrate the flag by punishing its desecration, for in doing so we dilute the freedom that this cherished emblem represents.”
The holding reaches laws tied to the message and leaves laws tied to other harms alone. Texas charged only desecration, with its message-dependent element, and it did not charge theft, trespass, disorderly conduct, or arson, although Johnson had received the flag from a fellow protester who took it from a flagpole outside one of the buildings the march targeted. The subsection on neutral laws returns to what such a charge would look like.
Four Justices dissented. Chief Justice Rehnquist, joined by Justices White and O’Connor, wrote that for more than 200 years the flag has been the visible symbol of the Nation without representing any party or philosophy, so it is not simply another idea in the marketplace of ideas. He compared the burning to the fighting words of Chaplinsky, said it was no essential part of any exposition of ideas and tended to incite a breach of the peace, and called it the equivalent of an inarticulate grunt or roar.18 Justice Stevens, dissenting separately, treated the flag’s symbolic value (freedom, equal opportunity, religious tolerance, and national unity) as a significant and legitimate interest. He saw the prosecution as aimed at the method Johnson chose and not at his message, and he compared the burning to spray-painting the same message on the Lincoln Memorial, which no one would doubt government may forbid.19 Justice Kennedy supplied the fifth vote and wrote separately that the Constitution compelled a result he did not like, one that follows from the fact that the flag protects even those who hold it in contempt.20 The majority and the dissents disagree about where Texas’s harm lies. The dissents placed it in the loss of a unique national asset, a loss independent of what the burner says. The majority placed it in the message, because the statute reached only mistreatment likely to give serious offense, and on that basis it held the interest related to expression. Congress tried to answer that reasoning with a statute, as the next subsection explains.
United States v. Eichman
Congress tried to avoid Johnson with a statute, and United States v. Eichman held that the federal Flag Protection Act of 1989 was aimed at suppressing the communicative impact of expressive conduct and could not be enforced. Johnson drew heavy public criticism, and Congress divided over the response. Some members, with the President’s support, wanted a constitutional amendment, while others believed a carefully drafted statute might survive review and ease the pressure for one.22 The statutory strategy prevailed, and the Act passed both Houses by overwhelming majorities. It made it a crime to knowingly mutilate, deface, physically defile, burn, maintain on the floor or ground, or trample upon any flag of the United States, and it excepted “any conduct consisting of the disposal of the flag when it has become worn or soiled.”23
The drafting was meant to answer Johnson on its own terms. The Texas statute punished mistreatment only when the actor knew it would seriously offend onlookers, so liability turned on the audience’s likely reaction. The new Act dropped that element, and the government argued that it safeguarded “the physical integrity of the flag under all circumstances” and reached damaging conduct “without regard to the actor’s motive, his intended message, or the likely effects of his conduct on onlookers.”24 A statute written that way resembled the draft-certificate law in O’Brien, which reached every knowing destruction of a certificate, and its defenders hoped that Justices Scalia and Kennedy, who had joined the Johnson majority, would accept a law that was arguably no longer content-based.25
Protesters violated the Act on purpose to test it. Two prosecutions, one in Washington, D.C., and one in Seattle, ended when the trial courts dismissed the charges as unconstitutional.26 The government conceded that the burning was expressive conduct but asked the Court to reconsider Johnson’s rejection of the claim that flag burning, like obscenity or fighting words, falls outside the First Amendment. Justice Brennan’s majority opinion declined, which left one question: whether the Act was sufficiently distinct from the Texas statute to be enforced. The same five Justices who had formed the Johnson majority (Brennan, Marshall, Blackmun, Scalia, and Kennedy) answered that it was not.27
The Act’s wider scope changed nothing that mattered. Justice Brennan acknowledged that the Act contained no explicit content-based limitation, yet he found that “it is nevertheless clear that the Government’s asserted interest is related to the suppression of free expression.”28 The interest in the physical integrity of a privately owned flag ultimately rested on the need to preserve the flag’s status as a symbol of the nation. On the majority’s reasoning, no one would care about the physical integrity of a piece of cloth except for the message the cloth communicates, so the message is what the Act protects.29 The exemption for worn or soiled flags supports that reading. A flag burned in dignified retirement is as thoroughly destroyed as a flag burned at a rally, so the line the Act draws follows the meaning of the act and not the damage to the cloth.
The contrast with O’Brien shows which fact carried the result. The draft-certificate interest survives when the message is subtracted, because certificates matter to the Selective Service System whether or not anyone reads their burning as protest. The flag Act’s interest does not survive that subtraction, because it exists only so long as the flag keeps its meaning. That puts the case on the second track of the gatekeeping question, which asks why the government cares about the conduct, and the answer does not change when a legislature rewrites the text. If it did, protection for expression would depend on drafting, and a ban on whatever symbol carries a disfavored message could return under new wording.
Reed v. Town of Gilbert later stated the principle generally. A law that is content based on its face draws strict scrutiny, and so does a law that is facially neutral but whose purpose and justification are content based, so a court must evaluate each question before it concludes that a law is content neutral. Reed cited Eichman as its example of the second kind: a regulation neutral on its face may be content based if its manifest purpose is to regulate speech because of the message it conveys.30 The same attention to purpose appears outside symbolic-speech cases. Concurring in a school-library case, Justice Blackmun stressed that officials may not remove books to restrict access to the political ideas in them when the removal is motivated simply by disapproval of those ideas.31
Because the Act’s interest was related to expression, it drew the most exacting scrutiny, and the government’s interest could not justify the burden for the reasons given in Johnson. Justice Brennan wrote that the Act “suffers from the same fundamental flaw: it suppresses expression out of concern for its likely communicative impact.” The government pointed to Congress’s recognition of a national consensus favoring a ban, and the Court replied that “any suggestion that the Government’s interest in suppressing speech becomes more weighty as popular opposition to that speech grows is foreign to the First Amendment.” Government remains free to create national symbols, promote them, and encourage respect for them, but it may not criminally proscribe expressive conduct because of its likely communicative impact. The opinion closed by observing that punishing desecration of the flag dilutes the very freedom that makes the emblem so revered, and worth revering.32
Justice Stevens, joined by Chief Justice Rehnquist and Justices White and O’Connor, dissented, and his argument is the best foil for the majority’s. He agreed that government may not prohibit an idea because society finds it offensive, but he argued that a method of expression may be prohibited when a legitimate societal interest unrelated to suppressing the speaker’s ideas supports the ban, the speaker remains free to express those ideas by other means, and the interest in choosing among methods weighs less than the interest behind the ban.33 On his view the interest in the flag’s symbolic value is the same whichever idea motivated a burning, so the prosecution does not depend on the object of the protest.34 The Act, in that account, merely removed one method of communicating protest from the public forum.35 The majority’s answer is that a symbol matters to government because of what it communicates, so removing one symbol that carries one message selects among messages. Even the dissent conceded “a direct relationship between the communicative value of the act of flag burning and the symbolic value of the object being burned,”36 and that concession grants the premise the majority relied on: the flag’s value and its message are linked.
Eichman closed the statutory route, and the campaign for a constitutional amendment resumed. The amendment that reached the floor of both Houses in 1990 fell short of the required two-thirds majority in each, and a 1995 version passed the House 312 to 120 but fell three votes short in the Senate.37 The reason is the one the case teaches. A holding grounded in the First Amendment cannot be undone by rewording a law, so only a change to the Constitution itself would alter the result.
Neutral Laws Still Apply
A protester who handles a flag can still be punished, because Johnson and Eichman remove only one ground of liability, the communicative impact of the act, and leave every other ground in place. The gatekeeping question marks the line. A law whose interest exists whether or not anyone reads the act as a message sits on the O’Brien track, and a law whose interest depends on the message does not.
The Court has put the line in one sentence: nonverbal expressive activity can be banned because of the action it entails but not because of the ideas it expresses, so burning a flag in violation of an ordinance against outdoor fires could be punished, while burning one in violation of an ordinance against dishonoring the flag could not.39 The act, the flag, and the protester’s message are identical under both ordinances. Only the interest differs. A blaze threatens fire safety whatever the burner means by it, while the interest in a flag’s honor exists only because of what burning the flag says. Time, place, or manner rules follow the same logic. They are valid only if justified without reference to the content of the regulated speech, and the O’Brien test differs little from the standard applied to them.40 The reasoning reaches pure speech too. A protester who spray-painted an anti-draft slogan on a courthouse wall would be punished for damaging public property, and writing the message in words would protect that person no better than burning a certificate protected O’Brien, because the government’s focus in each case is the manner of the act and not its content.41
Johnson itself shows where the neutral laws were. Texas charged only desecration, and not theft, trespass, disorderly conduct, or arson, even though Johnson had received the flag from a fellow protester who took it from a flagpole outside one of the buildings the march targeted.42 The flag-burning decisions do not forbid a state to protect its own property or other people’s property from vandalism, so a prosecution under those general laws would have rested on an interest in property that exists whether or not anyone reads the act as protest. Justice Stevens’s dissent made a version of the point with a protester who spray-paints a message on the Lincoln Memorial, which no one would doubt government may forbid.43 The example is sound as far as it goes, since a law against defacing a monument serves a property interest whatever the message. The majority and the dissent parted over whether the Texas statute was a law of that kind. The majority held that it was not, because the statute reached only mistreatment likely to cause serious offense, so whether Johnson violated it depended on the communicative impact of his conduct.44
The other exception follows from Johnson, which stated that the First Amendment does not forbid a State to prevent imminent lawless action and observed that Texas already had a statute specifically prohibiting breaches of the peace.45 What a state may not do is presume disorder from offense. The inquiry asks whether the expression is directed to inciting imminent lawless action and is likely to produce it, and fighting words reach only expression likely to provoke the average person to retaliation, which a generalized protest against government policy does not. A peace-based charge therefore stays available when the record shows a call to immediate violence and a likelihood that it will follow, and it drops out when the only evidence is that onlookers were offended.
A law’s neutral appearance does not settle which track it is on. The federal Act in Eichman contained no offense element, yet it failed because the government’s interest in the flag’s physical integrity depended on the flag’s status as a national symbol. The inquiry therefore looks to the interest the government actually serves, and a general law gets the O’Brien treatment only when that interest would exist even if the act carried no message. Whether a neutral law is being enforced against protesters for a reason tied to their message is a separate question, taken up in the section on protest and selective enforcement.
Hate Crimes and Cross Burning
The flag decisions removed one ground for punishing an act, the offense its message gives, and left government free to punish on grounds unrelated to the message. The same line governs three settings in which an act carries a message: a sentence enhanced because the defendant picked the victim for a bias reason, a ban on burning a cross to intimidate, and the limits that keep both from reaching mere belief. Virginia v. Black supplies the frame. The First Amendment protects symbolic and expressive conduct as well as actual speech, but the protection is not absolute, because government may still regulate certain categories of expression, among them true threats.
The least controversial setting is the penalty enhancement. In Wisconsin v. Mitchell, a unanimous Court upheld a Wisconsin statute that lengthened a defendant’s sentence whenever the defendant intentionally selected the victim because of the victim’s race. The underlying offense there was a battery, a crime on its own terms.47 A bias motive may count because of what it adds to the crime. Courts consider motive whenever they sentence, imposing lighter sentences for good motives and heavier ones for bad, so that murder for hire is often punished more severely than murder in a sudden rage. A crime committed because of the victim’s race also changes the harm the act causes and the likelihood that the offender will repeat it, and the enhancement answers those consequences of the act. A state may likewise punish spraying a Nazi emblem on a synagogue more severely than spray painting a nonsense word on a subway train, because the First Amendment offers no shelter to a criminal act merely because a political motive lay behind it.
The enhancement differs sharply from the ordinance struck down in R.A.V. v. City of St. Paul. There the defendants allegedly burned a cross in the fenced yard of a Black family, conduct punishable under any of a number of laws, yet the city charged its Bias-Motivated Crime Ordinance, which reached symbols that one knows will arouse anger, alarm, or resentment on the basis of race, color, creed, religion, or gender.48 The Court held the ordinance facially invalid because, even as narrowed to fighting words, it applied only to those that insult or provoke violence on the listed topics, and the city itself had conceded that the content of the cross-burning message was what the ordinance attempted to legislate.49 The St. Paul ordinance made the message the offense, while the Wisconsin statute left the offense defined by the act and used the motive only to set the penalty. R.A.V. itself drew that line: where 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.50 The Court observed that “St. Paul has sufficient means at its disposal to prevent such behavior without adding the First Amendment to the fire.”51
The first law resembles the Wisconsin statute, since it adjusts the penalty for a crime defined by the act. The second resembles the St. Paul ordinance, since the message alone makes the display an offense. A burning cross raises a separate question when it is used to frighten a particular person, and Virginia v. Black took that question up.
Virginia’s statute made burning a cross on the property of another a felony, and the Court held that the First Amendment permits a state to ban a true threat.52 True threats are statements in which the speaker means to communicate a serious expression of an intent to commit an act of unlawful violence against a particular individual or group. The speaker need not actually intend to carry out the threat, because the prohibition protects people from the fear of violence, from the disruption that fear engenders, and from the possibility that the threatened violence will occur. Intimidation in the proscribable sense is a type of true threat, one in which the speaker intends to place the victim in fear of bodily harm or death, and the history of cross burning shows that it is often used that way.53 The Virginia Supreme Court had read R.A.V. to doom the statute anyway, because a ban that singles out cross burning discriminates by content. The Court disagreed. R.A.V. did not forbid all content discrimination within a proscribable category, and it allows such discrimination when the basis consists entirely of the very reason the whole class is proscribable.54
Virginia chose a subset of threats for the feature that makes threats punishable in the first place, the fear of violence, and it was entitled to do so because burning a cross is “a particularly virulent form of intimidation” with a long history as a signal of impending violence.55 The statute also reaches intimidation whatever its target. It does not matter whether the cross burner intends to intimidate because of the victim’s race, gender, or religion or because of the victim’s political affiliation or union membership, and cross burners do not direct their intimidating conduct solely to racial or religious minorities.56 That breadth separates Virginia’s law from St. Paul’s ordinance, which singled out speech on specified disfavored topics, so the danger that a state is handicapping one side of a debate is slight.
The statute also provided that any burning of a cross was prima facie evidence of an intent to intimidate, and that provision did not survive.57 A plurality led by Justice O’Connor facially invalidated it, because treating the burning alone as enough to infer intent ignored the context that shows whether a particular burning was meant to intimidate or to express an ideology or anger, and so it chilled protected political speech.58 The holding follows from what the true-threat category requires. Intent to intimidate separates an unprotected threat from a protected display, so a presumption that supplies the intent from the act alone removes the one finding that does the sorting. The prosecutions in the case show why the sorting matters. Black was convicted for burning a cross while leading a Ku Klux Klan rally, and Elliott and O’Mara for burning one on the lawn of an African-American neighbor.59 A burning cross directed at a particular person often serves as a message of intimidation meant to inspire fear of bodily harm, while a cross burned as an ideological statement is protected speech.60
The Justices divided on that last point. Justice Scalia, joined by Justice Thomas, dissented from the facial invalidation. A law is not overbroad merely because protected speakers may be arrested and prosecuted, he argued, since overbreadth turns on whether they can be convicted, and the plurality had not shown that the Virginia law is substantially overbroad.61 Justice Souter, joined by Justices Kennedy and Ginsburg, would have struck down the whole statute, because a content-based ban on cross burning may be a subtle effort to suppress the white-supremacy message that even nonthreatening burnings broadcast.62 On his view the presumption skews jury deliberations toward conviction where the evidence of intent is thin, and so skews the statute toward suppressing ideas.63
Both settings rest on a limit that Dawson v. Delaware makes concrete. There the Court held that the First Amendment barred the prosecution from introducing, at a capital sentencing proceeding, the defendant’s membership in a racist organization called the Aryan Brotherhood, because the evidence had no relevance to the issues being decided.64 The victim, like the defendant, was white, so no racial hatred figured in the murder, and the evidence showed nothing more than the defendant’s abstract beliefs. Mitchell shows which fact moved the result. In Mitchell the bias motive was the reason the defendant selected the victim, so it explained the harm and the risk of repetition, while in Dawson the belief explained nothing about the crime. The prosecution could have avoided the result by tying the association to a sentencing question. It said its expert would have described the group as a white racist prison gang associated with drugs and violent escape attempts that advocates the murder of fellow inmates, and such specific evidence might have rebutted the defendant’s mitigating evidence of kindness to family and good time earned in prison programs. A state also may not punish people merely for belonging to a racist organization or for hate speech unconnected to any crime.65
Set beside the flag cases, these decisions ask the same question: does the state’s reason for punishing depend on the idea the act conveys? An enhancement for victim selection turns on the harm and the risk of repetition, a ban on intimidating cross burning turns on the fear of violence, and Dawson demands a link between belief and the offense. The Court has long held that nonverbal expressive activity can be banned because of the action it entails and cannot be banned because of the ideas it expresses.66 A prosecution that rests on disapproval of the message fails for the reason the flag statutes failed, however the offense is labeled.