delivered the opinion of the Court.
In the predawn hours of June 21,1990, petitioner and several other teenagers allegedly assembled a crudely made cross by taping together broken chair legs. They then allegedly burned the cross inside the fenced yard of a black family that lived across the street from the house where petitioner was staying. Although this conduct could have been pun*380ished under any of a number of laws,1 one of the two provisions under which respondent city of St. Paul chose to charge petitioner (then a juvenile) was the St. Paul Bias-Motivated Crime Ordinance, St. Paul, Minn., Legis. Code § 292.02 (1990), which provides:
“Whoever places on public or private property a symbol, object, appellation, characterization or graffiti, including, but not limited to, a burning cross or Nazi swastika, which 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 commits disorderly conduct and shall be guilty of a misdemeanor.”
Petitioner moved to dismiss this count on the ground that the St. Paul ordinance was substantially overbroad and impermissibly content based and therefore facially invalid under the First Amendment.2 The trial court granted this motion, but the Minnesota Supreme Court reversed. That court rejected petitioner’s overbreadth claim because, as construed in prior Minnesota cases, see, e. g., In re Welfare of S. L. J., 263 N. W. 2d 412 (Minn. 1978), the modifying phrase “arouses anger, alarm or resentment in others” limited the reach of the ordinance to conduct that amounts to “fighting words,” i. e., “conduct that itself inflicts injury or tends to incite immediate violence . . . ,” In re Welfare of R. A. V., 464 N. W. 2d 507, 510 (Minn. 1991) (citing Chaplin*381sky v. New Hampshire, 315 U. S. 568, 572 (1942)), and therefore the ordinance reached only expression “that the first amendment does not protect,” 464 N. W. 2d, at 511. The court also concluded that the ordinance was not impermissi-bly content based because, in its view, “the ordinance is a narrowly tailored means toward accomplishing the compelling governmental interest in protecting the community against bias-motivated threats to public safety and order.” Ibid. We granted certiorari, 501 U. S. 1204 (1991).
HH
In construing the St. Paul ordinance, we are bound by the construction given to it by the Minnesota court. Posadas de Puerto Rico Associates v. Tourism Co. of Puerto Rico, 478 U. S. 328, 339 (1986); New York v. Ferber, 458 U. S. 747, 769, n. 24 (1982); Terminiello v. Chicago, 337 U. S. 1, 4 (1949). Accordingly, we accept the Minnesota Supreme Court’s authoritative statement that the ordinance reaches only those expressions that constitute “fighting words” within the meaning of Chaplinsky. 464 N. W. 2d, at 510-511. Petitioner and his amici urge us to modify the scope of the Chaplinsky formulation, thereby invalidating the ordinance as “substantially overbroad,” Broadrick v. Oklahoma, 413 U. S. 601, 610 (1973). We find it unnecessary to consider this issue. Assuming, arguendo, that all of the expression reached by the ordinance is proscribable under the “fighting words” doctrine, we nonetheless conclude that the ordinance is facially unconstitutional in that it prohibits otherwise permitted speech solely on the basis of the subjects the speech addresses.3
*382The First Amendment generally prevents government from proscribing speech, see, e. g., Cantwell v. Connecticut, 310 U. S. 296, 309-311 (1940), or even expressive conduct, see, e. g., Texas v. Johnson, 491 U. S. 397, 406 (1989), because of disapproval of the ideas expressed. Content-based regulations are presumptively invalid. Simon & Schuster, Inc. v. Members of N. Y. State Crime Victims Bd., 502 U. S. 105, 115 (1991); id., at 124 (Kennedy, J., concurring in judgment); Consolidated Edison Co. of N. Y. v. Public Serv. Comm’n of N. Y, 447 U. S. 530, 536 (1980); Police Dept. of Chicago v. Mosley, 408 U. S. 92, 95 (1972). From 1791 to the present, however, our society, like other free but civilized societies, has permitted restrictions upon the content of speech in a *383few limited areas, which are “of such slight social value as a step to truth that any benefit that may be derived from them is clearly outweighed by the social interest in order and morality.” Chaplinsky, supra, at 572. We have recognized that “the freedom of speech” referred to by the First Amendment does not include a freedom to disregard these traditional limitations. See, e. g., Roth v. United States, 354 U. S. 476 (1957) (obscenity); Beauharnais v. Illinois, 343 U. S. 250 (1952) (defamation); Chaplinsky v. New Hampshire, supra (“ ‘fighting’ words”); see generally Simon & Schuster, supra, at 124 (Kennedy, J., concurring in judgment). Our decisions since the 1960’s have narrowed the scope of the traditional categorical exceptions for defamation, see New York Times Co. v. Sullivan, 376 U. S. 254 (1964); Gertz v. Robert Welch, Inc., 418 U. S. 323 (1974); see generally Milkovich v. Lorain Journal Co., 497 U. S. 1, 13-17 (1990), and for obscenity, see Miller v. California, 413 U. S. 15 (1973), but a limited categorical approach has remained an important part of our First Amendment jurisprudence.
We have sometimes said that these categories of expression are “not within the area of constitutionally protected speech,” Roth, supra, at 483; Beauharnais, supra, at 266; Chaplinsky, supra, at 571-572, or that the “protection of the First Amendment does not extend” to them, Bose Corp. v. Consumers Union of United States, Inc., 466 U. S. 485, 504 (1984); Sable Communications of Cal., Inc. v. FCC, 492 U. S. 115, 124 (1989). Such statements must be taken in context, however, and are no more literally true than is the occasionally repeated shorthand characterizing obscenity “as not being speech at all,” Sunstein, Pornography and the First Amendment, 1986 Duke L. J. 589, 615, n. 46. What they mean is that these areas of speech can, consistently with the First Amendment, be regulated because of their constitutionally proscribable content (obscenity, defamation, etc.)— not that they are categories of speech entirely invisible to the Constitution, so that they may be made the vehicles for *384content discrimination unrelated to their distinctively pro-scribable content. Thus, the government may proscribe libel; but it may not make the further content discrimination of proscribing only libel critical of the government. We recently acknowledged this distinction in Ferber, 458 U. S., at 763, where, in upholding New York’s child pornography law, we expressly recognized that there was no “question here of censoring a particular literary theme . . . See also id., at 775 (O’Connor, J., concurring) (“As drafted, New York’s statute does not attempt to suppress the communication of particular ideas”).
Our cases surely do not establish the proposition that the First Amendment imposes no obstacle whatsoever to regulation of particular instances of such proseribable expression, so that the government “may regulate [them] freely,” post, at 400 (White, J., concurring in judgment). That would mean that a city council could enact an ordinance prohibiting only those legally obscene works that contain criticism of the city government or, indeed, that do not include endorsement of the city government. Such a simplistic, all-or-nothing-at-all approach to First Amendment protection is at odds with common sense and with our jurisprudence as well.4 It is *385not true that “fighting words” have at most a “de minimis” expressive content, ibid., or that their content is in all respects “worthless and undeserving of constitutional protection,” post, at 401; sometimes they are quite expressive indeed. We have not said that they constitute “no part of the expression of ideas,” but only that they constitute “no essential part of any exposition of ideas.” Chaplinsky, supra, at 572 (emphasis added).
The proposition that a particular instance of speech can be proseribable on the basis of one feature (e. g., obscenity) but not on the basis of another (e. g., opposition to the city government) is commonplace and has found application in many contexts. We have long held, for example, that nonverbal expressive activity can be banned because of the action it entails, but not because of the ideas it expresses — so that burning a flag in violation of an ordinance against outdoor fires could be punishable, whereas burning a flag in violation of an ordinance against dishonoring the flag is not. See Johnson, 491 U. S., at 406-407. See also Barnes v. Glen Theatre, Inc., 501 U. S. 560, 569-570 (1991) (plurality opinion); id., at 573-574 (Scalia, J., concurring in judgment); id., at 581-582 (Souter, J., concurring in judgment); United *386States v. O’Brien, 391 U. S. 367, 376-377 (1968). Simüarly, we have upheld reasonable “time, place, or manner” restrictions, but only if they are “justified without reference to the content of the regulated speech.” Ward v. Rock Against Racism, 491 U. S. 781, 791 (1989) (internal quotation marks omitted); see also Clark v. Community for Creative NonViolence, 468 U. S. 288, 298 (1984) (noting that the OBrien test differs little from the standard applied to time, place, or manner restrictions). And just as the power to proscribe particular speech on the basis of a noncontent element (e. g., noise) does not entail the power to proscribe the same speech on the basis of a content element; so also, the power to proscribe it on the basis of one content element (e. g., obscenity) does not entail the power to proscribe it on the basis of other content elements.
In other words, the exclusion of “fighting words” from the scope of the First Amendment simply means that, for purposes of that Amendment, the unprotected features of the words are, despite their verbal character, essentially a “non-speech” element of communication. Fighting words are thus analogous to a noisy sound truck: Each is, as Justice Frankfurter recognized, a “mode of speech,” Niemotko v. Maryland, 340 U. S. 268, 282 (1951) (opinion concurring in result); both can be used to convey an idea; but neither has, in and of itself, a claim upon the First Amendment. As with the sound truck, however, so also with fighting words: The government may not regulate use based on hostility — or favoritism — towards the underlying message expressed. Compare Frisby v. Schultz, 487 U. S. 474 (1988) (upholding, against facial challenge, a content-neutral ban on targeted residential picketing), with Carey v. Brown, 447 U. S. 455 (1980) (invalidating a ban on residential picketing that exempted labor picketing).5
*387The concurrences describe us as setting forth a new First Amendment principle that prohibition of constitutionally proseribable speech cannot be “underinelusiv[e],” post, at 402 (White, J., concurring in judgment) — a First Amendment “absolutism” whereby “[wjithin a particular ‘proseribable’ category of expression, ... a government must either proscribe all speech or no speech at all,” post, at 419 (Stevens, J., concurring in judgment). That easy target is of the concurrences’ own invention. In our view, the First Amendment imposes not an “underinclusiveness” limitation but a “content discrimination” limitation upon a State’s prohibition of proseribable speech. There is no problem whatever, for example, with a State’s prohibiting obscenity (and other forms of proseribable expression) only in certain media or markets, for although that prohibition would be “underinclu-sive,” it would not discriminate on the basis of content. See, e. g., Sable Communications, 492 U. S., at 124-126 (upholding 47 U. S. C. § 223(b)(1), which prohibits obscene telephone communications).
Even the prohibition against content discrimination that we assert the First Amendment requires is not absolute. It applies differently in the context of proseribable speech than in the area of fully protected speech. The rationale of the general prohibition, after all, is that content discrimination “raises the specter that the Government may effectively drive certain ideas or viewpoints from the marketplace,” Simon & Schuster, 502 U. S., at 116; Leathers v. Medlock, 499 U. S. 439, 448 (1991); FCC v. League of Women Voters of Cal., 468 U. S. 364, 383-384 (1984); Consolidated Edison Co., 447 U. S., at 536; Police Dept. of Chicago v. Mosley, 408 U. S., *388at 95-98. But content discrimination among various instances of a class of proscribable speech often does not pose this threat.
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. 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. To illustrate: A State might choose to prohibit only that obscenity which is the most patently offensive in its 'prurience— i. e., that which involves the most lascivious displays of sexual activity. But it may not prohibit, for example, only that obscenity which includes offensive political messages. See Kucharek v. Hanaway, 902 F. 2d 513, 517 (CA7 1990), cert. denied, 498 U. S. 1041 (1991). And the Federal Government can criminalize only those threats of violence that are directed against the President, see 18 U. S. C. § 871 — since the reasons why threats of violence are 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. See Watts v. United States, 394 U. S. 705, 707 (1969) (upholding the facial validity of §871 because of the “overwhelmin[g] interest in protecting the safety of [the] Chief Executive and in allowing him to perform his duties without interference from threats of physical violence”). But the Federal Government may not criminalize only those threats against the President that mention his policy on aid to inner cities. And to take a final example (one mentioned by Justice Stevens, post, at 421-422), a State may choose to regulate price advertising in one industry but not in others, because the risk of fraud (one of the characteristics of commercial speech' that justifies depriving it of full First Amendment protection, see Virginia *389State Bd. of Pharmacy v. Virginia Citizens Consumer Council, Inc., 425 U. S. 748, 771-772 (1976)) is in its view greater there. Cf. Morales v. Trans World Airlines, Inc., 504 U. S. 374 (1992) (state regulation of airline advertising); Ohralik v. Ohio State Bar Assn., 436 U. S. 447 (1978) (state regulation of lawyer advertising). But a State may not prohibit only that commercial advertising that depicts men in a demeaning fashion. See, e. g., Los Angeles Times, Aug. 8, 1989, section 4, p. 6, col. 1.
Another valid basis for according differential treatment to even a content-defined subclass of proseribable speech is that the subclass 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,” Renton v. Playtime Theatres, Inc., 475 U. S. 41, 48 (1986) (quoting, with emphasis, Virginia State Bd. of Pharmacy, supra, at 771); see also Young v. American Mini Theatres, Inc., 427 U. S. 50, 71, n. 34 (1976) (plurality opinion); id., at 80-82 (Powell, J., concurring); Barnes, 501 U. S., at 586 (Souter, J., concurring in judgment). A State could, for example, permit all obscene live performances except those involving minors. Moreover, since words can in some circumstances violate laws directed not against speech but against conduct (a law against treason, for example, is violated by telling the enemy the Nation’s defense secrets), a particular content-based subcategory of a proseribable class of speech can be swept up incidentally within the reach of a statute directed at conduct rather than speech. See id., at 571 (plurality opinion); id., at 577 (Scalia, J., concurring in judgment); id., at 582 (Souter, J., concurring in judgment); FTC v. Superior Court Trial Lawyers Assn., 493 U. S. 411, 425-432 (1990); O’Brien, 391 U. S., at 376-377. Thus, for example, sexually derogatory “fighting words,” among other words, may produce a violation of Title VII’s general prohibition against sexual discrimination in employment practices, 42 U. S. G. §2000e-2; 29 GFR § 1604.11 (1991). See also 18 *390U. S. C. §242; 42 U. S. C. §§ 1981, 1982. 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.
These bases for distinction refute the proposition that the selectivity of the restriction is “even arguably ‘conditioned upon the sovereign’s agreement with what a speaker may intend to say.’ ” Metromedia, Inc. v. San Diego, 453 U. S. 490, 555 (1981) (Stevens, J., dissenting in part) (citation omitted). There may be other such bases as well. Indeed, to validate such 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 is such that there is no realistic possibility that official suppression of ideas is afoot. (We cannot 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.) Save for that limitation, the regulation of “fighting words,” like the regulation of noisy speech, may address some offensive instances and leave other, equally offensive, instances alone. See Posadas de Puerto Rico, 478 U. S., at 342-343.6
*391HH I-H
Applying these principles to the St. Paul ordinance, we conclude that, even as narrowly construed by the Minnesota Supreme Court, the ordinance is facially unconstitutional. Although the phrase in the ordinance, “arouses anger, alarm or resentment in others,” has been limited by the Minnesota Supreme Court’s construction to reach only those symbols or displays that amount to “fighting words,” the remaining, unmodified terms make clear that the ordinance applies only to “fighting words” that insult, or provoke violence, “on the basis of race, color, creed, religion or gender.” Displays containing abusive invective, no matter how vicious or severe, are permissible unless they are addressed to one of the specified disfavored topics. Those who wish to use “fighting ■words” in connection with other ideas — to express hostility, for example, on the basis of political affiliation, union membership, or homosexuality — are not covered. The First Amendment does not permit St. Paul to impose special prohibitions on those speakers who express views on disfavored subjects. See Simon & Schuster, 502 U. S., at 116; Arkansas Writers’ Project, Inc. v. Ragland, 481 U. S. 221, 229-230 (1987).
In its practical operation, moreover, the ordinance goes even beyond mere content discrimination, to actual viewpoint discrimination. Displays containing some words— odious racial epithets, for example — would be prohibited to proponents of all views. But “fighting words” that do not themselves invoke race, color, creed, religion, or gender— aspersions upon a person’s mother, for example — would seemingly be usable ad libitum in the placards of those arguing in favor of racial, colorj etc., tolerance and equality, but could not be used by those speakers’ opponents. One could hold up a sign saying, for example, that all “anti-*392Catholic bigots” are misbegotten; but not that all “papists” are, for that would insult and provoke violence “on the basis of religion.” St. Paul has no such authority to license one side of a debate to fight freestyle, while requiring the other to follow Marquis of Queensberry rules.
What we have here, it must be emphasized, is not a prohibition of fighting words that are directed at certain persons or groups (which would be facially valid if it met the requirements of the Equal Protection Clause); but rather, a prohibition of fighting words that contain (as the Minnesota Supreme Court repeatedly emphasized) messages of “bias-motivated” hatred and in particular, as applied to this case, messages “based on virulent notions of racial supremacy.” 464 N. W. 2d, at 508, 511. One must wholeheartedly agree with the Minnesota Supreme Court that “[i]t is the responsibility, even the obligation, of diverse communities to confront such notions in whatever form they appear,” id., at 508, but the manner of that confrontation cannot consist of selective limitations upon speech. St. Paul’s brief asserts that a general “fighting words” law would hot meet the city’s needs because only a content-specific measure can communicate to minority groups that the “group hatred” aspect of such speech “is not condoned by the majority.” Brief for Respondent 25. The point of the First Amendment is that majority preferences must be expressed in some fashion other than silencing speech on the basis of its content.
Despite the fact that the Minnesota Supreme Court and St. Paul acknowledge that the ordinance is directed at expression of group hatred, Justice Stevens suggests that this “fundamentally misreads” the ordinance. Post, at 438. It is directed, he claims, not to speech of a particular content, but to particular “injur[ies]” that are “qualitatively different” from other injuries. Post, at 424. This is wordplay. What makes the anger, fear, sense of dishonor, etc., produced by violation of this ordinance distinct from the anger, fear, sense of dishonor, etc., produced by other fighting words is *393nothing other than the fact that it is caused by a distinctive idea, conveyed by a distinctive message. The First Amendment cannot be evaded that easily. It is obvious that the symbols which will arouse “anger, alarm or resentment in others on the basis of race, color, creed, religion or gender” are those symbols that communicate a message of hostility based on one of these characteristics. St. Paul concedes in its brief that the ordinance applies only to “racial, religious, or gender-specific symbols” such as “a burning cross, Nazi swastika or other instrumentality of like import.” Brief for Respondent 8. Indeed, St. Paul argued in the Juvenile Court that “[t]he burning of a cross does express a message and it is, in fact, the content of that message which the St. Paul Ordinance attempts to legislate.” Memorandum from the Ramsey County Attorney to the Honorable Charles A. Flinn, Jr., dated July 13, 1990, in In re Welfare of R. A. V., No. 89-D-1231 (Ramsey Cty. Juvenile Ct.), p. 1, reprinted in App. to Brief for Petitioner C-l.
The content-based discrimination reflected in the St. Paul ordinance comes within neither any of the specific exceptions to the First Amendment prohibition we discussed earlier nor a more general exception for content discrimination that does not threaten censorship of ideas. It 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 proseribable. As explained earlier, see supra, at 386, the reason why fighting words are categorically excluded from the protection of the First Amendment is not that their content communicates any particular idea, but that their content embodies a particularly intolerable (and socially unnecessary) mode of expressing whatever idea the speaker wishes to convey. St. Paul has not singled out an especially offensive mode of expression — it has not, for example, selected for prohibition only those fighting words that communicate ideas in a threatening (as opposed to a merely obnoxious) manner. Rather, it has proscribed fight*394ing words of whatever manner that communicate messages of racial, gender, or religious intolerance. Selectivity of this sort creates the possibility that the city is seeking to handicap the expression of particular ideas. That possibility would alone be enough to render the ordinance presumptively invalid, but St. Paul's comments and concessions in this ease elevate the possibility to a certainty.
St. Paul argues that the ordinance comes within another of the specific exceptions we mentioned, the one that allows content discrimination aimed only at the “secondary effects” of the speech, see Renton v. Playtime Theatres, Inc., 475 U. S. 41 (1986). According to St. Paul, the ordinance is intended, “not to impact on [sic] the right of free expression of the accused,” but rather to. “protect against the victimization of a person or persons who are particularly vulnerable because of their membership in a group that historically has been discriminated against.” Brief for Respondent 28. Even assuming that an ordinance that completely proscribes, rather than merely regulates, a specified category of speech can ever be considered to be directed only to the secondary effects of such speech, it is clear that the St. Paul ordinance is not directed to secondary effects within the meaning of Renton. As we said in Boos v. Barry, 485 U. S. 312 (1988), “Listeners' reactions to speech are not the type of ‘secondary effects’ we referred to in Renton.” Id., at 321. “The emotive impact of speech on its audience is not a ‘secondary effect.’” Ibid. See also id., at 334 (opinion of Brennan, J.).7
*395It hardly needs discussion that the ordinance does not fall within some more general exception permitting all selectivity that for any reason is beyond the suspicion of official suppression of ideas. The statements of St. Paul in this very case afford ample basis for, if not full confirmation of, that suspicion.
Finally, St. Paul and its amici defend the conclusion of the Minnesota Supreme Court that, even if the ordinance regulates expression based on hostility towards its protected ideological content, this discrimination is nonetheless justified because it is narrowly tailored to serve compelling state interests. Specifically, they assert that the ordinance helps to ensure the basic human rights of members of groups that have historically been subjected'to discrimination, including the right of such group members to live in peace where they wish. We do not doubt that these interests are compelling, and that the ordinance can be said to promote them. But the “danger of censorship” presented by a facially content-based statute, Leathers v. Medlock, 499 U. S., at 448, requires that that weapon be employed only where it is “necessary to serve the asserted [compelling] interest,” Burson v. Freeman, 504 U. S. 191, 199 (1992) (plurality opinion) (emphasis added); Perry Ed. Assn. v. Perry Local Educators’ Assn., 460 U. S. 37, 45 (1983). The existence of adequate content-neutral alternatives thus “undercuts] significantly” any defense of such a statute, Boos v. Barry, supra, at 329, easting considerable doubt on the government’s protestations that “the asserted justification is in fact an accurate description of the purpose and effect of the law,” Burson, supra, at 213 (Kennedy, J., concurring). See Boos, supra, at 324-329; cf. Minneapolis Star & Tribune Co. v. Minnesota Comm’r of Revenue, 460 U. S. 575, 586-587 (1983). The dispositive question in this case, therefore, is whether content discrimination is reasonably necessary to achieve St. Paul’s compel*396ling interests; it plainly is not. An ordinance not limited to the favored topics, for example, would have precisely the same beneficial effect. In fact the only interest distinctively served by the content limitation is that of displaying the city council’s special hostility towards the particular biases thus singled out.8 That is precisely what the First Amendment forbids. The politicians of St. Paul are entitled to express that hostility — but not through the means of imposing unique limitations upon speakers who (however benightedly) disagree.
* * *
Let there be no mistake about our belief that burning a cross in someone’s front yard is reprehensible. But St. Paul has sufficient means at its disposal to prevent such behavior without adding the First Amendment to the fire.
The judgment of the Minnesota Supreme Court is reversed, and the case is remanded for proceedings not inconsistent with this opinion.
It is so ordered.
The conduct might have violated Minnesota statutes carrying significant penalties. See, e. g., Minn. Stat. §609.713(1) (1987) (providing for up to five years in prison for terroristic threats); § 609.568 (arson) (providing for up to five years and a $10,000 fine, depending on the value of the property intended to be damaged); § 609.695 (Supp. 1992) (criminal damage to property) (providing for up to one year and a $3,000 fine, depending upon the extent of the damage to the property).
Petitioner has also been charged, in Count I of the delinquency petition, with a violation of Minn. Stat. § 609.2231(4) (Supp. 1990) (racially motivated assaults). Petitioner did not challenge this count.
Contrary to Justice White’s suggestion, post, at 397-398, n. 1, petitioner’s claim is “fairly included” within the questions presented in the petition for certiorari, see this Court’s Rule 14.1(a). It was clear from the petition and from petitioner’s other filings in this Court (and in the courts below) that his assertion that the St. Paul ordinance “violat[es] over-breadth ... principles of the First Amendment,” Pet. for Cert, i, was not *382just a technical “overbreadth” claim — i. e., a claim that the ordinance violated the rights of too many third parties — but included the contention that the ordinance was “overbroad” in the sense of restricting more speech than the Constitution permits, even in its application to him, because it is content based. An important component of petitioner’s argument is, and has been all along, that narrowly construing the ordinance to cover only “fighting words” cannot cure this fundamental defect. Id., at 12,14,15-16. In his briefs in this Court, petitioner argued that a narrowing construction was ineffective because (1) its boundaries were vague, Brief for Petitioner 26, and because (2) denominating particular expression a "fighting word” because of the impact of its ideological content upon the audience is inconsistent with the First Amendment, Reply Brief for Petitioner 6; id., at 13 (“[The ordinance] is overbroad, viewpoint discriminatory and vague as ‘narrowly construed’”) (emphasis added). At oral argument, counsel for petitioner reiterated this second point: “It is . . . one of my positions, that in [punishing only some fighting words and not others], even though it is a subcategory, technically, of unprotected conduct, [the ordinance] still is picking out an opinion, a disfavored message, and making that dear through the State.” Tr. of Oral Arg. 8. In resting our judgment upon this contention, we have not departed from our criteria of what is "fairly included” within.the petition. See Arkansas Electric Cooperative Corp. v. Arkansas Pub. Serv. Comm’n, 461 U. S. 375, 382, n. 6 (1983); Brown v. Socialist Workers ’74 Campaign Comm., 459 U. S. 87, 94, n. 9 (1982); Eddings v. Oklahoma, 455 U. S. 104, 113, n. 9 (1982); see generally R. Stern, E. Gressman, & S. Shapiro, Supreme Court Practice 361 (6th ed. 1986).
Justice White concedes that a city council cannot prohibit only those legally obscene works that contain criticism of the city government, post, at 406, but asserts that to be the consequence, not of the First Amendment, but of the Equal Protection Clause. Such content-based discrimination would not, he asserts, “be rationally related to a legitimate government interest.” Ibid. But of course the only reason that government interest is not a “legitimate” one is that it violates the First Amendment. This Court itself has occasionally fused the First Amendment into the Equal Protection Clause in this fashion, but at least with the acknowledgment (which Justice White cannot afford to make) that the First Amendment underlies its analysis. See Police Dept. of Chicago v. Mosley, 408 U. S. 92, 95 (1972) (ordinance prohibiting only nonlabor picketing violated the Equal Protection Clause because there was no “appropriate governmental interest” supporting the distinction inasmuch as “the First Amendment means that government has no power to restrict expression because of its message, its ideas, its subject matter, or its content”); Carey v. *385Brown, 447 U. S. 455 (1980). See generally Simon & Schuster, Inc. v. Members of N. Y. State Crime Victims Bd., 502 U. S. 105, 124 (1991) (Kennedy, J., concurring in judgment).
Justice Stevens seeks to avoid the point by dismissing the notion of obscene antigovernment speech as “fantastical,” post, at 418, apparently believing that any reference to politics prevents a finding of obscenity. Unfortunately for the purveyors of obscenity, that is obviously false. A shockingly hardcore pornographic movie that contains a model sporting a political tattoo can be found, “taken as a whole, [to] lac[k] serious literary, artistic, political, or scientific value,” Miller v. California, 413 U. S. 15, 24 (1973) (emphasis added). Anyway, it is easy enough to come up with other illustrations of a content-based restriction upon “unprotected speech” that is obviously invalid: the antigovernment libel illustration mentioned earlier, for one. See supra, at 384. And of course the concept of racist fighting words is, unfortunately, anything but a “highly speculative hypothetical],” post, at 419.
Although Justice White asserts that our analysis disregards “established principles of First Amendment law,” post, at 416, he cites not a single case (and we are aware of none) that even involved, much less con*387sidered and resolved, the issue of content discrimination through regulation of "unprotected” speech — though we plainly recognized that as an issue in New York v. Ferber, 458 U. S. 747 (1982). It is of course contrary to all traditions of our jurisprudence to consider the law on this point conclusively resolved by broad language in cases where the issue was not presented or even envisioned.
Justice Stevens cites a string of opinions as supporting his assertion that “selective regulation of speech based on content” is not presumptively invalid. Post, at 421-422. Analysis reveals, however, that they do not support it. To begin with, three of them did not command a majority of the Court, Young v. American Mini Theatres, Inc., 427 U. S. 50, 63-73 (1976) (plurality opinion); FCC v. Pacifica Foundation, 438 U. S. 726, 744-748 (1978) (plurality opinion); Lehman v. Shaker Heights, 418 U. S. 298 (1974) (plurality opinion), and two others did not even discuss the First Amendment, Morales v. Trans World Airlines, Inc., 504 U. S. 374 (1992); Jacob Siegel Co. v. FTC, 327 U. S. 608 (1946). In any event, all that their contents establish is what we readily concede: that presumptive invalidity does not mean invariable invalidity, leaving room for such exceptions as reasonable and viewpoint-neutral content-based discrimination in nonpublic forums, see Lehman, supra, at 301-304; see also Cornelius v. NAACP Legal Defense & Ed. Fund, Inc., 473 U. S. 788, 806 (1985), or with respect to certain speech by government employees, see Broadrick v. Oklahoma, *391413 U. S. 601 (1973); see also Civil Service Comm’n v. Letter Carriers, 413 U. S. 548, 564-567 (1973).
St. Paul has not argued in this case that the ordinance merely regulates that subclass of fighting words which is most likely to provoke a violent response. But even if one assumes (as appears unlikely) that the categories selected may be so described, that would not justify selective regulation under a “secondary effects” theory. The only reason why such expressive conduct would be especially correlated with violence is that it conveys a particularly odious message; because the “chain of causation” thus necessarily “run[s] through the persuasive effect of the expressive component” of the conduct, Barnes v. Glen Theatre, Inc., 501 U. S. 560, 586 (1991) (Souter, J., concurring in judgment), it is clear that the St. Paul *395ordinance regulates on the basis of the “primary” effect of the speech— i e., its persuasive (or repellant) force.
A plurality of the Court reached a different conclusion with regard to the Tennessee antieleetioneering statute considered earlier this Term in Burson v. Freeman, 504 U. S. 191 (1992). In light of the “logical connection” between electioneering and the State’s compelling interest in preventing voter intimidation and election fraud — an inherent connection borne out by a “long history” and a “widespread and time-tested consensus,” id,., at 206, 208, n. 10, 211 — the plurality concluded that it was faced with one of those “rare case[s]” in which the use of a facially content-based restriction was justified by interests unrelated to the suppression of ideas, id., at 211; see also id., at 213 (Kennedy, J., concurring). Justice White and Justice Stevens are therefore quite mistaken when they seek to convert the Burson plurality’s passing comment that “[t]he First Amendment does not require States to regulate for problems that do not exist,” id., at 207, into endorsement of the revolutionary proposition that the suppression of particular ideas ean b'e justified when only those ideas have been a source of trouble in the past. Post, at 405 (White, J., concurring in judgment); post, at 434 (Stevens, J., concurring in judgment).