OPINION OF THE COURT
ALITO, Circuit Judge:The plaintiffs in this case challenge the constitutionality of a public school district’s “anti-harassment” policy, arguing that it violates the First Amendment’s guarantee of freedom of speech.1 The District Court, concluding that the policy prohibited no more speech than was already unlawful under federal and state anti-discrimination laws, held that the policy is constitutional and entered judgment for the school district. We reverse.
I.
A.
In August 1999, the State College Area School District (“SCASD”) adopted an Anti-Harassment Policy (“the Policy”). The full text of the Policy is reproduced in the Appendix to this opinion; we will briefly review the most relevant portions here.
The Policy begins by setting forth its goal — “providing all students with a safe, secure, and nurturing school environment” — and noting that “[disrespect among members of the school community is unacceptable behavior which threatens to disrupt the school environment and well being of the individual.” The second paragraph contains what appears to be the Policy’s operative definition of harassment:
Harassment means verbal or physical conduct based on one’s actual or perceived race, religion, color, national origin, gender, sexual orientation, disability, or other personal characteristics, and which has the purpose or effect of substantially interfering with a student’s educational performance or creating an intimidating, hostile or offensive environment.
The Policy continues by providing several examples of “harassment”:
Harassment can include any unwelcome verbal, written or physical conduct *203which offends, denigrates or belittles an individual because of any of the characteristics described above. Such conduct includes, but is not limited to, unsolicited derogatory remarks, jokes, demeaning comments or behaviors, slurs, mimicking, name calling, graffiti, innuendo, gestures, physical contact, stalking, threatening, bullying, extorting or the display or circulation of written material or pictures.
These examples are followed by a lengthy section captioned “Definitions,” which defines various types of prohibited harassment, including “Sexual harassment,” “Racial and color harassment,” “Harassment on the basis of religion,” “Harassment based on national origin,” “Disability harassment,” and “Other harassment” on the basis of characteristics such as “clothing, physical appearance, social skills, peer group, intellect, educational program, hobbies or values, etc.” The definitions state that harassment “can include unwelcome verbal, written or physical conduct directed at” the particular characteristic. Examples of specific types of harassment are also provided. For example, “Racial and color harassment” is said to include “nicknames emphasizing stereotypes, racial slurs, comments on manner of speaking, and negative references to racial customs.” Religious harassment reaches “derogatory comments regarding surnames, religious tradition, or religious clothing, or religious slurs or graffiti.” National origins harassment includes “negative comments regarding surnames, manner of speaking, customs, language, or ethnic slurs.” Harassment on the basis of sexual orientation extends to “negative name calling and degrading behavior.” Disability harassment encompasses “imitating manner of speech or movement.”
The Policy provides that “[a]ny harassment of a student by a member of the school community is a violation of this policy.”2 It establishes procedures for the reporting, informal mediation, and formal resolution of complaints. In addition, the Policy sets a list of punishments for harassment, “including but not limited to warning, exclusion, suspension, expulsion, transfer, termination, discharge ..., training, education, or counseling.”
B.
Plaintiff David Saxe is a member of the Pennsylvania State Board of Education and serves as an unpaid volunteer for SCASD. He is the legal guardian of both student-plaintiffs, who are enrolled in SCASD schools. After the Anti-Harassment Policy was adopted, Saxe filed suit in District Court, alleging that the Policy was facially unconstitutional. under the First Amendment’s free speech clause.3 In his Complaint, he alleged that
[a]ll Plaintiffs openly and sincerely identify themselves as Christians. They be-' lieve, and their religion teaches, that homosexuality is a sin. Plaintiffs further believe that they have a right to speak out about the sinful nature and harmful effects of homosexuality. Plaintiffs also feel compelled by their religion to speak out on other topics, especially moral issues.
(App.27.) Plaintiffs further alleged that they feared that they were likely to be punished under the Policy for speaking out about their religious beliefs, engaging in symbolic activities reflecting those beliefs, and distributing religious literature. (App.27-28.) They sought to have the Policy declared unconstitutionally vague and *204overbroad and its operation permanently enjoined.
The District Court found that Saxe had standing to mount a facial challenge but granted SCASD’s motion to dismiss on the pleadings, holding that the Policy was facially constitutional. See Saxe v. State College Area School District, 77 F.Supp.2d 621 (M.D.Pa.1999). The Court found that the Policy’s operative definition of harassment was contained in its second paragraph, which, as the Court read it, prohibited “language or conduct which is based on specified characteristics and which has the effect of ‘substantially interfering with a student’s educational performance’ or which creates a hostile educational atmosphere.” Id. at 625. The Court went on to observe that this standard is similar to “that used by courts and agencies to define harassment for purposes of Title VII, Title IX, the Pennsylvania Human Relations Act, etc.” Id. Consequently, the Court held that the Policy does not prohibit “anything that is not already prohibited by law” and therefore cannot be unconstitutional. Id. at 626. Rejecting the plaintiffs’ vagueness argument, the Court asserted that “a more precise definition of harassment, like Justice Stewart’s famous description of ‘pornography,’ may be virtually impossible.” Id. at 625. Plaintiffs appealed.
II.
The District Court dismissed the plaintiffs’ free speech claims based on its conclusion that “harassment,” as defined by federal and state anti-discrimination statutes, is not entitled to First Amendment protection. The Court rejected the plaintiffs’ characterization of the Policy as a “hate speech code,” holding instead that it merely prohibits harassment that is already unlawful under state and federal law. The Court observed:
Harassment has never been considered to be protected activity under the First Amendment. In fact, the harassment prohibited under the Policy already is unlawful. The Policy is a tool which gives SCASD the ability to take action itself against harassment which may subject it to civil liability.
Saxe, 77 F.Supp.2d at 627.
We disagree with the District Court’s reasoning. 'There is no categorical “harassment exception” to the First Amendment’s free speech clause. Moreover, the SCASD Policy prohibits a substantial amount of speech that would not constitute actionable harassment under either federal or state law.
A.
Because the District Court based its holding on a determination that the Policy simply replicated existing law, we begin by briefly reviewing the scope of the applicable anti-harassment statutes. At the federal level, discriminatory harassment in the public schools is governed primarily by two statutes. Title VI of the Civil Rights Act of 1964 provides that “[n]o person in the United States shall, on the ground of race, color, or national origin, be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any program or activity receiving Federal financial assistance.” 42 U.S.C. § 2000d. Title IX of the Education Amendments of 1972 further provides that “[n]o person ... shall, on the basis of sex, be excluded from participation in, be denied the benefits of, or be subjected to discrimination under any educational program or activity receiving federal financial assistance.” 20 U.S.C. § 1681(a). Although less often involved in harassment cases, the Rehabilitation Act of 1973, 29 U.S.C. § 794, makes it unlawful for programs receiving federal assistance to discriminate on the basis of disability or age.4
*205The federal courts have held that these statutes create a private right of action similar to that available under Title VII, which prohibits discrimination in the workplace. Most significantly for this case, the Supreme Court has recognized that a public school student may bring suit against a school under Title IX for so-called “hostile environment” harassment. Davis v. Monroe County Board of Education, 526 U.S. 629, 119 S.Ct. 1661, 143 L.Ed.2d 839 (1999); Franklin v. Gwinnett County Public Schools, 503 U.S. 60, 74-75, 112 S.Ct. 1028, 117 L.Ed.2d 208 (1992).
The concept of “hostile environment” harassment originated in a series of Title VII cases involving sexual harassment in the workplace. In Meritor Savings Bank v. Vinson, 477 U.S. 57, 106 S.Ct. 2399, 91 L.Ed.2d 49 (1986), the Supreme Court held that Title VII prohibits abusive and discriminatory conduct that creates a “hostile environment” — that is, harassment so severe or pervasive as “to alter the conditions of the victim’s employment and create an abusive working environment.” Id. at 67, 106 S.Ct. 2399. In Harris v. Forklift Systems, Inc., 510 U.S. 17, 114 S.Ct. 367, 126 L.Ed.2d 295 (1993), the Court clarified that in order for conduct to constitute harassment under a “hostile environment” theory, it must both: (1) be viewed subjectively as harassment by the victim and (2) be objectively severe or pervasive enough that a reasonable person would agree that it is harassment. See id. at 21-22, 114 S.Ct. 367. The Court emphasized that the objective prong of this inquiry must be evaluated by looking at the “totality of the circumstances.” “These may include,” the Court observed, “the frequency of the discriminatory conduct; its severity; whether it is physically threatening or humiliating, or a mere offensive utterance; and whether it unreasonably interferes with an employee’s work performance.” Id. at 23, 114 S.Ct. 367. See also Oncale v. Sundowner Offshore Services, Inc., 523 U.S. 75, 80, 118 S.Ct. 998, 140 L.Ed.2d 201 (1998) (“Conduct that is not severe or pervasive enough to create an objectively hostile or abusive work environment — an environment that a reasonable person would find hostile or abusive — is beyond Title VII’s purview.”). In defining the contours of this concept, the Court has repeatedly stated that Title YII is not violated by the “mere utterance of an ... epithet which engenders offensive feelings in an employee” or by mere “ ‘discourtesy or rudeness,’ unless so severe or pervasive as to constitute an objective change in the conditions of employment.” Faragher v. City of Boca Raton, 524 U.S. 775, 787, 118 S.Ct. 2275, 141 L.Ed.2d 662 (1998).
The Supreme Court has extended an analogous cause of action to students under Title IX. Originally, such claims were limited to cases involving harassment of a student by a teacher or other agent of the school. See Franklin v. Gwinnett County Public Schools, supra. However, in 1999, in Davis v. Monroe County Board of Education, supra, the Court held that Title IX also permits a plaintiff to recover damages from a federally funded educational institution for certain cases of student-on-student sexual harassment. To recover in such a case,
a plaintiff must establish sexual harassment of students that is so severe, pervasive, and objectively offensive, and *206that so undermines and detracts from the victims’ educational experience, that the victim students are effectively denied equal access to an institution’s resources and opportunities.
Id. at 651, 119 S.Ct. 1661. This determination “ ‘depends on a constellation of surrounding circumstances, expectations, and relationships,’ including, but not limited to, the ages of the harasser and the victim, and the number of individuals involved.” Id. (quoting Oncale, 523 U.S. at 82, 118 S.Ct. 998). The Court stressed that “[djamages are not available for simple acts of teasing and name-calling among school children, even where these comments target differences in gender.” Id. at 652, 119 S.Ct. 1661. Rather, private damages actions against the school are limited to cases in which the school “acts with deliberate indifference to known acts of harassment,” and those acts have “a systemic effect on educational programs and activities.” Id. at 633, 653, 119 S.Ct. 1661.5
B.
With this framework in mind, we now turn to the District Court’s assertion that “harassment has never been considered to be protected activity under the First Amendment.” The District Court’s categorical pronouncement exaggerates the current state of the case law in this area.
There is of course no question that non-expressive, physically harassing conduct is entirely outside the ambit of the free speech clause. But there is also no question that the free speech clause protects a wide variety of speech that listeners may consider deeply offensive, including statements that impugn another’s race or national origin or that denigrate religious beliefs. See, e.g., Brandenburg v. Ohio, 395 U.S. 444, 89 S.Ct. 1827, 23 L.Ed.2d 430 (1969); Cantwell v. Connecticut, 310 U.S. 296, 60 S.Ct. 900, 84 L.Ed. 1213 (1940). When laws against harassment attempt to regulate oral or written expression on such topics, however detestable the views expressed may be, we cannot turn a blind eye to the First Amendment implications. “Where pure expression is involved,” anti-discrimination law “steers into the territory of the First Amendment.” DeAngelis v. El Paso Mun. Police Officers Ass’n, 51 F.3d 591, 596 (5th Cir.1995).
This is especially true because, as the Fifth Circuit has noted, when anti-discrimination laws are “applied to ... harassment claims founded solely on verbal insults, pictorial or literary matter, the statutefs] impose[ ] content-based, viewpoint-discriminatory restrictions on speech.” DeAngelis, 51 F.3d at 596-97. Indeed, a disparaging comment directed at an individual’s sex, race, or some other personal characteristic has the potential to create an “hostile environment”—and thus come within the ambit of anti-discrimination laws—precisely because of its sensitive subject matter and because of the odious viewpoint it expresses.6
*207This sort of content- or viewpoint-based restriction is ordinarily subject to the most exacting First Amendment scrutiny. This point was dramatically illustrated in R.A.V. v. City of St. Paul, 505 U.S. 377, 112 S.Ct. 2538, 120 L.Ed.2d 305 (1992), in which the Supreme Court struck down a municipal hate-speech ordinance prohibiting “fighting words” that aroused “anger, alarm or resentment on the basis of race, color, creed, religion or gender.” Id. at 377, 112 S.Ct. 2538. While recognizing that fighting words generally are unprotected by the First Amendment, the Court nevertheless found that the ordinance unconstitutionally discriminated on the basis of content and viewpoint:
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, color, etc. tolerance and equality, but could not be used by that speaker’s opponents.
Id. at 391, 112 S.Ct. 2538. Striking down the law, the Court concluded that “[t]he point of the First Amendment is that majority preferences must be expressed in some fashion other than silencing speech on the basis of content.” Id. at 392, 112 S.Ct. 2538.
Loosely worded anti-harassment laws may pose some of the same problems as the St. Paul hate speech ordinance: they may regulate deeply offensive and potentially disruptive categories of speech based, at least in part, on subject matter and viewpoint. Although the Supreme Court has written extensively on the scope of workplace harassment, it has never squarely addressed whether harassment, when it takes the form of pure speech, is exempt from First Amendment protection. See Weller v. Citation Oil & Gas Corp., 84 F.3d 191, 194 n. 6 (5th Cir.1996) (noting that the Supreme Court has “provided] little guidance whether conduct targeted for its expressive content ... may be regulated under Title VII”); Aguilar v. Avis Rent A Car Sys., Inc., 21 Cal.4th 121, 87 Cal.Rptr.2d 132, 980 P.2d 846, 863 (Cal.1999) (Werdegar, J., concurring) (“No decision by the United States Supreme Court has, as yet, declared that the First Amendment permits restrictions on speech creating a hostile work environment.”).7
SCASD relies heavily on a passage in R.A.V. in which the Court suggested in dictum that at least some harassing speech does not warrant First Amendment protection:
[Sjince words 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 proscribable class of speech can be swept up incidentally within the reach of a statute directed at conduct rather than speech [citing Barnes v. Glen Theatre, Inc., 501 U.S. 560, 571, 111 S.Ct. 2456, 115 L.Ed.2d 504 (1991); FTC v. Superior Court Trial Lawyers Ass’n, 493 U.S. 411, 425-32, 110 S.Ct. 768, 107 L.Ed.2d 851 (1990); and United States v. O’Brien, 391 U.S. 367, 376-77, 88 S.Ct. *2081673, 20 L.Ed.2d 672 (1968) ]. 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. 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.
R.A.V., 505 U.S. at 389, 112 S.Ct. 2538 (other citations omitted) (emphasis added).
This passage suggests that government may constitutionally prohibit speech whose non-expressive qualities promote discrimination. For example, a supervisor’s statement “sleep with me or you’re fired” may be proscribed not on the ground of any expressive idea that the statement communicates, but rather because it facilitates the threat of discriminatory conduct. Despite the purely verbal quality of such a threat, it surely is no more “speech” for First Amendment purposes than the robber’s demand “your money or your life.” Accord NLRB v. Gissel Packing Co., 395 U.S. 575, 617, 618, 89 S.Ct. 1918, 23 L.Ed.2d 547 (1969) (holding that employer’s “threat of retaliation” on basis of union membership was “without the protection of the First Amendment”) (citation and internal quotation marks omitted).8 Similarly, we see no constitutional problem with using an employer’s offensive speech as evidence of motive or intent in a case involving an allegedly discriminatory employment action. Accord Dawson v. Delaware, 503 U.S. 159, 112 S.Ct. 1093, 117 L.Ed.2d 309 (1992) (“The Constitution does not erect a per se barrier to the admission of evidence concerning one’s beliefs and associations ... simply because those beliefs and associations are protected by the First Amendment.”).
*209The previously quoted passage from R.A.V., however, does not necessarily mean that anti-discrimination laws are categorically immune from First Amendment challenge when they are applied to prohibit speech solely on the basis of its expressive content. See DeAngelis, 51 F.3d at 596 n. 7; John E. Nowak & Ronald D. Rotunda, Constitutional Law § 16.39, at 1116 (5th ed.1995). “Harassing” or discriminatory speech, although evil and offensive, may be used to communicate ideas or emotions that nevertheless implicate First Amendment protections. As the Supreme Court has emphatically declared, “[i]f 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 offensive or disagreeable.” Texas v. Johnson, 491 U.S. 397, 414, 109 S.Ct. 2533, 105 L.Ed.2d 342 (1989).
For this reason, we cannot accept SCASD’s contention that the application of anti-harassment law to expressive speech can be justified as a regulation of the speech’s “secondary effects.” R.AV. did acknowledge that content-discriminatory speech restrictions may be permissible when the content classification merely “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.’ ” R.AV, 505 U.S. at 389, 112 S.Ct. 2538 (quoting Renton v. Playtime Theatres, Inc., 475 U.S. 41, 48, 106 S.Ct. 925, 89 L.Ed.2d 29 (1986)). The Supreme Court has made it clear, however, that the government may not prohibit speech under a “secondary effects” rationale based solely on the emotive impact that its offensive content may have on a listener: “Listeners’ reactions to speech are not the type of ‘secondary effects’ we referred to in Renton .... The emotive impact of speech on its audience is not a ‘secondary effect.’ ” Boos v. Barry, 485 U.S. 312, 321, 108 S.Ct. 1157, 99 L.Ed.2d 333 (1988); see also United States v. Playboy Entertainment Group, 529 U.S. 803, 120 S.Ct. 1878, 1885, 146 L.Ed.2d 865 (2000) (“The overriding justification for the regulation is concern for the effect of the subject matter on [listeners].... This is the essence of content-based regulation.”); Forsyth County v. Nationalist Movement, 505 U.S. 123, 134, 112 S.Ct. 2395, 120 L.Ed.2d 101 (1992) (“Listeners’ reaction to speech is not a content-neutral basis for regulation.”). Nor do we believe that the restriction of expressive speech on the basis of its content may be characterized as a mere “time, place and manner” regulation. See Reno v. ACLU, 521 U.S. 844, 879, 117 S.Ct. 2329, 138 L.Ed.2d 874 (1997) (“time, place and manner” analysis not applicable when statute “regulates speech on the basis of its content”); Pacific Gas & Elec. Co. v. Public Util. Comm’n, 475 U.S. 1, 20, 106 S.Ct. 903, 89 L.Ed.2d 1 (1986) (“[flor a time, place, or manner regulation to be valid, it must be neutral as to the content of the speech”); Consolidated Edison Co. v. Public Serv. Comm’n, 447 U.S. 530, 536, 100 S.Ct. 2326, 65 L.Ed.2d 319 (1980) (“a constitutionally permissible time, place, or manner restriction may not be based upon either the content or subject matter of speech”).
In short, we see little basis for the District Court’s sweeping assertion that “harassment” — at least when it consists of speech targeted solely on the basis of its expressive content — “has never been considered to be protected activity under the First Amendment.” Such a categorical rule is without precedent in the decisions of the Supreme Court or this Court, and it belies the very real tension between anti-harassment laws and the Constitution’s guarantee of freedom of speech.
We do not suggest, of course, that no application of anti-harassment law to expressive speech can survive First Amendment scrutiny. Certainly, preventing discrimination in the workplace — and in the schools — is not only a legitimate, but a compelling, government interest. See, e.g., Board of Directors of Rotary Internation*210al v. Rotary Club of Duarte, 481 U.S. 537, 549, 107 S.Ct. 1940, 95 L.Ed.2d 474 (1987). And, as some courts and commentators have suggested, speech may be more readily subject to restrictions when a school or workplace audience is “captive” and cannot avoid the objectionable speech. See, e.g., Aguilar, 87 Cal.Rptr.2d 132, 980 P.2d at 871-73 (Werdegar, J., concurring). We simply note that we have found no categorical rule that divests “harassing” speech, as defined by federal anti-discrimination statutes, of First Amendment protection.
C.
In any event, we need not map the precise boundary between permissible anti-discrimination legislation ahd impermissible restrictions on First Amendment rights today. Assuming for present purposes that the federal anti-discrimination laws are constitutional in all of their applications to pure speech, we note that the SCASD Policy’s reach is considerably broader.
For one thing, the Policy prohibits harassment based on personal characteristics that are not protected under federal law. Titles VI and IX, taken together with the other relevant federal statutes, cover only harassment based on sex, race, color, national origin, age and disability. The Policy, in contrast, is much broader, reaching, at the extreme, a catch-all category of “other personal characteristics” (which, the Policy states, includes things like “clothing,” “appearance,” “hobbies and values,” and “social skills”). Insofar as the policy attempts to prevent students from making negative comments about each others’ “appearance,” “clothing,” and “social skills,” it may be brave, futile, or merely silly. But attempting to proscribe negative comments about “values,” as that term is commonly used today, is something else altogether. By prohibiting disparaging speech directed at a person’s “values,” the Policy strikes at the heart of moral and political discourse — the lifeblood of constitutional self government (and democratic education) and the core concern of the First Amendment. That speech about “values” may offend is not cause for its prohibition, but rather the reason for its protection: “a principal ‘function of free speech under our system of government is to invite dispute. It may indeed best serve its high purpose when it induces a condition of unrest, creates dissatisfaction with conditions as they are, or even stirs people to anger.’ ” Texas v. Johnson, 491 U.S. 397, 408-09, 109 S.Ct. 2533, 105 L.Ed.2d 342 (1989) (quoting Terminiello v. Chicago, 337 U.S. 1, 4, 69 S.Ct. 894, 93 L.Ed. 1131 (1949)). No court or legislature has ever suggested that unwelcome speech directed at another’s “values” may be prohibited under the rubric of anti-discrimination.
We do not suggest, of course, that a public school may never adopt regulations more protective than existing law; it may, provided that those regulations do not offend the Constitution. Such regulations cannot be insulated from First Amendment challenge, however, based on the argument that they do no more than prohibit conduct that is already unlawful.
Moreover, the Policy’s prohibition extends beyond harassment that objectively denies a student equal access to a school’s education resources. Even on a narrow reading, the Policy unequivocally prohibits any verbal or physical conduct that is based on an enumerated personal characteristic and that “has the purpose or effect of substantially interfering with a student’s educational performance or creating an intimidating, hostile or offensive environment.” (emphasis added). Unlike federal anti-harassment law, which imposes liability only when harassment has “a systemic effect on educational programs and activities,” Davis, 526 U.S. at 633, 119 S.Ct. 1661 (emphasis added), the Policy extends to speech that merely has the “purpose” of harassing another. This formulation, by focusing on the speaker’s motive rather than the effect of speech on the learning *211environment, appears to sweep in those “simple acts of teasing and name-calling” that the Davis Court explicitly held were insufficient for liability.
D.
The District Court justifies its ruling by a syllogism: (1) the SCASD Policy covers only speech that is already prohibited under federal and state anti-harassment laws; (2) such prohibited speech is not entitled to First Amendment protection; (3) therefore, the Policy poses no First Amendment problems. This reasoning is flawed in both its major and minor premises. First, the Policy — even narrowly interpreted — covers substantially more speech than applicable federal and state laws. Second, the courts have never embraced a categorical “harassment exception” from First Amendment protection for speech that is within the ambit of federal anti-discrimination laws.
III.
Accordingly, we must examine whether the Policy may be justified as a permissible regulation of speech within the schools.
A.
We begin by reviewing the Supreme Court’s cases demarcating the scope of a student’s right to freedom of expression while in school.9 The Court set out the framework for student free speech claims in Tinker v. Des Moines Independent Community School District, 393 U.S. 503, 89 S.Ct. 733, 21 L.Ed.2d 731 (1969). In Tinker, a group of students was suspended for wearing black armbands to protest American involvement in the Vietnam War. The Court held that the wearing of the armbands to make a political statement was “closely akin to ‘pure speech’ ” and thus was constitutionally protected. Id. at 505, 89 S.Ct. 733. Taking as its premise that “[i]t can hardly be argued that either students or teachers shed their constitutional rights to freedom of speech or expression at the schoolhouse gate,” id. at 506, 89 S.Ct. 733, the Court reasoned that
[t]he school officials banned and sought to punish petitioners for a silent, passive expression of opinion, unaccompanied by any disorder or disturbance on the part of the petitioners. There is here no evidence whatever of the petitioners’ interference, actual or nascent, with the school’s work or of collision with the rights of other students to be secure and left alone. Accordingly, this case does not concern speech or action that intrudes upon the work of the school or the rights of other students.
Id. at 504, 89 S.Ct. 733. Significantly, the Court emphasized that “undifferentiated fear or apprehension of disturbance is not enough to overcome the right to freedom of expression.” Id. at 508, 89 S.Ct. 733.
Under Tinker, then, regulation of student speech is generally permissible only when the speech would substantially disrupt or interfere with the work of the school or the rights of other students. As subsequent federal cases have made clear, Tinker requires á specific and significant fear of disruption, not just some remote apprehension of disturbance. In Chandler v. McMinnville School District, 978 F.2d 524 (9th Cir.1992), for example, a middle school punished students who wore “SCAB” buttons to protest replacement teachers during a strike. Because the school had failed to present any evidence that the buttons were “inherently disruptive” to school activities, the court held that students could proceed with their First Amendment claim. In Chalifoux v. New Caney Independent School District, 976 F.Supp. 659 (S.D.Tex.1997), a high *212school student challenged his school’s policy against gang-related apparel. The school applied the ban to prohibit the plaintiff, a devout Catholic, from wearing a rosary to school on the ground that some gangs had adopted the rosary as then-identifying symbol. The court held that the ban failed to satisfy Tinker's substantial disruption test:
[Although Plaintiffs wore them rosaries outside their shirts-for several months, they were never misidentified as gang members nor approached by gang members. There also was no evidence that they attracted the attention of other students because of their rosaries.... Accordingly, the Court finds that there was insufficient evidence of actual disruption at New Caney High School, or that there was substantial reason for NCISD to anticipate a disruption, to justify the infringement on Plaintiffs’ religiously-motivated speech.
Chalifoux, 976 F.Supp. at 667. Finally, in Clark v. Dallas Independent School District, 806 F.Supp. 116, 120 (N.D.Tex.1992), the court held that a high school could not prohibit its students from distributing religious tracts on school grounds. Again citing Tinker, the court held that “Defendants have failed to establish that Plaintiffs’ distribution of the religious tracts gave rise to a material or substantial disruption of the operation” of the school. Id. at 120. Noting that the only evidence of disruption was the objection of several other students, the court observed that “[i]f school officials were permitted to prohibit expression to which other students objected, absent any further justification, the officials would have a license to prohibit virtually every type of expression.” Id.
The Tenth Circuit’s recent decision in West v. Derby Unified School District No. 260, 206 F.3d 1358 (10th Cir.2000), which reached a different result, nevertheless confirms Tinker’s requirements of specificity and concreteness. In West, a middle school student was suspended for drawing a Confederate flag in math class under a school policy providing that a “student shall not racially harass or intimidate another student by name calling, using racial or derogatory slurs,[or] wearing or possession of items depicting or implying racial hatred or prejudice.” Id. at 1361. The Court upheld the suspension under Tinker’s substantial disruption standard, finding that the school had demonstrated a concrete threat of substantial disruption:
[B]ased upon recent past events, Derby School District officials • had reason to believe that a student’s display of the Confederate flag might cause disruption and interfere with the rights of other students to be secure and let alone.... The district experienced a series of racial incidents [including “hostile confrontations” and at least one fight] in 1995, some of which were related to the Confederate flag.... The Racial Harassment policy enacted in response to this situation was clearly something more than a mere desire to avoid the discomfort and unpleasantness that always accompany an unpopular viewpoint. The history of racial tension in the district made administrators’ and parents’ concerns about future substantial disruptions from possession of Confederate flag symbols at school reasonable.
Id. at 1366 (citation omitted). As West makes clear, the mere desire to avoid “discomfort” or “unpleasantness” is not enough to justify restricting student speech under Tinker. However, if a school can point to a well-founded expectation of disruption—especially one based on past incidents arising out of similar speech—the restriction may pass constitutional muster.
Since Tinker, the Supreme Court has carved out a number of narrow categories of speech that a school may restrict even without the threat of substantial disruption. In Bethel School District No. 403 v. Fraser, 478 U.S. 675, 106 S.Ct. 3159, 92 L.Ed.2d 549 (1986), the Court upheld the school’s suspension of a high school stu*213dent who, at a school assembly, nominated a peer for class office through “an elaborate, graphic, and explicit sexual metaphor.” Id. at 677, 106 S.Ct. 3159. Holding that the student’s expression was not protected by the First Amendment, the Court reasoned that
[t]he schools, as instruments of the state, may determine that the essential lessons of civil, mature conduct cannot be conveyed in a school that tolerates lewd, indecent, or offensive speech and conduct such as that indulged in by this confused boy.
Id. at 683, 106 S.Ct. 3159. Distinguishing Cohen v. California, 403 U.S. 15, 91 S.Ct. 1780, 29 L.Ed.2d 284 (1971), in which the Court struck down an adult’s conviction for wearing a jacket bearing an obscenity in a public courthouse, the Court explained that
[i]t does not follow ... that, simply because the use of an offensive form of expression may not be prohibited to adults making what the speaker considers a political point, the same latitude must be permitted to children in public school.... “[T]he First Amendment gives a high school student the classroom right to wear Tinker’s armband, but not Cohen’s jacket.”
Fraser, 478 U.S. at 683, 106 S.Ct. 3159 (citations omitted). According to Fraser, then, there is no First Amendment protection for “lewd,” “vulgar,” “indecent,” and “plainly offensive” speech in school. Fraser permits a school to prohibit words that “offend for the same reasons that obscenity offends” — a dichotomy neatly illustrated by the comparison between Cohen’s jacket and Tinker’s armband. Fraser, 478 U.S. at 685, 106 S.Ct. 3159 (quoting FCC v. Pacifica Foundation, 438 U.S. 726, 746, 98 S.Ct. 3026, 57 L.Ed.2d 1073 (1978)); see also Hazelwood School District v. Kuhlmeier, 484 U.S. 260, 286 n. 2, 108 S.Ct. 562, 98 L.Ed.2d 592 (Brennan, J., dissenting) (Fraser exception limited “to the appropriateness of the manner in which the message is conveyed, not of the message’s content”); East High Gay/Straight Alliance v. Board of Educ. of Salt Lake City School Dist., 81 F.Supp.2d 1166, 1193 (D.Utah 1999) (“Fraser speaks to the form and manner of student speech, not its substance. It addresses the mode of expression, not its content or viewpoint.”).
Finally, in Hazelwood School District v. Kuhlmeier, 484 U.S. 260, 108 S.Ct. 562, 98 L.Ed.2d 592 (1988), the Court upheld, against First Amendment challenge, a principal’s deletion of student articles on teen pregnancy from a school-sponsored newspaper. Distinguishing Tinker, the Court noted the school had not opened the newspaper up as a public forum and therefore could “exercis[e] editorial control over the style and content of student speech in school-sponsored expressive activities as long as [its] actions are reasonably related to legitimate pedagogical concerns.” Id. at 273, 108 S.Ct. 562 (emphasis added). As the Court reasoned,
[t]he question whether the First Amendment requires a school to tolerate particular student speech — the question that we addressed in Tinker — is different from the question whether the First Amendment requires a school affirmatively to promote particular student speech. The former question addresses educators’ ability to silence a student’s personal expression that happens to occur on the school premises. The latter question concerns educators’ authority over school-sponsored publications, theatrical productions, and other expressive activities that students, parents, and members of the public might reasonably perceive to bear the imprimatur of the school_Educators are entitled to exercise greater control over this second form of student expression....
Id. at 270-71, 108 S.Ct. 562. In Rosenberger v. Rector & Visitors of University of Virginia, 515 U.S. 819, 115 S.Ct. 2510, 132 L.Ed.2d 700 (1995), the Court made clear that Hazelwood’s, permissive “legitimate pedagogical concern” test governs only when a student’s school-sponsored speech *214could reasonably be viewed as speech of the school itself:
[W]hen the State is the speaker, it may make content-based choices. When the University determines the content of the education it provides, it is thé University speaking, and we have permitted the government to regulate the content of what is or is not expressed when it is the speaker or when it enlists private entities to convey its own message.... It does not follow, however ... that viewpoint-based restrictions are proper when the University does not itself speak or subsidize transmittal of a message it favors but instead encourage[s] a diversity of views from private speakers. A holding that the University may not discriminate based on the viewpoint of private persons whose speech it facilitates does not restrict the University’s own speech, which is controlled by different principles. See, e.g., ... Hazelwood School Dist. v. Kuhlmeier, [484 U.S. at 270-72, 108 S.Ct. 562].
Rosenberger, 515 U.S. at 834, 115 S.Ct. 2510. Similarly, a post-Hazelwood case from the Seventh Circuit illustrates that school “sponsorship” of student speech is not lightly to be presumed. See Hedges v. Wauconda Comm. Unit Sch. Dist. No. 118, 9 F.3d 1295, 1299 (7th Cir.1993). In striking down a blanket prohibition against distributing religious materials on school grounds, the Hedges Court rejected the argument that the ban was justified under Hazelwood because observers might “infer that the school endorses whatever it permits”:
[The School District] proposes to throw up its hands, declaring that because misconceptions are possible it may silence its pupils, that the best defense against misunderstanding is censorship.... Public belief that the government is partial does not permit the government to become partial. Students therefore may hand out literature even if the recipients would misunderstand its provenance. The school’s proper response is to educate the audience rather than squelch the speaker.
Hedges, 9 F.3d at 1299; see also Burch v. Barker, 861 F.2d 1149, 1159 (9th Cir.1988) (“under ground newspaper” distributed on school grounds could not reasonably be viewed as school-sponsored).
To summarize: Under Fraser, a school may categorically prohibit lewd, vulgar or profane language. Under Hazelwood, a school may regulate school-sponsored speech (that is, speech that a reasonable observer would view as the school’s own speech) on the basis of any legitimate pedagogical concern. Speech falling outside of these categories is subject to Tinker’s general rule: it may be regulated only if it would substantially disrupt school operations or interfere with the right of others. See Chandler, 978 F.2d at 529; Pyle v. South Hadley Sch. Comm., 861 F.Supp. 157, 166 (D.Mass.1994).
IV.
We turn now to the SCASD Policy itself. Saxe levies facial challenges against the Policy on both overbreadth and vagueness grounds. Because we hold that the Policy, even narrowly read, is unconstitutionally overbroad, we do not reach the merits of Saxe’s vagueness claim.
A.
A regulation is unconstitutional on its face on overbreadth grounds where there is a “a likelihood that the statute’s very existence will inhibit free expression” by “inhibiting the speech of third parties who are not before the Court.” Members of City Council v. Taxpayers for Vincent, 466 U.S. 789, 799, 104 S.Ct. 2118, 80 L.Ed.2d 772 (1984). To render a law unconstitutional, the overbreadth must be “not only real but substantial in relation to the statute’s plainly legitimate sweep.” Broadrick v. Oklahoma, 413 U.S. 601, 615, 93 S.Ct. 2908, 37 L.Ed.2d 830 (1973).
On first reading, the Policy on its face appears both unconstitutionally vague *215and overbroad. As an initial matter, the Policy contains several separate passages, each of which could be read as embodying its operative definition of banned speech. The Policy's second paragraph sets forth one definition:
Harassment means verbal or physical conduct based on one’s actual or perceived race, religion, color, national origin, gender, sexual orientation, disability, or other personal characteristics, and which has the purpose or effect of substantially interfering with a student’s educational performance or creating an intimidating, hostile or offensive environment.
This, however, is immediately followed two paragraphs later by a statement that harassment under the Policy “can include any unwelcome verbal, written or physical conduct which offends, denigrates or belittles an individual because of any of the characteristics described above.” In addition, in a separate section, the Policy purports to set out “definitions” for various categories of harassment that do not always coincide with the above-quoted language. Religious harassment, for example, is defined as “unwelcome verbal, written or physical conduct directed at the characteristics of a person’s religion, such as derogatory comments regarding surnames, religious tradition, or religious clothing, or religious slurs, or graffiti.”
Certainly, some of these purported definitions of harassment are facially over-broad. No one would suggest that a school could constitutionally ban “any unwelcome verbal ... conduct which offends ... an individual because of’ some enumerated personal characteristics. Nor could the school constitutionally restrict, without more, any “unwelcome verbal ... conduct directed at the characteristics of a person’s religion.” The Supreme Court has held time and again, both within and outside of the school context, that the mere fact that someone might take offense at the content of speech is not sufficient justification for prohibiting it. See Tinker, 393 U.S. at 509, 89 S.Ct. 733 (school may not prohibit speech based on the “mere desire to avoid the discomfort and unpleasantness that always accompany an unpopular viewpoint”); Texas v. Johnson, 491 U.S. 397, 414, 109 S.Ct. 2533, 105 L.Ed.2d 342 (1989) (“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.”); Street v. New York, 394 U.S. 576, 592, 89 S.Ct. 1354, 22 L.Ed.2d 572 (1969) (“It is firmly settled that ... the public expression of ideas may not be prohibited merely because the ideas are themselves offensive to some of their hearers.”); see also Doe v. University of Michigan, 721 F.Supp. 852, 863 (E.D.Mich.1989) (striking down university speech code: “Nor could the University proscribe speech simply because it was found to be offensive, even gravely so, by large numbers of people.”).
Before declaring the Policy unconstitutional, however, we must first determine whether it is susceptible to a reasonable limiting construction: “the elementary rule is that every reasonable construction must be resorted to, in order to save a statute from unconstitutionality.” 10 Stretton v. Disciplinary Bd. of the Supreme Court of Pennsylvania, 944 F.2d 137, 144 (3d Cir.1991) (citations *216omitted); see also Hoffman Estates v. Flipside, Hoffman Estates, 455 U.S. 489, 494 n. 4, 102 S.Ct. 1186, 71 L.Ed.2d 362 (1982) (“In evaluating a facial challenge to a state law, a federal court must, of course, consider any limiting construction.”); Broadrick, 413 U.S. at 617 n. 16, 93 S.Ct. 2908 (“a federal court must determine what a state statute means before it can judge its facial unconstitutionality”).
When the Policy is read as a whole, it appears that, its operative definition of prohibited harassment is contained in the above-quoted second paragraph, which requires that speech either “substantially interfer[e] with a student’s educational performance or creat[e] an intimidating, hostile or offensive environment.” The Policy’s fourth paragraph and “Definitions” section could reasonably be read as merely listing examples of conduct that might (but would not necessarily) violate this operative definition. On this narrow reading, the second paragraph would supply the Policy’s “formal” definition of prohibited harassment, but the other sections of the Policy could still be relevant in clarifying vague or ambiguous terms in that operative definition.
So narrowed, the Policy would require the following elements before speech could be deemed harassing: (1) verbal or phj^sical conduct (2) that is based on one’s actual or perceived personal characteristics and (3) that has the purpose or effect of either (3a) substantially interfering with a student’s educational performance or (3b) creating an intimidating hostile, or offensive environment.
It is apparent from these elements that SCASD cannot take solace in the relatively more permissive Fraser or Hazelwood standards. First, the Policy does not confine itself merely to vulgar or lewd speech; rather, it reaches any speech that interferes or is intended to interfere with educational performance or that creates or is intended to create a hostile environment. While some Fraser-type speech may fall within this definition, the Policy’s scope is clearly broader. Second, the Policy does not contain any geographical or contextual limitations; rather, it purports to cover “[a]ny harassment of a student by a member of the school community.” Thus, its strictures presumably apply whether the harassment occurs in a school sponsored assembly, in the classroom, in the hall between classes, or in a playground or athletic facility.11 Obviously, the Policy covers far more than just Hazelwood-type school-sponsored speech; it also sweeps in private student speech that merely “happens to occur on the school premises.” Hazelwood, 484 U.S. at 271, 108 S.Ct. 562. As a result, SCASD cannot rely on Hazel-wood’s, more lenient “legitimate pedagogical concern” test in defending the Policy from facial attack.
In short, the Policy, even narrowly read, prohibits a substantial amount of non-vulgar, non-sponsored student speech. SCASD must therefore satisfy the Tinker test by showing that the Policy’s restrictions are necessary to prevent substantial disruption or interference with the work of the school or the rights of other students. Applying this test, we conclude that the Policy is substantially overbroad.
As an initial matter, the Policy punishes not only speech that actually causes disruption, but also speech that merely intends to do so: by its terms, it covers speech “which has the purpose or effect of’ interfering with educational perfor-*217manee or creating a hostile environment. This ignores Tinkers requirement that a school must reasonably believe that speech will cause actual, material disruption before prohibiting it.
In addition, even if the “purpose” component is ignored, we do not believe that prohibited “harassment,” as defined by the Policy, necessarily rises to the level of a substantial disruption under Tinker. We agree that the Policy’s first prong, which prohibits speech that would “substantially interfere] with a student’s educational performance,” may satisfy the Tinker standard. The primary function of a public school is to educate its students; conduct that substantially interferes with the mission is, almost by definition, disruptive to the school environment.
The Policy's second criterion, however— which prohibits speech that “creat[es] an intimidating, hostile or offensive environment” — poses a more difficult problem. There are several possible grounds on which SCASD could attempt to justify this prohibition. First, SCASD could argue that it has an interest in avoiding liability for harassment under Franklin and Davis. However, because the Policy prohibits substantially more conduct than would give rise to liability under these cases, this justification is unavailing.
Second, SCASD could argue that speech creating a “hostile environment” may be banned because it “intrudes upon ... the rights of other students.” Tinker, 393 U.S. at 504, 89 S.Ct. 733. The precise scope of Tinker's “interference with the rights of others” language is unclear; at least one court has opined that it covers only independently tortious speech like libel, slander or intentional infliction of emotional distress. See Slotterback v. Interboro Sch. Dist., 766 F.Supp. 280, 289 n. 8 (E.D.Pa.1991); see also Kuhlmeier v. Hazelwood Sch. Dist., 795 F.2d 1368, 1375 (8th Cir.), rev’d on other grounds, 484 U.S. 260, 108 S.Ct. 562, 98 L.Ed.2d 592 (1988). In any case, it is certainly not enough that the speech is merely offensive to some listener. See, e.g., Rivera, 721 F.Supp. at 1191. Because the Policy’s “hostile environment” prong does not, on its face, require any threshold showing of severity or pervasiveness, it could conceivably be applied to cover any speech about some enumerated personal characteristics the content of which offends someone.12 This could include much “core” political and religious speech: the Policy’s “Definitions” section lists as examples of covered harassment “negative” or “derogatory” speech about such contentious issues as “racial customs,” “religious tradition,” “language,” “sexual orientation,” and “values.” Such speech, when it does not pose a realistic threat of substantial disruption, is within a student’s First Amendment rights.
Finally, SCASD might argue that the “hostile environment” prohibition is required to maintain an orderly and non-disruptive educational environment. However, as Tinker made clear, the “undifferentiated fear or apprehension of disturbance” is not enough to justify a restriction on student speech. Although SCASD correctly asserts that it has a compelling interest in promoting an educational environment that is safe and conducive to learning, it fails to provide any particularized reason as to why it anticipates substantial disruption from the broad swath of student speech prohibited under the Policy.
The Policy, then, appears to cover substantially more speech than could be prohibited under Tinker's substantial disruption test. Accordingly, we hold that the Policy is unconstitutionally overbroad.
*218V.
For the foregoing reasons, the judgment of the District Court is reversed.
. Plaintiffs also assert that the Policy violates the free speech guarantee of the Pennsylvania Constitution. However, plaintiffs fail to present any authority to show that Pennsylvania’s guarantees are any broader than the First Amendment's. Accordingly, we confine our discussion to the plaintiffs' federal constitutional claims.
. The school community, by die Policy’s terms, "includes, but is not limited to, all students, school employees, contractors, unpaid volunteers, school board members, and other visitors.” "School employees” include, but are not limited to, “all teachers, support staff, administrators, bus drivers, custodians, cafeteria workers, coaches, volunteers, and agents of the school.”
. In their complaint, plaintiffs also asserted a claim under the free exercise clause of the First Amendment, but they do not press this claim on appeal.
. The District Court also referred to two state statutes: the Pennsylvania Human Relations Act (PHRA) and the Pennsylvania criminal harassment statute. We do not believe that either of these statutes is particularly relevant to this appeal. The PHRA, 43 P.S.A. §§ 951 *205et seq., prohibits discrimination in employment, housing and public accommodations on the basis of "race, color, religious creed, ancestry, age, sex, national origin or non-job related handicap or disability.” 43 P.S.A. § 953. It has not been construed, however, to create a cause of action for "hostile environment” harassment of a public school student. Pennsylvania’s criminal harassment statute makes it a criminal offense when a person, with intent to harass, annoy or alarm another person, subjects or threatens to subject that person to unwelcome physical contact; follows that person in or about a public place; or behaves in a manner which alarms or seriously annoys that person and that serves no legitimate purpose. 18 P.S.A. § 2709. Clearly, this law covers a much narrower range of conduct than is implicated by the SCASD Policy.
. Although both Franklin and Davis dealt with sexual harassment under Title IX, we believe that their reasoning applies equally to harassment on the basis of the personal characteristics enumerated in Title VI and other relevant federal anti-discrimination statutes. Accord Monteiro v. Tempe Union High Sch. Dist., 158 F.3d 1022, 1032-33 (9th Cir.1998) (applying Title VI to student-on-student racial harassment).
. Most commentators including those who favor and oppose First Amendment protection for harassing speech, agree that federal anti-discrimination law regulates speech on the basis of content and viewpoint. See, e.g., Deborah Epstein, Can a "Dumb Ass Woman" Achieve Equality in the Workplace? Running the Gauntlet of Hostile Environment Harassing Speech, 84 Geo. L.J. 399, 433 (1996); Eugene Volokh, How Harassment Law Restricts Free Speech, 47 Rutgers L. Rev. 563, 571-72 (1995); Suzanne Sangree, Title VII Prohibitions Against Hostile Environment Sexual Harassment and the First Amendment: No Collision in Sight, 47 Rutgers L. Rev. 461, 477 (1995); Richard H. Fallon, Sexual Harassment, Content Neutrality, and the First Amendment Dog That Didn’t Bark, 1994 Sup. Ct. L. Rev. 1, 8 (1994); Kingsley R. Browne, Title VII as Censorship: Hostile-Environment *207Harassment and the First Amendment, 52 Ohio State L.J. 481, 481 (1991); Marcy Strauss, Sexist Speech in the Workplace, 25 Harv. C.R.-C.L. L. Rev. 1, 32-33 (1990). But see Charles R. Calleros, Title VII and the First Amendment: Content-Neutral Regulation, Disparate Impact, and the “Reasonable Person", 58 Ohio St. L.J. 1217 (1997).
. Hishon v. King & Spalding, 467 U.S. 69, 104 S.Ct. 2229, 81 L.Ed.2d 59 (1984), which SCASD cites for the proposition that Title VII’s prohibitions do not offend the First Amendment, is inapposite. Hishon, which was decided years before the Supreme Court even recognized the existence of a "hostile environment” cause of action under that statute, only addressed the constitutionality of the statute’s application to quid pro quo harassment.
. The cases cited in R.A.V. each upheld a restriction on expressive conduct that was based solely on secondary effects of the speech that were merely incidental to its expressive content. In none of these cases, however, did the Court imply that the government may prohibit speech based on a desire to suppress the ideas it communicates. In Barnes, the Court found that the state’s legitimate interest in preventing public nudity permitted it to enforce a public indecency statute against a nude dancing establishment:
[W]e do not think that when Indiana applies its statute to the nude dancing in these nightclubs it is proscribing nudity because of the erotic message conveyed by the dancers.... The perceived evil that Indiana seeks to address is not erotic dancing, but public nudity. The appearance of people of all shapes, sizes and ages in the nude at a beach, for example, would convey little if any erotic message, yet the State still seeks to prevent it. Public nudity is the evil the State seeks to prevent, whether or not it is combined with expressive activity.
Barnes, 501 U.S. at 570, 111 S.Ct. 2456. Similarly, in Superior Court Trial Lawyers, the Court upheld, against First Amendment challenge, the application of the Sherman Act against boycotters based on the boycott's economic effects:
A nonviolent and totally voluntary boycott may have a disruptive effect on local economic conditions. This Court has recognized the strong governmental interest in certain forms of economic regulation, even though such regulation may have an incidental effect on rights of speech and association.
Superior Court Trial Lawyers, 493 U.S. at 428 n. 12, 110 S.Ct. 768 (quoting NAACP v. Claiborne Hardware Co., 458 U.S. 886, 912, 102 S.Ct. 3409, 73 L.Ed.2d 1215 (1982)). Finally, in O’Brien, the Court found no First Amendment impediment to prosecuting anti-war protestors who had violated federal law by burning their draft cards:
[E]ven on the assumption that the alleged, communicative element in O’Brien's conduct is sufficient to bring into play the First Amendment, it does not necessarily follow that the destruction of a registration certificate is constitutionally protected activity.... The many functions performed by Selective Service certificates establish beyond doubt that Congress has a legitimate and substantial interest in preventing their wanton and unrestrained destruction and assuring their continuing availability by punishing people who knowingly and wil-fully destroy or mutilate them.
O’Brien, 391 U.S. at 376, 380, 88 S.Ct. 1673. Accord Wisconsin v. Mitchell, 508 U.S. 476, 487-88, 113 S.Ct. 2194, 124 L.Ed.2d 436 (1993) (noting that conduct not targeted on the basis of its expressive content may be regulated under Title VII).
. We recognize that the SCASD Policy restricts the speech, not only of students, but also of teachers, volunteers and other adult members of the "school community.” Because we conclude, however, that the Policy fails under the less stringent standards for the restriction of student speech, we need not address this matter further.
. Saxe’s citation to Hynes v. Mayor & Council of Borough of Oradell, 425 U.S. 610, 96 S.Ct. 1755, 48 L.Ed.2d 243 (1976), ostensibly for the proposition that federal courts may not give a narrowing construction to a local statute, is inapposite. In Hynes, the New Jersey Supreme Court had already authoritatively construed the scope of the challenged statute. The U.S. Supreme Court held that the state court’s narrowing construction failed to solve the law’s vagueness problems and that, in light of the existing authoritative interpretation, the federal courts were without power to further limit the statute. See 425 U.S. at 622, 96 S.Ct. 1755. Here, in contrast, the SCASD Policy has not been authoritatively construed by the state courts, and we are therefore required to give it a reasonable narrowing construction if necessary to save it from unconstitutionality.
. Indeed, Saxe even suggests that the Policy could even be read to cover conduct occurring outside of school premises. This reading is not implausible based on the Policy's plain language, and would raise additional constitutional questions. See, e.g., Boucher v. School Board of the School District of Greenfield, 134 F.3d 821, 828 (7th Cir.1998) ("school officials' authority over off-campus expression is much more limited than it is over expression on school grounds”); Klein v. Smith, 635 F.Supp. 1440 (D.Me.1986) (student's vulgarity directed at teacher off school premises was "too attenuated to support discipline”).
. Such a reading would be consistent with the Policy's very broad statement of purpose, which notes that "[m] embers of the school community are expected to treat each other with mutual respect” and that "[d)isrespect among members of the school community is unacceptable behavior.”