B.L. v. Mahanoy Area School District

PRECEDENTIAL UNITED STATES COURT OF APPEALS FOR THE THIRD CIRCUIT __________ No. 19-1842 __________ B.L., a minor, by and through her father LAWRENCE LEVY and her mother BETTY LOU LEVY v. MAHANOY AREA SCHOOL DISTRICT, Appellant __________ On Appeal from the United States District Court for the Middle District of Pennsylvania (M.D. Pa. No. 3:17-cv-01734) Hon. A. Richard Caputo, United States District Judge __________ Argued November 12, 2019 Before: AMBRO, KRAUSE, and BIBAS, Circuit Judges (Filed: June 30, 2020) Arleigh P. Helfer, III Theresa E. Loscalzo Schnader Harrison Segal & Lewis 1600 Market Street Suite 3600 Philadelphia, PA 19103 Mary Catherine Roper American Civil Liberties Union of Pennsylvania P.O. Box 60173 Philadelphia, PA 19102 Sara J. Rose [Argued] American Civil Liberties Union P.O. Box 23058 Pittsburgh, PA 15222 Molly M. Tack-Hooper American Civil Liberties Union of Washington Foundation 901 Fifth Avenue Suite 630 Seattle, WA 19102 Counsel for Appellees David W. Brown Michael I. Levin [Argued] Levin Legal Group, P.C. 1800 Byberry Road 1301 Masons Mill Business Park Huntingdon Valley, PA 19006 John G. Dean Elliott Greenleaf & Dean 201 Penn Avenue Suite 202 Scranton, PA 18503 2 Counsel for Appellant Francisco M. Negrón, Jr. National School Boards Association 1680 Duke Street Room 523 Alexandria, VA 22314 Counsel for Amici Curiae National School Boards Association; Pennsylvania School Boards Association; Delaware School Boards Association; New Jersey School Boards Association; Pennsylvania Principals Association; National Association of Elementary School Principals; National Association of Secondary School Principals; and AASA, The School Superintendents Association Sophia Cope Electronic Frontier Foundation 815 Eddy Street San Francisco, CA 94109 Counsel for Amici Curiae Electronic Frontier Foundation, Student Press Law Center, Pennsylvania Center for the First Amendment, and Brechner Center for Freedom of Information Marieke T. Beck-Coon Foundation for Individual Rights in Education 510 Walnut Street Suite 1250 Philadelphia, PA 19106 Counsel for Amicus Curiae Foundation for Individual Rights in Education 3 __________ OPINION OF THE COURT __________ KRAUSE, Circuit Judge. Public school students’ free speech rights have long de- pended on a vital distinction: We “defer to the school[]” when its “arm of authority does not reach beyond the schoolhouse gate,” but when it reaches beyond that gate, it “must answer to the same constitutional commands that bind all other institu- tions of government.” Thomas v. Bd. of Educ., 607 F.2d 1043, 1044–45 (2d Cir. 1979). The digital revolution, however, has complicated that distinction. With new forms of communica- tion have come new frontiers of regulation, where educators assert the power to regulate online student speech made off school grounds, after school hours, and without school re- sources. This appeal takes us to one such frontier. Appellee B.L. failed to make her high school’s varsity cheerleading team and, over a weekend and away from school, posted a picture of her- self with the caption “fuck cheer” to Snapchat. J.A. 484. She was suspended from the junior varsity team for a year and sued her school in federal court. The District Court granted sum- mary judgment in B.L.’s favor, ruling that the school had vio- lated her First Amendment rights. We agree and therefore will affirm. I. BACKGROUND B.L. is a student at Mahanoy Area High School (MAHS). As a rising freshman, she tried out for cheerleading and made 4 junior varsity. The next year, she was again placed on JV. To add insult to injury, an incoming freshman made the varsity team. B.L. was frustrated: She had not advanced in cheerleading, was unhappy with her position on a private softball team, and was anxious about upcoming exams. So one Saturday, while hanging out with a friend at a local store, she decided to vent those frustrations. She took a photo of herself and her friend with their middle fingers raised and posted it to her Snapchat story.1 The snap was visible to about 250 “friends,” many of whom were MAHS students and some of whom were cheer- leaders, and it was accompanied by a puerile caption: “Fuck school fuck softball fuck cheer fuck everything.” J.A. 484. To that post, B.L. added a second: “Love how me and [another student] get told we need a year of jv before we make varsity but that’s [sic] doesn’t matter to anyone else? .”2 J.A. 485. One of B.L.’s teammates took a screenshot of her first snap and sent it to one of MAHS’s two cheerleading coaches. That coach brought the screenshot to the attention of her co-coach, 1 “Snapchat is a social media application for smartphones that allows users to send private text, photo, and video mes- sages to other users.” J.A. 6. Snaps can be viewed only tem- porarily and “cannot be accessed from the web.” Id. 2 The “upside-down smiley face” emoji “indicate[s] silliness, sarcasm, irony, passive aggression, or frustrated resignation.” Upside-Down Face Emoji, Dictionary.com, https://www.dic- tionary.com/e/emoji/upside-down-face-emoji (last visited June 25, 2020). 5 who, it turned out, was already in the know: “Several students, both cheerleaders and non-cheerleaders,” had approached her, “visibly upset,” to “express their concerns that [B.L.’s] [s]naps were inappropriate.” J.A. 7 (citations omitted). The coaches decided B.L.’s snap violated team and school rules, which B.L. had acknowledged before joining the team, requiring cheerleaders to “have respect for [their] school, coaches, . . . [and] other cheerleaders”; avoid “foul language and inappropriate gestures”; and refrain from sharing “negative information regarding cheerleading, cheerleaders, or coaches . . . on the internet.” J.A. 439. They also felt B.L.’s snap violated a school rule requiring student athletes to “con- duct[] themselves in such a way that the image of the Mahanoy School District would not be tarnished in any manner.” J.A. 486. So the coaches removed B.L. from the JV team. B.L. and her parents appealed that decision to the athletic director, school principal, district superintendent, and school board. But to no avail: Although school authorities agreed B.L. could try out for the team again the next year, they upheld the coaches’ decision for that year. Thus was born this lawsuit. B.L. sued the Mahanoy Area School District (School Dis- trict or District) in the United States District Court for the Mid- dle District of Pennsylvania. She advanced three claims under 42 U.S.C. § 1983: that her suspension from the team violated the First Amendment; that the school and team rules she was said to have broken are overbroad and viewpoint discrimina- tory; and that those rules are unconstitutionally vague. The District Court granted summary judgment in B.L.’s fa- vor. It first ruled that B.L. had not waived her speech rights by agreeing to the team’s rules and that her suspension from the 6 team implicated the First Amendment even though extracurric- ular participation is merely a privilege. Turning to the merits, the Court ruled that B.L.’s snap was off-campus speech and thus not subject to regulation under Bethel School District No. 403 v. Fraser, 478 U.S. 675 (1986). And, finding that B.L.’s snap had not caused any actual or foreseeable substan- tial disruption of the school environment, the Court ruled her snap was also not subject to discipline under Tinker v. Des Moines Independent Community School District, 393 U.S. 503 (1969). The Court therefore concluded that the School District had violated B.L.’s First Amendment rights, rendering unnec- essary any consideration of her overbreadth, viewpoint dis- crimination, or vagueness claims. It entered judgment in B.L.’s favor, awarding nominal damages and requiring the school to expunge her disciplinary record. This appeal fol- lowed. II. DISCUSSION3 The First Amendment provides that “Congress shall make no law . . . abridging the freedom of speech.” U.S. Const. amend. I. Over time, those deceptively simple words have spun off a complex doctrinal web. The briefs here are a testa- ment to that complexity, citing a wealth of cases involving not 3 The District Court had jurisdiction under 28 U.S.C. §§ 1331 and 1343(a), and we have jurisdiction under 28 U.S.C. § 1291. We review a grant of summary judgment de novo and apply the same standard as the district court. J.S. ex rel. Snyder v. Blue Mountain Sch. Dist., 650 F.3d 915, 925 (3d Cir. 2011) (en banc). 7 only student speech but also public employee speech, obscen- ity, indecency, and many other doctrines. At its heart, though, this appeal requires that we answer just two questions. The first is whether B.L.’s snap was protected speech. If it was not, our inquiry is at an end. But if it was, we must then decide whether B.L. validly waived that protection. Although navigating those questions requires some stopovers along the way, we ultimately conclude that B.L.’s snap was protected and that she did not waive her right to post it. A. B.L.’s Speech Was Entitled to First Amendment Protection We must first determine what, if any, protection the First Amendment affords B.L.’s snap. To do so, we begin by can- vassing the Supreme Court’s student speech cases. Next, we turn to a threshold question on which B.L.’s rights depend: whether her speech took place “on” or “off” campus. Finally, having found that B.L.’s snap was off-campus speech, we as- sess the School District’s arguments that it was entitled to pun- ish B.L. for that speech under Fraser, Tinker, and several other First Amendment doctrines. 1. Students’ broad free speech rights and the on- versus off-campus distinction For over three-quarters of a century, the Supreme Court has recognized that although schools perform “important, delicate, and highly discretionary functions,” there are “none that they may not perform within the limits of the Bill of Rights.” W. Va. State Bd. of Educ. v. Barnette, 319 U.S. 624, 637 (1943). And the free speech rights of minors are subject to “scrupulous protection,” lest we “strangle the free mind at its source and 8 teach youth to discount important principles of our government as mere platitudes.” Id. In Tinker v. Des Moines Independent Community School District, 393 U.S. 503 (1969), the Court reiterated that students do not “shed their constitutional rights to freedom of speech or expression at the schoolhouse gate.” Id. at 506. Expanding on Barnette, Tinker also held that student speech rights are “not confined to the supervised and ordained discussion” of the classroom; instead, they extend to all aspects of “the process of attending school,” whether “in the cafeteria, or on the play- ing field, or on the campus during authorized hours.” Id. at 512–13. Without “a specific showing of constitutionally valid reasons to regulate their speech,” then, “students are entitled to freedom of expression,” id. at 511, and cannot be punished for “expressions of feelings with which [school officials] do not wish to contend,” id. (quoting Burnside v. Byars, 363 F.2d 744, 749 (5th Cir. 1966)). To these broad rights, Tinker added a narrow exception “in light of the special characteristics of the school environment.” 393 U.S. at 506. Some forms of speech, the Court recognized, can “interfere[] . . . with the rights of other students to be secure and to be let alone.” Id. at 508. So as part of their obligation “to prescribe and control conduct in the schools,” id. at 507, school officials may regulate speech that “would ‘materially and substantially interfere with the requirements of appropriate discipline in the operation of the school,’” id. at 509 (quoting Burnside, 363 F.2d at 749). To exercise that regulatory power, however, schools must identify “more than a mere desire to avoid the discomfort and unpleasantness that always 9 accompany an unpopular viewpoint” and more than “undiffer- entiated fear or apprehension of disturbance.” Id. at 508–09. Tinker thus struck a balance, reaffirming students’ rights but recognizing a limited zone of heightened governmental au- thority. But that authority remains the exception, not the rule. Where Tinker applies, a school may prohibit student speech only by showing “a specific and significant fear of disruption,” J.S. ex rel. Snyder v. Blue Mountain Sch. Dist., 650 F.3d 915, 926 (3d Cir. 2011) (en banc) (quoting Saxe v. State Coll. Area Sch. Dist., 240 F.3d 200, 211 (3d Cir. 2001)), and where it does not, a school seeking to regulate student speech “must answer to the same constitutional commands that bind all other insti- tutions of government,” Thomas, 607 F.2d at 1045. In each of three later cases, the Court identified a limited area in which schools have leeway to regulate student speech without meeting Tinker’s substantial disruption standard. In Bethel School District No. 403 v. Fraser, 478 U.S. 675 (1986), it held that to “inculcate the habits and manners of civility,” schools may “prohibit the use of vulgar and offensive terms.” Id. at 681, 683 (citation omitted). In Hazelwood School Dis- trict v. Kuhlmeier, 484 U.S. 260 (1988), it held that officials may regulate student speech in the context of “school-spon- sored . . . expressive activities that students, parents, and mem- bers of the public might reasonably perceive to bear the impri- matur of the school,” provided “their actions are reasonably re- lated to legitimate pedagogical concerns.” Id. at 271–73. And in Morse v. Frederick, 551 U.S. 393 (2007), given educators’ “important—indeed, perhaps compelling[—]interest” in “de- terring drug use by schoolchildren,” id. at 407 (internal quota- tion marks and citation omitted), the Court held that schools 10 may “restrict student expression that they reasonably regard as promoting illegal drug use,” id. at 408. Although each of these cases added a wrinkle, none dis- turbed the basic framework on which Tinker relied. Fraser could not have been disciplined had he “delivered the same speech in a public forum outside the school context.” Morse, 551 U.S. at 405. Kuhlmeier’s editorial authority applies “only when a student’s school-sponsored speech could reasonably be viewed as speech of the school itself,” which “is not lightly to be presumed.” Saxe, 240 F.3d at 213–14. And central to Morse was not only the speech’s relationship to the school day—that it was made “during school hours” and “at a school- sanctioned activity,” 551 U.S. at 400–01 (citation omitted)— but also that juvenile drug use “cause[s] severe and permanent damage to the health and well-being of young people,” id. at 407. The Court’s case law therefore reveals that a student’s First Amendment rights are subject to narrow limitations when speaking in the “school context” but “are coextensive with [those] of an adult” outside that context. J.S., 650 F.3d at 932. 2. B.L.’s snap was “off-campus” speech To define B.L.’s speech rights with precision, therefore, we must ask whether her snap was “on-” or “off-campus” speech—terms we use with caution, for the schoolyard’s phys- ical boundaries are not necessarily coextensive with the “school context,” J.S., 650 F.3d at 932. After reviewing the line separating on- from off-campus speech, we hold B.L.’s speech falls on the off-campus side. 11 It is “well established” that the boundary demarcating schools’ heightened authority to regulate student speech “is not constructed solely of the bricks and mortar surrounding the school yard.” Layshock ex rel. Layshock v. Hermitage Sch. Dist., 650 F.3d 205, 216 (3d Cir. 2011) (en banc). That is the only conclusion to be drawn from the fact that the Supreme Court, in defining the scope of schools’ authority, has consist- ently focused not on physical boundaries but on the extent to which schools control or sponsor the forum or the speech. See Morse, 551 U.S. at 400–01; Kuhlmeier, 484 U.S. at 270–71; Fraser, 478 U.S. at 677, 680. And that focus makes sense: Just as the school context “is not confined to . . . the classroom,” Tinker, 393 U.S. at 512, neither can it be confined to the school’s physical grounds because exclusive dependence on “real property lines,” Layshock, 650 F.3d at 221 (Jordan, J., concurring), would exclude “part[s] of the process of attending school” that occur beyond those lines, Tinker, 393 U.S. at 512. Equally well established, however, is that “the ‘school yard’ is not without boundaries and the reach of school author- ities is not without limits.” Layshock, 650 F.3d at 216. School officials, in other words, may not “reach into a child’s home and control his/her actions there to the same extent that it can control that child when he/she participates in school sponsored activities.” Id. Permitting such expansive authority would twist Tinker’s limited accommodation of the “special charac- teristics of the school environment,” 393 U.S. at 506, into a broad rule reducing the free speech rights of all young people who happen to be enrolled in public school. The courts’ task, then, is to discern and enforce the line sep- arating “on-” from “off-campus” speech. That task has been 12 tricky from the beginning. See, e.g., Thomas, 607 F.2d at 1045–47, 1050–52 (declining to apply Tinker to a student pub- lication because, although a few articles were written and stored at school, the publication was largely “conceived, exe- cuted, and distributed outside the school”). But the difficulty has only increased after the digital revolution. Students use social media and other forms of online communication with remarkable frequency. Sometimes the conversation online is a high-minded one, with students “participating in issue- or cause-focused groups, encouraging other people to take action on issues they care about, and finding information on protests or rallies.” Br. of Amici Curiae Electronic Frontier Foundation et al. 13. Other times, that conversation is mundane or plain silly. Either way, the “omnipresence” of online communica- tion poses challenges for school administrators and courts alike. Layshock, 650 F.3d at 220–21 (Jordan, J., concurring); see J.S., 650 F.3d at 940 (Smith, J., concurring). Although the Supreme Court has not addressed the on- and off-campus divide in the context of online speech, it has laid down invaluable road markers that guide our way. The Court first addressed the internet’s “vast democratic forums” in Reno v. ACLU, 521 U.S. 844, 868 (1997). Reno recognized that the internet poses unique challenges but also offers unique advantages, “provid[ing] relatively unlimited, low-cost capac- ity for communication of all kinds” and content “as diverse as human thought,” id. at 870 (citation omitted). In applying the First Amendment to this technology, the Court was careful not to discard existing doctrines. Instead, it applied those doctrines faithfully, trusting that even faced with a “new marketplace,” “[t]he interest in encouraging freedom of expression in a dem- ocratic society outweighs any theoretical but unproven benefit 13 of censorship.” Id. at 885. It took a similar approach in Pack- ingham v. North Carolina, 137 S. Ct. 1730 (2017), recognizing both the “vast potential” and serious risks connected with the “revolution of historic proportions” wrought by new commu- nicative technologies. Id. at 1736. As in Reno, in Packingham the Court met new technologies with settled precedent, “exer- cis[ing] extreme caution before suggesting that the First Amendment provides scant protection for access to vast net- works” in “the modern public square.” Id. at 1736–37. The lesson from Reno and Packingham is that faced with new technologies, we must carefully adjust and apply—but not discard—our existing precedent. The thrust of that lesson is not unique to the First Amendment context. But it may be of special importance there because each new communicative technology provides an opportunity for “unprecedented” regu- lation. Packingham, 137 S. Ct. at 1737. And even when it is unclear whether the government will seize upon such an op- portunity, the lack of clarity itself has a harmful “chilling effect on free speech.” Reno, 521 U.S. at 872. Updating the line between on- and off-campus speech may be difficult in the so- cial media age, but it is a task we must undertake. Thankfully, significant groundwork has been laid. In 2011, we decided two appeals as a full Court, J.S. and Layshock, both of which involved a student’s fake MySpace profile ridiculing a school official using crude language. Although the profiles were created away from school, they were not far removed from the school environment: They attacked school officials, used photos copied from the schools’ websites, were shared with students, caused gossip at school and, in Layshock, were viewed on school computers. J.S., 650 F.3d at 920–23; 14 Layshock, 650 F.3d at 207–09. Even so, in both decisions we treated the profiles as “off-campus” speech. In J.S., we em- phasized that the speech occurred “outside the school, during non-school hours,” and deemed irrelevant that a printout of the profile had been brought into the school at the principal’s re- quest. 650 F.3d at 932–33. We went further in Layshock, re- jecting the arguments that the profile was “on-campus” speech because the profile was “aimed at the School District Commu- nity and . . . accessed on campus,” 650 F.3d at 216, and because the student had “enter[ed]” the school’s website to copy the principal’s photo, id. at 214–16. J.S. and Layshock yield the insight that a student’s online speech is not rendered “on campus” simply because it involves the school, mentions teachers or administrators, is shared with or accessible to students, or reaches the school environment. That was true in the analog era, see, e.g., Thomas, 607 F.2d at 1050–52; see also Porter v. Ascension Parish Sch. Bd., 393 F.3d 608, 611–12, 616–17 (5th Cir. 2004), and it remains true in the digital age. Applying these principles to B.L.’s case, we easily con- clude that her snap falls outside the school context. This is not a case in which the relevant speech took place in a “school- sponsored” forum, Fraser, 478 U.S. at 677, or in a context that “bear[s] the imprimatur of the school,” Kuhlmeier, 484 U.S. at 271. Nor is this a case in which the school owns or operates an online platform. Cf. Oral Arg. Tr. 25 (discussing a “school listserv”). Instead, B.L. created the snap away from campus, over the weekend, and without school resources, and she shared it on a social media platform unaffiliated with the school. And while the snap mentioned the school and reached 15 MAHS students and officials, J.S. and Layshock hold that those few points of contact are not enough. B.L.’s snap, therefore, took place “off campus.”4 3. The punishment of B.L.’s off-campus speech violated the First Amendment We next ask whether the First Amendment allowed the School District to punish B.L. for her off-campus speech. The District defends its decision under (i) Fraser, (ii) Tinker, and 4 Our concurring colleague asserts that it is “a fundamental principle of judicial restraint” that we must avoid analyzing constitutional issues beyond those implicated by “the precise facts” before us. Concurr. 1 (quoting Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 450 (2008) (dis- cussing the disfavored nature of facial challenges)). We take no issue with that general principle. Indeed, that principle ex- plains why, although we had to tease out the on- and off-cam- pus distinction enough to be confident about how to categorize B.L.’s speech, we have refrained from opining about how that distinction should be applied in future cases. We fail to see how our choice not to analyze hypothetical questions—for in- stance, the exact boundaries of “school-supervised channels” for “all forms of social media students use that schools moni- tor,” or which types of speech “constitute[] ‘harassment’ in the school and social media context,” id. at 3—shows a lack of ju- dicial restraint. Just as in all areas of constitutional law, future cases requiring additional analysis will supply the “facts” nec- essary “to draw . . . clear and administrable line[s].” Id. 16 (iii) a series of First Amendment doctrines beyond the student speech context. We address each in turn. i. B.L.’s punishment cannot be justified under Fraser The School District principally defends its actions based on its power “to enforce socially acceptable behavior” by banning “vulgar, lewd, obscene, or plainly offensive” speech by stu- dents. Appellant’s Br. 7–8. Under Fraser, such speech re- ceives “no First Amendment protection . . . in school.” Saxe, 240 F.3d at 213 (emphasis added). But the District’s argument runs aground on our precedent holding that Fraser does not apply to off-campus speech. J.S., 650 F.3d at 932–33; Layshock, 650 F.3d at 216–17, 219. As a panel, we may not revisit that precedent absent “intervening authority,” Reich v. D.M. Sabia Co., 90 F.3d 854, 858 (3d Cir. 1996), which neither party identifies here. See Morse, 551 U.S. at 405 (“Had Fraser delivered the same speech in a public forum outside the school context, it would have been protected.”). To prevail under Fraser, therefore, the School District must explain why J.S. and Layshock do not supply the decisional rule. Its attempts to do so come in several varieties but share the same thrust: that we should apply Fraser to off-campus speech where the speech or punishment involved an extracur- ricular activity. We are unpersuaded. To begin, the argument collides with our precedent. In Layshock, among several other punishments, the student was “banned from all extracurricular activities.” 650 F.3d at 210. But at no point did we suggest any relevant distinction among the punishments he had received. Quite the opposite: 17 Although we acknowledged the Second Circuit had suggested a lesser degree of First Amendment protection for punishments related to extracurricular activities, see Doninger ex rel. Don- inger v. Niehoff, 527 F.3d 41, 53 (2d Cir. 2008), in Layshock we declined to follow that analysis and even clarified that our discussion of Doninger was not a “suggest[ion] that we agree[d] with that court’s conclusion,” 650 F.3d at 218. All that mattered to us in Layshock was that the school had “pun- ish[ed]” the student for his speech, see id. at 214, 216, as B.L. was undoubtedly punished for hers.5 Even apart from Layshock’s guidance, we see no sound rea- son why we should graft an extracurricular distinction onto our case law. Yes, students have “a reduced expectation of pri- vacy” under the Fourth Amendment when they participate in extracurricular athletics. Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 657, 661–62, 665 (1995). But the School 5 The District Court assumed without deciding that B.L.’s claim fell within the First Amendment retaliation framework, which requires a plaintiff to show that “(1) he engaged in con- stitutionally protected conduct, (2) the defendant engaged in retaliatory action sufficient to deter a person of ordinary firm- ness from exercising his constitutional rights, and (3) a causal link [existed] between the constitutionally protected conduct and the retaliatory action.” Baloga v. Pittston Area Sch. Dist., 927 F.3d 742, 752 (3d Cir. 2019) (alteration in original) (inter- nal quotation marks and citation omitted). The parties here dis- pute only whether B.L.’s speech was constitutionally pro- tected. For the same reasons as the District Court, we conclude that we need not decide whether the retaliation framework is appropriate in this context. 18 District’s reliance on that line of cases is misplaced. In the Fourth Amendment context, “the ultimate measure of the con- stitutionality of a governmental search is ‘reasonableness,’” a standard which “is judged by balancing [the search’s] intrusion on the individual’s Fourth Amendment interests against its pro- motion of legitimate governmental interests.” Id. at 652–53 (citation omitted). The First Amendment, however, abhors “ad hoc balancing of relative social costs and benefits.” United States v. Stevens, 559 U.S. 460, 470 (2010); accord, e.g., Brown v. Entm’t Merchants Ass’n, 564 U.S. 786, 792 (2011). That line dividing First from Fourth Amendment doctrine is foundational, and we will not blur it here. The same goes for the argument that B.L. had no “constitu- tionally protected property right to participate in extracurricu- lar activities,” Appellant’s Br. 17. Be that as it may,6 due pro- cess case law—which also “depends upon a balancing of the individual rights and the governmental interests affected,” Main Rd. v. Aytch, 522 F.2d 1080, 1090 (3d Cir. 1975)—is an equally poor fit in the First Amendment context. To prevail on a free speech claim, a plaintiff need not show that his interests in speaking outweigh the government’s interests in suppress- ing the speech. Such a rule would “revise the ‘judgment [of] 6 We have suggested that students have no cognizable prop- erty interest in extracurricular activities, Angstadt v. Midd-W. Sch. Dist., 377 F.3d 338, 344 (3d Cir. 2004), a suggestion ech- oed by several other circuits, see, e.g., Lowery v. Euverard, 497 F.3d 584, 588 (6th Cir. 2007); Davenport ex rel. Daven- port v. Randolph Cty. Bd. of Educ., 730 F.2d 1395, 1397 (11th Cir. 1984). We take no position here on the wisdom or cor- rectness of that proposition. 19 the American people,’ embodied in the First Amendment, ‘that the benefits of its restrictions on the Government outweigh the costs.’”7 Entm’t Merchants, 564 U.S. at 792 (alteration in orig- inal) (quoting Stevens, 559 U.S. at 470). The School District next offers up an analogy: that students who join extracurriculars “represent their schools much in the way that government employees represent their employer.” Appellant’s Br. 30. So by going out for the team, it posits, students subject their speech rights to coaches’ whims so long as their speech does not involve “a matter of public concern.” Id. (citing Pickering v. Bd. of Educ., 391 U.S. 563, 568 (1968)). This argument, however, depends on dicta from the Sixth Cir- cuit, which went on to clarify that it was not “grafting a public- concern requirement onto” student speech doctrine and had in- voked the Pickering doctrine only to discuss whether “disrup- tion will occur when a subordinate challenges the authority of his or her superior.” See Lowery v. Euverard, 497 F.3d 584, 598 n.5 (6th Cir. 2007). And neither “the Supreme Court nor any other federal court of appeals has held [the personal mat- ter/public concern] distinction applicable in student speech cases.” Pinard v. Clatskanie Sch. Dist. 6J, 467 F.3d 755, 766 (9th Cir. 2006). The reason is simple: As we have recognized, 7 Similarly unavailing is the School District’s argument that Pennsylvania law permits regulation of students’ “conduct and deportment” only when they are “under the supervision of the board of school directors and teachers,” see 24 P.S. § 5-510, but authorizes regulation of extracurricular activities without that limitation, see id. § 5-511. Whether or not that is true is wholly beside the point, as state law cannot excuse a violation of the federal constitution. See U.S. Const. art. VI, cl. 2. 20 students’ free speech rights are not limited to matters of public concern. See, e.g., J.S., 650 F.3d at 926 (“Although Tinker dealt with political speech, the opinion has never been confined to such speech.”); see also Pinard, 467 F.3d at 766 (“[N]either Tinker nor its progeny limited students’ rights solely to the ex- ercise of political speech or speech that touches on a matter of public concern.”). Above all, we cannot depart from J.S. and Layshock with- out undermining the values those cases sought to protect. What was “unseemly and dangerous” about the efforts to apply Fra- ser to off-campus speech was not the punishments the students received, but that those punishments were used to “control” students’ free expression in an area traditionally beyond regu- lation. Layshock, 650 F.3d at 216. Those concerns apply with equal force where a school seeks to control student speech us- ing even modest measures, much less participation in extracur- ricular activities, which “are an important part of an overall ed- ucational program,” Br. of Amicus Curiae Foundation for In- dividual Rights in Education 7–8 (citation omitted). Thus, whatever the school’s preferred mode of discipline, it impli- cates the First Amendment so long as it comes in response to the student’s exercise of free speech rights. No one challenges that is exactly what happened to B.L. As a result, we can no more hold that B.L. abdicated her First Amendment right to speak as a cheerleader than we could re- turn to bygone days in which a police officer was thought to have a “right to talk politics . . . [but not] to be a policeman.” See O’Hare Truck Serv., Inc. v. City of Northlake, 518 U.S. 712, 716–17 (1996) (quoting McAuliffe v. Mayor, 29 N.E. 517, 517 (Mass. 1892)). Instead, we conclude, Fraser did not 21 authorize the School District’s punishment of B.L. for her off- campus speech. ii. Nor can B.L.’s punishment be justified under Tinker The School District falls back on Tinker, arguing that B.L.’s snap was likely to substantially disrupt the cheerleading program. But as we have explained, although B.L.’s snap in- volved the school and was accessible to MAHS students, it took place beyond the “school context,” J.S., 650 F.3d at 932. We therefore confront the question whether Tinker applies to off-campus speech. That is a question we have avoided answering to date. In Layshock, the school defended its decision to punish the stu- dent only under Fraser. See 650 F.3d at 216. And in J.S., we were able to “assume, without deciding,” that Tinker applied to speech like J.S.’s, 650 F.3d at 926, because we held that the school had not “reasonably forecast[] a substantial disruption of or material interference with the school,” id. at 931. But the question is once again squarely before us,8 and for three rea- sons we conclude we must answer it today. 8 One of the amici supporting B.L. suggests we follow J.S. by assuming Tinker applies and holding that her snap did not satisfy the substantial disruption standard. Br. of Amicus Cu- riae Foundation for Individual Rights in Education 17. An- other set of amici on B.L.’s side takes a different view, con- tending that Tinker’s substantial disruption standard “should not apply to off-campus speech.” Br. of Amici Curiae Elec- tronic Frontier Foundation et al. 4 (capitalization altered). For 22 First, our choice to sidestep the issue in J.S. adhered to the maxim that, where possible, we should avoid difficult consti- tutional questions in favor of simpler resolutions. There, it was sensible to avoid the issue because we could resolve the case by applying well-settled precedent addressing the substantial disruption standard in the context of the school environment. See, e.g., Sypniewski v. Warren Hills Reg’l Bd. of Educ., 307 F.3d 243, 254–57 (3d Cir. 2002); Saxe, 240 F.3d at 211– 12. But that is not the case here. The School District’s defense of its decision to punish B.L. focuses not on disruption of the school environment at large, but on disruption in the extracur- ricular context—specifically, the cheerleading program B.L. decried in her snap. And, as the parties’ and amici’s dueling citations reveal, the question of how to measure the potentially disruptive effect of student speech on particular extracurricular activities has bedeviled our sister circuits,9 and it is not one we her part, B.L. takes a middle path: She argues that “[f]un- damental First Amendment principles plainly forbid giving schools the power to censor student speech outside of school,” Appellee’s Br. 12, but as the appellee, she unsurprisingly adds we “need not answer that question in this case” because “even if it were clear that schools may punish offensive, off-campus speech under Tinker (which it is not),” the substantial disrup- tion standard was not met here, id. at 12, 22. And on the other side of the “v.,” both the School District and the amici that sup- port it argue that Tinker applies to off-campus speech like B.L.’s. See, e.g., Appellant’s Br. 23 & n.1, 41; Br. of Amici Curiae National School Boards Association et al. 18–23. 9 Compare, e.g., Pinard, 467 F.3d at 760–61, 768–69 (hold- ing that students’ distribution of a petition seeking their 23 have addressed to date. So were we to leapfrog Tinker’s ap- plicability in favor of substantial disruption analysis, we would still face complex and unresolved constitutional questions.10 coach’s resignation did not give rise to a reasonable forecast of substantial disruption, in part because the students had reported the coach’s misbehavior in a “responsibly tailored” way (cita- tion omitted)), with, e.g., id. at 769–70 (holding that the ath- letes’ refusal to board a bus before a game “substantially dis- rupted and materially interfered with a school activity”), Low- ery, 497 F.3d at 593–94 (holding that a similar petition request- ing a coach’s termination qualifies as substantially disruptive because of its effect on “team morale and unity”), and Wildman ex rel. Wildman v. Marshalltown Sch. Dist., 249 F.3d 768, 769–72 (8th Cir. 2001) (holding that a student athlete’s letter calling for teammates to criticize their coach disturbed the goal of providing “an educational environment conducive to learn- ing team unity and sportsmanship and free from disruptions and distractions that could hurt or stray the cohesiveness of the team”). 10 Our concurring colleague argues that this case is “straightforward” under Tinker’s substantial disruption stand- ard, Concurr. 3, because school authorities conceded there was “no reason to believe that the Snap would disrupt classroom or school activities,” e.g., Appellee’s Br. 8. But that is not the School District’s argument. Rather, the District contends that B.L.’s snap was disruptive because it undercut the “team mo- rale” and “chemistry” on which the cheerleading program de- pends and because, in the unique context of extracurricular ac- tivities, this is enough to satisfy Tinker. Appellant’s Br. 41. That contention finds some grounding in opinions from other 24 Second, when we decided J.S., the social media revolution was still in its infancy, and few appellate courts had grappled with Tinker’s application to off-campus online speech. In avoiding the issue, we afforded our sister circuits the chance to Courts of Appeals holding that because school athletics pro- grams rely heavily on “team unity,” “cohesiveness,” and “sportsmanship,” Wildman, 249 F.3d at 771, and advance a “narrower” set of goals than does the education system as a whole, Lowery, 497 F.3d at 589, student speech that under- mines those values satisfies Tinker’s substantial disruption standard. See Lowery, 497 F.3d at 593–94; Wildman, 249 F.3d at 769–72. But see Pinard, 467 F.3d at 768–69. Here, B.L. does not dispute that her speech would undermine team morale and chemistry: She openly criticized the program and ques- tioned her coaches’ decisionmaking, causing a number of teammates and fellow students to be “visibly upset” and to ap- proach the coaches with their “concerns,” J.A. 7 (citations omitted). She did so, moreover, in the context of a sport in which team members rely on each other for not only emotional and moral support, but also physical safety. In this context, we cannot so comfortably conclude that assuming Tinker’s ap- plicability and analyzing substantial disruption would yield a ready answer or a rule we could cogently explain for the benefit of future cases. And while our colleague makes some refer- ence to these issues in a footnote, we do not think they can be swept aside under the umbrella “that courts may consider all the ways in which student speech may be disruptive,” Con- curr. 4 n.1. At bottom, we think it unwise to explore these un- resolved questions without assessing the threshold question whether Tinker applies to B.L.’s speech in the first place. 25 coalesce around an approach and the Supreme Court the chance to resolve the issue. Nearly a decade later, however, we see not only that social media has continued its expansion into every corner of modern life, but also that no dominant ap- proach has developed. All the while, we have relegated district courts in this Circuit to confronting this issue without clear guidance, prompting them to turn elsewhere for support, see, e.g., Dunkley v. Bd. of Educ., 216 F. Supp. 3d 485, 492–94 (D.N.J. 2016), and to voice their growing frustration. As one of our district judges put it, “a district court in this Circuit takes up a student off-campus speech case for review with consider- able apprehension and anxiety.” R.L. ex rel. Lordan v. Cent. York Sch. Dist., 183 F. Supp. 3d 625, 635 (M.D. Pa. 2016). Finally, while legal uncertainty of any kind is undesirable, uncertainty in this context creates unique problems. Obscure lines between permissible and impermissible speech have an independent chilling effect on speech. See, e.g., Ashcroft v. Free Speech Coal., 535 U.S. 234, 244 (2002) (reasoning that the “uncertain reach” of a law punishing speech would “chill speech within the First Amendment’s vast and privileged sphere”). And because local officials are liable for constitu- tional violations only where “every reasonable official would understand that what he is doing is unlawful,” Russell v. Rich- ardson, 905 F.3d 239, 251 (3d Cir. 2018) (internal quotation marks and citation omitted), the unresolved issue of Tinker’s scope has left a significant obstacle in the path of any student seeking to vindicate her free speech rights through a § 1983 suit. See, e.g., Longoria ex rel. M.L. v. San Benito Indep. Con- sol. Sch. Dist., 942 F.3d 258, 267 (5th Cir. 2019) (holding that because the court had “declin[ed] to adopt a ‘specific rule,’” its case law applying Tinker to off-campus speech “does not 26 constitute clearly-established binding law that should have placed the defendants on notice about the constitutionality of their actions”). The time has come for us to answer the question. We begin by canvassing the decisions of our sister circuits. We then con- sider the wisdom of their various approaches, tested against Tinker’s precepts. Finally, we adopt and explain our own, con- cluding that Tinker does not apply to off-campus speech and reserving for another day the First Amendment implications of off-campus student speech that threatens violence or harasses others. a. Other courts’ approaches Our sister circuits have approached this issue in three ways. One group applies Tinker where it was reasonably foreseeable that a student’s off-campus speech would reach the school en- vironment. That test sprung from trying circumstances: In Wisniewski ex rel. Wisniewski v. Board of Education, 494 F.3d 34 (2d Cir. 2007), a student created an instant messaging icon showing “a pistol firing a bullet at a person’s head, above which were dots representing splattered blood,” and beneath which were the words “Kill Mr. VanderMolen,” the student’s teacher. Id. at 35–36. That icon was visible to the student’s “buddies,” and he sent messages displaying it to fellow stu- dents. Id. at 36. In upholding his suspension, the Second Cir- cuit held that it was appropriate to apply Tinker because “it was reasonably foreseeable that the IM icon would come to the at- tention of school authorities,” id. at 39, and that the violence- threatening speech satisfied Tinker’s substantial disruption standard, id. at 38–39. The Eighth Circuit, in another case in- volving a threat of violence, took the same approach. See 27 D.J.M. ex rel. D.M. v. Hannibal Pub. Sch. Dist. No. 60, 647 F.3d 754, 757–59, 765–67 (8th Cir. 2011) (emphasizing that “student creativity and . . . ability . . . can[not] flourish if vio- lence threatens the school environment”). But from those cases involving threats of violence, the “rea- sonable foreseeability” standard spread far and wide. Multiple circuits have applied it in cases involving sexual or racial har- assment. See C.R. ex rel. Rainville v. Eugene Sch. Dist. 4J, 835 F.3d 1142, 1146, 1151 (9th Cir. 2016); S.J.W. ex rel. Wil- son v. Lee’s Summit R-7 Sch. Dist., 696 F.3d 771, 773, 777–78 (8th Cir. 2012). And the Second Circuit has applied it in a case involving neither violence nor harassment: In Doninger, the court used it to assess the punishment of a student who urged others to contact a school official to protest a concert’s post- ponement. 527 F.3d at 44–45, 48–52. The Eighth Circuit has likewise suggested that the standard governs all forms of off- campus speech, not just violent threats and harassment. S.J.W., 696 F.3d at 777. Another group of circuits applies Tinker to off-campus speech with a sufficient “nexus” to the school’s “pedagogical interests.” Kowalski v. Berkeley Cty. Schs., 652 F.3d 565, 573 (4th Cir. 2011). Kowalski involved a student who created a MySpace page harassing a fellow student. Id. at 567–68. In assessing the student’s suspension, the Fourth Circuit empha- sized that student-on-student harassment “can cause victims to become depressed and anxious, to be afraid to go to school, and to have thoughts of suicide.” Id. at 572 (citation omitted). Concluding that schools “must be able to prevent and punish harassment and bullying in order to provide a safe school en- vironment,” id., the court held that the speech bore a “sufficient 28 nexus with the school” justifying Tinker’s application, id. at 577. The Ninth Circuit has also applied the nexus test in a case involving off-campus sexual harassment. C.R., 835 F.3d at 1150–51. Finally, some circuits have applied Tinker to off-campus speech without articulating a governing test or standard. See, e.g., Bell v. Itawamba Cty. Sch. Bd., 799 F.3d 379, 394 (5th Cir. 2015) (en banc) (declining to “adopt a specific rule” but applying Tinker to a student who “intentionally direct[ed] at the school community [a] rap recording containing threats to, and harassment and intimidation of, two teachers”); Wynar v. Douglas Cty. Sch. Dist., 728 F.3d 1062, 1069 (9th Cir. 2013) (declining to “divine and impose a global standard for . . . off- campus speech” but holding that Tinker reaches off-campus speech presenting “an identifiable threat of school violence”). b. Issues with these approaches We sympathize with our sister circuits, which have faced the unenviable task of assessing students’ free speech rights against the backdrop of “school officials’ need to provide a safe school environment,” LaVine v. Blaine Sch. Dist., 257 F.3d 981, 983 (9th Cir. 2001), and find much to commend in their thoughtful opinions. Ultimately, however, we find their ap- proaches unsatisfying in three respects. First, “bad facts make bad law,” United States v. Joseph, 730 F.3d 336, 337 (3d Cir. 2013), and one unmistakable trend from the case law is that the most challenging fact patterns have produced rules untethered from the contexts in which they arose. The Second Circuit provides a case in point. It is un- derstandable that the court in Wisniewski, focusing on the 29 threat of violence bound up in the student’s speech, upheld the school’s authority to discipline him. See 494 F.3d at 39–40. As other courts have recognized, “we live in a time when school violence is an unfortunate reality that educators must confront on an all too frequent basis,” LaVine, 257 F.3d at 987, and in doing so, they “must be vigilant” and “react to potential threats before violence erupts,” Bell, 799 F.3d at 393. But in Doninger, the Second Circuit reflexively applied Wisniewski’s reasonable foreseeability test to a fact pattern of a very differ- ent sort: a student’s protest of a school’s decision to postpone an event. What began as a narrow accommodation of unusu- ally strong interests on the school’s side, cf. Wynar, 728 F.3d at 1069 (distinguishing “an identifiable threat of school vio- lence” from “myriad” other fact patterns), became a broad rule governing all off-campus expression. A similar dynamic took place with the “nexus” test, in that specialized concerns related to “harassment and bullying in the school environment,” Kow- alski, 652 F.3d at 572, produced a rule making off-campus free speech rights depend on the speech’s connection to a school’s “pedagogical interests,” id. at 573. Second, and as a result of this expansionary dynamic, our sister circuits have adopted tests that sweep far too much speech into the realm of schools’ authority. Start with reason- able foreseeability. Technology has brought unprecedented in- terconnectivity and access to diverse forms of speech. In the past, it was merely a possibility, and often a remote one, that the speech of a student who expressed herself in the public square would “reach” the school. But today, when a student speaks in the “modern public square” of the internet, Packing- ham, 137 S. Ct. at 1737, it is highly possible that her speech will be viewed by fellow students and accessible from school. 30 And in some situations, it is a virtual certainty: Depending on the settings favored by that student’s “friends” or “followers,” her message may automatically pop up on the face of class- mates’ phones in the form of notifications from Instagram, Fa- cebook, Twitter, Snapchat, or any number of other social plat- forms. Implicit in the reasonable foreseeability test, therefore, is the assumption that the internet and social media have ex- panded Tinker’s schoolhouse gate to encompass the public square. That assumption is not one we can accept, though, be- cause it subverts the longstanding principle that heightened au- thority over student speech is the exception rather than the rule. And it contradicts the Supreme Court’s instruction, in cases like Packingham and Reno, to apply legal precedent faithfully even when confronted with new technologies.11 11 By way of example, imagine a student who, off campus and over the weekend, writes a blog post identifying every teacher he thinks is incompetent. Imagine that he then shares the post on a social media platform where it is visible to many fellow students. It is a near certainty that the post will “reach campus,” Doninger, 527 F.3d at 48: Students are likely to chat about it in the lunchroom, view it surreptitiously in class, or even share it with school officials. But that type of downstream “reach[ing]” the “campus,” id., is “different in kind” from a student’s choice to “stand[] up during a lecture” and share sim- ilar thoughts about the teacher’s incompetence. See Lee Gold- man, Student Speech and the First Amendment: A Comprehen- sive Approach, 63 Fla. L. Rev. 395, 407 n.92 (2011) (citation omitted). If it is to remain a limited carveout from students’ general “free speech rights,” see J.S., 650 F.3d at 932, Tinker must apply only to the latter. See also id. at 939 (Smith, J., 31 The nexus test suffers from similar overbreadth. In holding that schools have regulatory authority over any speech, whether on or off campus, that “interfere[s] with the work and discipline of the school,” Kowalski, 652 F.3d at 574, it col- lapses Tinker’s scope of application and rule into one analytical step. The result is tautological: Schools can regulate off-cam- pus speech under Tinker when the speech would satisfy Tinker. And the effect is to erase the dividing line between speech in “the school context” and beyond it, J.S., 650 F.3d at 927, a line which is vital to young people’s free speech rights. Worse, in extending Tinker wherever there is a “nexus” to “pedagogical interests,” Kowalski, 652 F.3d at 573, the test raises the specter of officials’ asserting the power to regulate “any student speech that interferes with [the] school’s educational mission,” a power that “can easily be manipulated in dangerous ways.” J.S., 650 F.3d at 927 (internal quotation marks omitted) (quot- ing Morse, 551 U.S. at 423 (Alito, J., concurring)). Such an expansion of schools’ regulatory power would have “ominous implications” indeed. J.S., 650 F.3d at 939–40 (Smith, J., con- curring) (exploring the consequences not only for students, but also for adults, of extending Tinker to off-campus speech). Third, other circuits’ approaches have failed to provide clarity and predictability. This is true for those that have “de- clined to adopt a rule,” e.g., Bell, 799 F.3d at 394, leaving “stu- dents, teachers, and school administrators” without “clear guidance,” Longoria, 942 F.3d at 265 (citation omitted). But it is also true for those that have crafted a rule. In layering a concurring) (discussing the hypothetical of a student who writes an off-campus blog post taking a position that causes fellow students to react on campus). 32 foreseeability requirement on top of Tinker, the Second and Eighth Circuits have made it difficult for students speaking off campus to predict when they enjoy full or limited free speech rights. After all, a student can control how and where she speaks but exercises little to no control over how her speech may “come to the attention of the school authorities,” D.J.M., 647 F.3d at 766 (quoting Wisniewski, 494 F.3d at 39). The nexus test, too, affords little clarity, leaving students to wonder what types of speech might implicate a school’s “pedagogical interests,” Kowalski, 652 F.3d at 573. And in the First Amend- ment context, courts must pursue ex ante clarity not for clar- ity’s own sake, but to avoid chilling potential speech and to give government officials notice of the constitutional bounda- ries they may not cross. In the end, although the courts to address this issue have done so thoughtfully, we conclude that their approaches sweep in too much speech and distort Tinker’s narrow exception into a vast font of regulatory authority. We must forge our own path. c. Our approach We hold today that Tinker does not apply to off-campus speech—that is, speech that is outside school-owned, -oper- ated, or -supervised channels and that is not reasonably inter- preted as bearing the school’s imprimatur. In so holding, we build on a solid foundation, for in his concurrence in J.S., now- Chief Judge Smith, joined by four colleagues, embraced this rule, explaining “that the First Amendment protects students engaging in off-campus speech to the same extent it protects speech by citizens in the community at large.” 650 F.3d at 936. That rule is true to the spirit of Tinker, respects students’ rights, 33 and provides much-needed clarity to students and officials alike. From the outset, Tinker has been a narrow accommodation: Student speech within the school context that would “materi- ally and substantially interfere[] with the requirements of ap- propriate discipline,” Tinker, 393 U.S. at 505 (citation omit- ted), is stripped of the constitutional shield it enjoys “outside [that] context,” Morse, 551 U.S. at 405. Tinker’s focus on dis- ruption makes sense when a student stands in the school con- text, amid the “captive audience” of his peers. Fraser, 478 U.S. at 684. But it makes little sense where the student stands outside that context, given that any effect on the school envi- ronment will depend on others’ choices and reactions. Recent technological changes reinforce, not weaken, this conclusion. Like all who have approached these issues, we are “mindful of the challenges school administrators face,” includ- ing the need to manage the school environment in the digital age. Layshock, 650 F.3d at 222 (Jordan, J., concurring). We are equally mindful, however, that new communicative tech- nologies open new territories where regulators might seek to suppress speech they consider inappropriate, uncouth, or pro- vocative. And we cannot permit such efforts, no matter how well intentioned, without sacrificing precious freedoms that the First Amendment protects. The consensus in the analog era was that controversial off-campus speech was not subject to school regulation, see, e.g., Porter, 393 F.3d at 611–12, 615– 16; Thomas, 607 F.2d at 1050–52, and Reno and Packingham 34 require that we adhere to that principle even as the speech moves online.12 Holding Tinker inapplicable to off-campus speech also of- fers the distinct advantage of offering up-front clarity to stu- dents and school officials. To enjoy the free speech rights to which they are entitled, students must be able to determine when they are subject to schools’ authority and when not. A test based on the likelihood that speech will reach the school environment—even leaving aside doubts about what it means to “reach” the “school environment”—fails to provide that clarity. The same is true for a test dependent on whether the student’s speech has a sufficient “nexus” to unspecified peda- gogical interests or would substantially disrupt the school 12 Several circuits have applied Tinker to speech that the speaker brought into the campus environment. See, e.g., Boim v. Fulton Cty. Sch. Dist., 494 F.3d 978, 980–85 (11th Cir. 2007) (upholding a suspension of a student who, in class, showed another student a violent story she had written at home); Boucher v. Sch. Bd., 134 F.3d 821, 822, 827–29 (7th Cir. 1998) (applying Tinker to a student newspaper written out- side school but distributed “in bathrooms, in lockers and in the cafeteria”). Our holding tracks those cases because they do not involve “off-campus” speech at all. A student who brings a printed story into campus and shows it to fellow students has expressed herself inside the school context regardless whether she wrote the story at home or in class. So too with a student who opens his cellphone and shows a classmate a Facebook post from the night before. 35 environment.13 But a test based on whether the speech occurs in a context owned, controlled, or sponsored by the school is much more easily applied and understood. That clarity bene- fits students, who can better understand their rights, but it also benefits school administrators, who can better understand the limits of their authority and channel their regulatory energies in productive but lawful ways. Nothing in this opinion questions school officials’ “com- prehensive authority” to regulate students when they act or speak within the school environment. J.S., 650 F.3d at 925 (quoting Tinker, 393 U.S. at 507). Tinker applies, as it always has, to any student who, on campus, shares or reacts to contro- versial off-campus speech in a disruptive manner. That author- ity is not insignificant, and it goes a long way toward address- ing the concern, voiced by the School District and our concur- ring colleague, that holding Tinker is limited to on-campus speech will “sow . . . confusion” about what to do when a stu- dent’s controversial off-campus speech “provoke[s] significant disruptions within the school,” Concurr. 6. The answer is straightforward: The school can punish any disruptive speech or expressive conduct within the school context that meets Tinker’s standards—no matter how that disruption was “pro- voke[d].” It is the off-campus statement itself that is not sub- ject to Tinker’s narrow recognition of school authority. But at least in the physical world, that is nothing new, and no one, including our colleague, has second-guessed that longstanding 13 Our divided precedent shows it is often not easy to pre- dict whether speech will satisfy Tinker’s substantial disruption standard. Compare, e.g., J.S., 650 F.3d at 928–31, with, e.g., id. at 943–50 (Fisher, J., dissenting). 36 principle or suggested that a student who advocated a contro- versial position on a placard in a public park one Saturday would be subject to school discipline. We simply hold today that the “online” nature of that off-campus speech makes no constitutional difference. See supra pages 11–16. Nor are we confronted here with off-campus student speech threatening violence or harassing particular students or teach- ers. A future case in the line of Wisniewski, D.J.M., Kowalski, or S.J.W., involving speech that is reasonably understood as a threat of violence or harassment targeted at specific students or teachers, would no doubt raise different concerns and require consideration of other lines of First Amendment law. Cf. Layshock, 650 F.3d at 209–10, 219 (holding that the student’s parody MySpace page was protected speech even though the school had deemed it “[h]arassment of a school administra- tor”); J.S., 650 F.3d at 922, 933 (holding the same even though the school’s principal had contacted the police to press harass- ment charges). And while we disagree with the Tinker-based theoretical approach that many of our sister circuits have taken in cases involving students who threaten violence or harass others, our opinion takes no position on schools’ bottom-line power to discipline speech in that category. After all, student speech falling into one of the well-recognized exceptions to the First Amendment is not protected, cf. Doe v. Pulaski Cty. Spe- cial Sch. Dist., 306 F.3d 616, 619, 621–27 (8th Cir. 2002) (en banc) (upholding a school’s punishment of a student who wrote a threatening letter under the “true threat” doctrine); speech outside those exceptions may be regulated if the government can satisfy the appropriate level of scrutiny, see, e.g., Williams- Yulee v. Fla. Bar, 135 S. Ct. 1656, 1665–72 (2015); cf. Oral Arg. Tr. 28 (exploring whether actions taken to prevent 37 student-on-student harassment could satisfy strict scrutiny); and, perhaps most relevant, the Supreme Court has recognized that a sufficiently weighty interest on the part of educators can justify a narrow exception to students’ broader speech rights, see Morse, 551 U.S. at 407–08. We hold only that off-campus speech not implicating that class of interests lies beyond the school’s regulatory authority. True, our rule leaves some vulgar, crude, or offensive speech beyond the power of schools to regulate. Yet we return to Tinker and find in its pages wisdom and comfort: [O]ur Constitution says we must take this risk, and our history says that it is this sort of hazardous freedom— this kind of openness—that is the basis of our national strength and of the independence and vigor of Ameri- cans who grow up and live in this relatively permissive, often disputatious, society. 393 U.S. at 508–09 (internal citation omitted); see Barnette, 319 U.S. at 641 (encouraging courts to “apply the limitations of the Constitution with no fear that freedom to be intellectu- ally and spiritually diverse or even contrary will disintegrate the social organization”). Tinker’s careful delineation of schools’ authority, like these principles, is no less vital even in today’s digital age to ensure “adequate breathing room for valuable, robust speech.” J.S., 650 F.3d at 941 (Smith, J., concurring). For these reasons, we hold that Tinker does not apply to off-campus speech and thus cannot justify the decision to punish B.L. 38 iii. None of the School District’s remaining arguments justifies its punishment of B.L. Moving beyond student speech,14 the School District ad- vances a few arguments for why B.L.’s snap enjoyed no First Amendment protection at all. Each is unsuccessful. First, the School District contends that “vulgar language [i]s ‘low-value speech’ that c[an] be restricted ‘to a greater ex- tent than would otherwise be permissible.’” Appellant’s Br. 35 (quoting C.H. ex rel. Z.H. v. Oliva, 226 F.3d 198, 211 (3d Cir. 2000) (Alito, J., dissenting)). But in doing, the District relies on a dissenting opinion, and in any event its selective quotation omits the prepositional phrase “[i]n the public schools” and our citation of Fraser, see C.H., 226 F.3d at 211 (Alito, J., dissent- ing), both of which make clear we were not making a broad statement that non-obscene profanity enjoys reduced First Amendment protection. Had we made such a statement, it would have defied decades of settled law. See, e.g., Cohen v. California, 403 U.S. 15, 20 (1971). Second, the School District argues B.L.’s snap was unpro- tected because it “expressed no opinion.” Appellant’s Br. 34– 35. In support, it quotes B.L., who, when asked whether she was “trying to send a message,” replied she “was just mad 14 The School District does not suggest it had a right to reg- ulate B.L.’s snap under Kuhlmeier or Morse. Nor could it: No reasonable listener could have concluded that B.L.’s snap amounted to “speech of the school itself,” Saxe, 240 F.3d at 213–14, or speech “promoting illegal drug use,” Morse, 551 U.S. at 403. 39 about everything.” Id. at 34 (quoting J.A. 65). This argument borders on the frivolous. The “particular four-letter word” B.L. used “is perhaps more distasteful than most others of its genre,” but “one man’s vulgarity is another’s lyric,” Cohen, 403 U.S. at 25, and here, B.L. used it to vent her frustrations with the cheerleading program. There is no doubt B.L.’s snap was “imbued with elements of communication,” Troster v. Pa. State Dep’t of Corr., 65 F.3d 1086, 1090 (3d Cir. 1995) (cita- tion omitted), and thus deserving of First Amendment protec- tion. Finally, the School District argues that “profane speech is not protected when aimed at minors.” Appellant’s Reply 2 (capitalization altered). Again, the District misses the mark. Its argument relies on FCC v. Pacifica Foundation, 438 U.S. 726 (1978), a case involving the sui generis context of radio broadcasting, which is “uniquely accessible to children,” id. at 749. But nowhere did Pacifica suggest that indecent speech falls outside the First Amendment. Moreover, B.L.’s snap was no more indecent, or targeted at an “intended audience [of] mi- nors,” Appellant’s Reply 3, than the MySpace profiles we held were entitled to First Amendment protection in J.S. and Layshock. For these reasons, we hold that B.L.’s snap was not subject to regulation under Tinker or Fraser and instead enjoyed the full scope of First Amendment protections. B. B.L. Did Not Waive Her Free Speech Rights The School District next argues that by agreeing to certain school and team rules, B.L. waived her First Amendment right to post the “fuck cheer” snap. We disagree. 40 To begin, we note that the District Court ruled that requir- ing B.L. to waive her First Amendment rights as a condition of joining the team violated the unconstitutional conditions doc- trine, see Koontz v. St. Johns River Water Mgmt. Dist., 570 U.S. 595, 604–06 (2013), and that both B.L. and an amicus urge us to affirm that ruling. No doubt, for the government to condition participation in a beneficial program on a waiver of First Amendment rights raises serious constitutional concerns, particularly where the government “seek[s] to leverage [bene- fits] to regulate speech outside the contours of the program it- self.” Agency for Int’l Dev. v. All. for Open Soc’y Int’l, Inc., 570 U.S. 205, 214–15 (2013); see also, e.g., FCC v. League of Women Voters of Cal., 468 U.S. 364, 399–401 (1984). At the same time, however, the line between constitutional and un- constitutional conditions “is hardly clear,” Agency for Int’l Dev., 570 U.S. at 215, and there are a wide range of extracur- ricular activities and student roles that may make conditions on speech more or less connected to the needs of the program. Fortunately, we need not decide on which side of the line this case falls because we conclude that B.L. did not waive her right to the speech at issue here. All rights, including free speech rights, can be waived. Curtis Publ’g Co. v. Butts, 388 U.S. 130, 142–43 (1967). But waivers “must be voluntary, knowing, . . . intelligent, . . . [and] established by ‘clear’ and ‘compelling’ evidence,” Erie Tele- comms., Inc. v. City of Erie, 853 F.2d 1084, 1094 (3d Cir. 1988) (citation omitted), and courts must “indulge in every reasona- ble presumption against waiver,” id. at 1095 (quoting John- son v. Zerbst, 304 U.S. 458, 464 (1938)). Applying those standards, we conclude that B.L.’s snap does not clearly “fall within the scope,” United States v. Wilson, 707 F.3d 412, 414 41 (3d Cir. 2013) (citation omitted), of any of the rules on which the School District relies. We begin with the “Respect Rule” governing MAHS cheer- leaders: Please have respect for your school, coaches, teachers, other cheerleaders and teams. Remember, you are rep- resenting your school when at games, fundraisers, and other events. Good sportsmanship will be enforced[;] this includes foul language and inappropriate gestures. J.A. 439. B.L.’s snap contained foul language and disrespected her school and team. But the rule’s language suggests it applies only “at games, fundraisers, and other events,” a suggestion echoed by its invocation of “[g]ood sportsmanship.” Id. That would not cover a weekend post to Snapchat unconnected with any game or school event and before the cheerleading season had even begun. And common sense supports this reading: It is hard to believe a reasonable student would understand that by agreeing to the Respect Rule, she was waiving all rights to malign the school once safely off campus and in the world at large. Indeed, one of the cheerleading coaches recognized that the rule “doesn’t say anything about not being able to use foul language or inappropriate gestures . . . away from school.” J.A. 90. So this rule is of no help to the School District. The “Negative Information Rule” is likewise inapplicable. It states “[t]here will be no toleration of any negative infor- mation regarding cheerleading, cheerleaders, or coaches placed on the internet.” J.A. 439. Unlike the Respect Rule, this rule by its terms reaches off-campus speech. But it reaches only “information,” id., a term denoting matters of fact, see, e.g., Information, Merriam Webster’s Collegiate Dictionary 42 (10th ed. 1997) (“the communication or reception of knowledge or intelligence”; “knowledge obtained from inves- tigation, study, or instruction”), not mere expressions of opin- ion or emotion. We are hard pressed to find in the words “fuck cheer” any discernable negative information about the cheer- leading program. And although B.L.’s second snap contains information about the varsity team’s acceptance of an incom- ing freshman, nothing in the record suggests B.L.’s punishment was based on that snap or the information it revealed. So this rule, too, provides no basis for a finding of waiver. The School District’s last recourse is the “Personal Conduct Rule” in MAHS’s student handbook. It provides: Participation on an athletic team or cheerleading squad in the Mahanoy Area School District is a privilege and the participants must earn the right to represent Maha- noy Schools by conducting themselves in such a way that the image of the Mahanoy School District would not be tarnished in any manner. Any participant whose conduct is judged to reflect a discredit upon him- self/herself, the team, or the Mahanoy Schools, whether or not such activity takes place during or outside school hours during the sports season, will be subject to disci- plinary action as determined by the coach, the athletic director and/or the school principal. J.A. 486. This rule does not lend itself to a finding of waiver for two reasons. First, it applies only “during the sports sea- son,” id., but B.L. posted her snap after the previous season had ended and before practices for the next season had begun. Sec- ond, the rule’s language gives few clear markers, applying wherever a student’s behavior would “tarnish[]” the school’s “image” in “any manner,” J.A. 486. That language is too 43 obscure, and too dependent on the whims of school officials, to give rise to a knowing and voluntary waiver of B.L.’s rights to speak as she did. We therefore hold that B.L.’s snap was not covered by any of the rules on which the School District relies and reject its contention that B.L. waived her First Amendment rights. * * * The heart of the School District’s arguments is that it has a duty to “inculcate the habits and manners of civility” in its stu- dents. Appellant’s Br. 24 (citation omitted). To be sure, B.L.’s snap was crude, rude, and juvenile, just as we might expect of an adolescent. But the primary responsibility for teaching ci- vility rests with parents and other members of the community. As arms of the state, public schools have an interest in teaching civility by example, persuasion, and encouragement, but they may not leverage the coercive power with which they have been entrusted to do so. Otherwise, we give school adminis- trators the power to quash student expression deemed crude or offensive—which far too easily metastasizes into the power to censor valuable speech and legitimate criticism. Instead, by enforcing the Constitution’s limits and upholding free speech rights, we teach a deeper and more enduring version of respect for civility and the “hazardous freedom” that is our national treasure and “the basis of our national strength.” Tinker, 393 U.S. at 508–09. III. CONCLUSION For the foregoing reasons, we will affirm the judgment of the District Court. 44 AMBRO, Circuit Judge, concurring in the judgment I concur in the judgment affirming the District Court’s grant of summary judgment to B.L. on the narrow ground that our holdings in Layshock ex rel. Layshock v. Hermitage School District, 650 F.3d 205 (3d Cir. 2011) (en banc), and J.S. ex rel. Snyder v. Blue Mountain School District, 650 F.3d 915 (3d Cir. 2011) (en banc), mandate that outcome. I dissent from the majority’s holding that, on the facts before us, the holding in Tinker v. Des Moines Independent Community School District, 393 U.S. 503 (1969)—that schools may regulate student speech only if it “substantially disrupt[s] the work and discipline of the school,” id. at 513—does not apply to “off-campus” speech. I dissent because it is a fundamental principle of judicial restraint that courts should “neither anticipate a question of constitutional law in advance of the necessity of deciding it nor formulate a rule of constitutional law broader than is required by the precise facts to which it is to be applied.” Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 450 (2008) (citing Ashwander v. Tenn. Valley Auth., 297 U.S. 288, 346–47 (1936)) (quotation marks omitted). Cf. Golden v. Zwickler, 394 U.S. 103, 108 (1969) (“For adjudication of constitutional issues[,] concrete legal issues[] presented in actual cases, not abstractions[,] are requisite.”) (citation and quotation marks omitted). In Tinker the Supreme Court held that public school students do not shed their freedom of speech at the “schoolhouse gate,” 393 U.S. at 506, and their expression may not be suppressed unless, to repeat, school officials reasonably conclude that it will “materially and substantially disrupt the work and discipline of the school,” id. at 513. Our Court in two en banc rulings expressly declined to hold that Tinker does not apply to off-campus speech and applied 1 Tinker’s reasoning to those cases. See Layshock, 650 F.3d at 219 (“We need not now define the precise parameters of when the arm of authority can reach beyond the schoolhouse gate because . . . the district court found that [the student’s] conduct did not disrupt the school.”); id. at 220 (Jordan, J., concurring) (stating that the majority did not decide whether Tinker applies off campus and arguing that it does); J.S., 650 F.3d at 928–31, 933 (assuming Tinker governs and applying it; “[n]either the Supreme Court nor this Court has ever allowed schools to punish students for off-campus speech that is not school-sponsored or at a school-sponsored event and that caused no substantial disruption at school”). In both en banc cases we held in favor of students who had been suspended from school, and disciplined in other ways, for creating websites, while not on school property and not using school computers, mocking in appalling terms school officials. We concluded that the schools could not “punish a student for expressive conduct that originated outside of the schoolhouse, did not disturb the school environment and was not related to any school sponsored event.” Layshock, 650 F.3d at 207. B.L. concedes we need not decide whether Tinker’s test applies off campus. See, e.g., Appellee’s Br. 22 (“It is an open question whether public schools can ever punish students’ out-of-school speech—even if the Tinker standard is satisfied. . . . The Court need not answer that question in this case.”). Nonetheless, my colleagues in the majority hold that “Tinker does not apply to off-campus speech—that is, speech that is outside school-owned, -operated, or -supervised channels and that is not reasonably interpreted as bearing the school’s imprimatur[,]” Maj. Op. 33, and leave open the door for schools to regulate off-campus student speech if it threatens violence or harasses particular students or teachers, id. at 37. However, the case before us does not involve “school-supervised channels,” nor does it concern speech that 2 carried the school imprimatur, or was violent or threatening. So it comes as no surprise that the majority does not give guidance on how its new rule is to be applied. How do we define school-supervised channels? Do these channels include all forms of social media students use that schools monitor? What type of speech constitutes “harassment” in the school and social media context? Indeed there are no facts before us to draw a clear and administrable line for this new rule that Tinker does not apply to off-campus speech. The case before us is straightforward—B.L.’s Snap is not close to the line of student speech that schools may regulate. B.L. was suspended from her school’s cheerleading team as punishment for a Snap that said “fuck cheer,” which she created on her own smartphone, on her own time on a weekend, while off-campus, and not participating in any school-sponsored activity. The Snap did not mention the School District, the school, or any individuals, and did not feature any team uniforms, school logos, or school property. It caused complaints by a few other cheerleaders but no “substantial disruptions,” and the coaches testified that they did not expect the Snap would substantially disrupt any activities in the future.1 1 My colleagues cite Lowery v. Euverard, 497 F.3d 584 (6th Cir. 2007), and Wildman ex rel. Wildman v. Marshalltown School District, 249 F.3d 768 (8th Cir. 2001), among other cases, to argue that B.L.’s case is actually a nuanced one because it involves student athletics. However, both Lowery, 497 F.3d at 593–94, and Wildman, 249 F.3d at 771, expressly applied Tinker’s “substantial disruption” test and considered the effect of the students’ speech on team morale in deciding whether it caused a disruption. In my view, there is nothing controversial about the notion that 3 We have already rejected the School District’s principal argument, specifically that Bethel School District Number 403 v. Fraser, 478 U.S. 675 (1986), allows schools to punish students for their offensive or profane speech when the speech takes place off campus, outside of school activities, and without the use of school resources. J.S., 650 F.3d at 920, 923, 925, 932–33 & n.12; Layshock, 650 F.3d at 209, 219. And none of the other narrow exceptions to Tinker apply. B.L.’s Snap did not bear the imprimatur of the school in the way a school-sponsored newspaper does, see Hazelwood Sch. Dist. v. Kuhlmeier, 484 U.S. 260, 262 (1988), and she did not send her Snap from a school- supervised or -sanctioned event nor to anyone at such an event, see Morse v. Frederick, 551 U.S. 393, 396–97 (2007). My colleagues correctly point out that the School District’s remaining arguments also are unavailing. That students have a reduced expectation of privacy under the Fourth Amendment when they participate in extracurricular athletics, see Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 657, 661–62, 665 (1995), has no bearing on our First Amendment jurisprudence. We have never and decline now to “graft an extracurricular distinction onto our [First Amendment] case law.” Maj. Op. 18. I agree. Nor am I aware of any other circuit court that has adopted such a distinction. courts may consider all the ways in which student speech may be disruptive, including its effect on student activities such as sports and sportsmanship. That is indeed what the District Court did here; it considered all the alleged disruptive effects of B.L.’s speech and concluded that under Tinker, J.S., and Layshock, B.L.’s speech was not disruptive. I agree with the District Court and would affirm on the same ground. 4 Thus Tinker and its progeny, and our en banc decisions in Layshock and J.S., dictate that the School District violated B.L.’s First Amendment rights. That is all we had to say. Instead, ours is the first Circuit Court to hold that Tinker categorically does not apply to off-campus speech. A few Circuits have flirted with such a holding and have declined to apply Tinker to off-campus speech on a case-by- case basis. See, e.g., Porter v. Ascension Par. Sch. Bd., 393 F.3d 608, 615, 619–20 (5th Cir. 2004) (declining to apply Tinker where student at home drew a picture of school being attacked, and that picture inadvertently ended up on campus, because it was off-campus speech not directed at the school and the student took no step to bring the speech on campus); Thomas v. Bd. of Educ., 607 F.2d 1043, 1051 (2d Cir. 1979) (holding that school violated students’ speech rights by suspending them for publishing an underground lewd newspaper that was printed and distributed off campus, even if an occasional article was composed on campus, because the newspaper was “off-campus expression”). However, those same Circuit Courts have subsequently applied Tinker to off- campus speech. See, e.g., Wisniewski v. Bd. of Educ. of Weedsport Cent. Sch. Dist., 494 F.3d 34, 39–40 (2d Cir. 2007) (applying Tinker to uphold punishment of student who sent instant messages to fellow students from home computer during non-school hours depicting teacher being shot because the student’s hostile off-campus speech posed a reasonably foreseeable threat of disruption in school); Bell v. Itawamba Cty. Sch. Bd., 799 F.3d 379, 396 (5th Cir. 2015) (en banc), cert. denied, 136 S. Ct. 1166 (2016) (declining to “adopt any rigid standard,” but applying Tinker to a student who posted off site a song recording that threatened and harassed two teachers); see also Doninger v. Niehoff, 527 F.3d 41, 50–53 (2d Cir. 2008) (applying Tinker to uphold punishment of student whose blog demeaned school administrators for cancelling a school concert, and clarifying 5 that Thomas v. Board of Education did not stand for the proposition that off-campus speech may never be punished). The bottom line is that Circuit Courts facing harder and closer calls have stayed their hand and declined to rule categorically that Tinker does not apply to off-campus speech. Yet we do so here in a case bereft of substantial disruptions within the school. I fear that our decision will sow further confusion. For example, how does our holding apply to off- campus racially tinged student speech? Can a school discipline a student who posts off-campus Snaps reenacting and mocking the victims of police violence where those Snaps are not related to school, not taken or posted on campus, do not overtly threaten violence and do not target any specific individual, yet provoke significant disruptions within the school? Hard to tell. We promulgate a new constitutional rule based on facts that do not require us to entertain hard questions such as these. The craft of judging has a restraining principle: Do not decide today what can be decided tomorrow, for tomorrow it may not need to be decided. We twist that tenet today by a wide-reaching holding for facts outside the question my colleagues call. In J.S., despite a well-reasoned concurrence urging that Tinker not apply to off-campus student speech, J.S., 650 F.3d at 936–41 (Smith, J.), our en banc decisions in both it and Layshock declined to go that far. Yet a panel does so today with no more compelling context than either en banc case. Our task is to balance tolerance for expressive conduct with the need for order in our schools. The test in Tinker— whether student speech reasonably “forecast[s] substantial disruption of or material interference with school activities,” 393 U.S. at 514—is the law we applied en banc, and it no doubt works here to rule in B.L.’s favor. Why go further until it is needed? 6 Hence, while I join the judgment in today’s case, I dissent from its holding. 7