Note from the Editor:
This article discusses the Supreme Court’s opinion in Minnesota Voters Alliance v. Mansky and criticizes the Court’s dicta suggesting that it would uphold well written bans on political apparel in polling places.
The Federalist Society takes no positions on particular legal and public policy matters. Any expressions of opinion are those of the author. Whenever we publish an article that advocates for a particular position, as here, we offer links to other perspectives on the issue, including ones opposed to the position taken in the article. We also invite responses from our readers. To join the debate, please email us at firstname.lastname@example.org.
Oftentimes what we are allowed to say depends on where we say it. When we speak on government-controlled property, our free-speech rights are limited by the government’s need to maintain the property for its intended use, for example, as a school, a military base, or a polling place on Election Day. In Minnesota Voters Alliance v. Mansky, the Supreme Court considered a law that forbade the wearing of “political badge[s], political button[s], or other political insignia[s]” “at or about the polling place.” The law was challenged by a group of voters who wanted to enter their polling places wearing shirts and buttons that contained political messages.
The Supreme Court, in a 7-2 decision written by Chief Justice John Roberts, concluded that the law was unconstitutional. Neither the law itself nor any authoritative construction provided sufficient guidance to enable officials consistently to interpret the scope of the ban on “political” apparel. Accordingly, the law was subject to arbitrary applications—a cardinal sin for laws that limit speech.
This outcome was unsurprising. The real story, however, is not in the law that the Court struck down, but in the laws that the Court hinted that it would uphold. Although the Court struck down the Minnesota law, the Court signaled that other bans on political apparel in polling places could very well be constitutional so long as those other laws avoid the vagueness that rendered Minnesota’s political-apparel ban invalid. Such dicta are unfortunate, for they would allow the government to limit voters’ political speech without pointing to any harm likely to result from the speech. Longstanding precedent, however, indicates that the government may not limit speech—even in nonpublic forums, such as polling places—unless the ban is a “reasonable” way of ensuring that the government-controlled property will be able to be used for its non-speech purpose.
The government’s power to limit speech in polling places should therefore depend on how reasonable the limit is in preserving polling places’ character as places to cast votes. States should be able to limit speech to guard against voter intimidation and even to preserve a calm, reflective environment in which voters can make their selections. Intimidation and chaos are, after all, inconsistent with reasoned reflection on candidates and public issues. On the other hand, political speech that does not threaten those values should be protected. In particular, wearing clothing with a campaign slogan or insignia is fully consistent with the purposes of a polling place. Political-apparel bans should, therefore, be unconstitutional except as applied to conduct that presents a reasonable risk of voter intimidation or disorder.
I. Interiors of Polling Places Are Nonpublic Forums
Political speech is the core concern of the First Amendment. Accordingly, regulations of political speech usually trigger strict scrutiny, requiring the government to justify such regulations by showing that they are narrowly tailored to a compelling government interest. It is indisputable that a regulation like Minnesota’s would be unconstitutional if it were applied to limit political speech on privately owned land or in a public forum. Nonpublic forums, however, are treated differently. In that category of government-owned property—places not set aside for speech or typically used to engage in debate—strict scrutiny does not apply to government restrictions on speech. Rather, speech restrictions in nonpublic forums are constitutional if they are “reasonable in light of the purpose served by the forum and are viewpoint neutral.”
The first step in determining the constitutionality of Minnesota’s political-apparel ban, then, was to determine whether the inside of a polling place should be considered a public forum.
Never before had the Court considered the constitutionality of a law restricting speech within polling places. The closest precedent was Burson v. Freeman, which involved a Tennessee law that prohibited the display or distribution of campaign material and the solicitation of votes in the area around polling places. The Burson plurality applied strict scrutiny because the ban encompassed streets and sidewalks within 100 feet of polling places (an area the plurality considered to be “quintessential public forums”), but upheld the law. The plurality held that the deterrence of “voter intimidation and election fraud” was a compelling interest, and that the 100-foot “campaign-free zone” was a permissible means of achieving that interest.
If the Tennessee law was a permissible restriction of campaigning outside polling places, as Burson held, a fortiori a similar restriction on campaigning should be permissible inside polling places, where the government’s interests would be even more compelling. The Minnesota law challenged in Minnesota Voters Alliance, however, banned more speech than the Tennessee law did—including speech that was extremely unlikely to produce either voter intimidation or election fraud. Whereas the Tennessee law was principally concerned with limiting the distribution of campaign material and solicitation of votes—activities that could be hard for unwilling targets to avoid, and which might involve physical approaches and therefore a significant prospect of intimidation—the Minnesota law prohibited the merely “passive” conduct of wearing clothing with a political message or logo. It is very hard to imagine that the wearing of such clothing would lead to voter intimidation or election fraud, and even harder to imagine that apparel bans would be narrowly tailored ways of avoiding those problems. Accordingly, if political-apparel bans were evaluated under strict scrutiny, they would likely fail.
In Minnesota Voters Alliance, however, the Court held that while the traditional public forums of streets and sidewalks are often used for speech, the interiors of polling places are not: “A polling place . . . is, at least on Election Day, government-controlled property set aside for the sole purpose of voting.” Accordingly, the inside of a polling place was held to be a nonpublic forum, and the political-apparel ban triggered not strict scrutiny, but the much more lenient test applicable to nonpublic forums: whether the ban was reasonable in light of the purpose of the forum (voting) and free of viewpoint discrimination. And because the law was viewpoint neutral (at least on its face), the key question was reasonableness.
II. Vague Bans on “Political” Speech Are Unreasonable
The Supreme Court struck down Minnesota’s political-apparel ban on the narrow ground that the government had not adequately defined the kinds of apparel that were subject to the prohibition. The ban was therefore too “indeterminate” and likely to lead to “erratic application.”
Minnesota’s political-apparel ban provided that a “political badge, political button, or other political insignia may not be worn at or about the polling place.” Election judges at each polling place were authorized to determine if a particular piece of clothing fell within the prohibition, and to refer the matter to other officials if the offending voter refused to remove the item. Offenders were allowed to vote, but they were subject to a fine of up to $300.
The challengers who brought the case to the Supreme Court were individuals and associations of individuals who wished to wear clothing promoting the Tea Party and buttons stating “Please I.D. Me.” When they attempted to vote, some of the challengers were asked to cover up their political apparel, and those who refused to do so had their names recorded for referral and possible prosecution. One of the challengers was twice barred from voting until he removed or covered up the political apparel, despite the fact that the Minnesota law did not permit election officials to turn away voters who persisted in wearing political apparel.
In an attempt to clarify the meaning of “political”—and to make clear that the statute did not ban all speech having to do with government, politics, or elections—the Minnesota secretary of state issued a guidance document to the state’s election officials, stating that the prohibited political apparel “includes, but [is] not limited to” any item containing “the name of a political party” or “the name of a candidate at any election”; “[a]ny item in support of or opposition to a ballot question at any election”; “[i]ssue-oriented material designed to influence or impact voting (including specifically the ‘Please I.D. Me’ buttons)”; and “[m]aterial promoting a group with recognizable political views (such as the Tea Party, MoveOn.org, and so on).” Overall, Minnesota argued that the ban did not apply to all material that might be considered “political,” but rather to material having to do with the choices faced by the voters at the election.
The Court was troubled by the vagueness of the ban, even as clarified by the secretary of state. The government wished to apply the ban on issue-oriented material to issues that had been raised in the election campaigns. But, as the Court pointed out, “[a] rule whose fair enforcement requires an election judge to maintain a mental index of the platforms and positions of every candidate and party on the ballot is not reasonable.” The issue-related ban also led to some seeming inconsistencies. For example, at oral argument, the state represented that a voter could not wear a shirt that contained the text of the Second Amendment, but could wear one that contained the text of the First Amendment.
The ban on materials associated with groups holding “recognizable political views” was hopelessly vague. As the Court noted, “[a]ny number of associations, educational institutions, businesses, and religious organizations could have an opinion on an ‘issue confronting voters in a given election.’” The state’s attempt to limit the ban to groups holding views that were “well-known” to the “typical observer” raised more problems, for the application of those standards “may turn in significant part on the background knowledge and media consumption of the particular election judge.” For these reasons, the Court concluded that the statute was unconstitutionally vague.
Even the two dissenters did not disagree that the ban’s vagueness created constitutional concerns. Justice Sotomayor, joined by Justice Breyer, urged the Court to certify the case to the Minnesota Supreme Court so that the state court could construe the ban to avoid the law’s apparent vagueness and arbitrary distinctions. In the dissenters’ view, the state court’s construction “likely would obviate the hypothetical line-drawing problems” identified by the majority of the Supreme Court.
III. Would a Well Written Ban on Political Apparel Be Reasonable?
There were two ways that the Minnesota political-apparel ban could have been held to fail the test of reasonableness applicable to speech restrictions in nonpublic forums. The first way, adopted by the Court, was to hold that the law was unreasonable because it was vague. Potential speakers could not determine what messages would be determined to be “political,” and the vagueness of the law meant that the officials charged with its enforcement had too much discretion—which could have been used to discriminate against disfavored viewpoints. The alternative approach would have addressed a more fundamental objection to the ban: It simply prohibited too much speech—far more than necessary to serve the government interests at issue.
Minnesota Voters Alliance suggested in dicta that a “more lucid” statute than Minnesota’s would have been constitutional, and it pointed to statutes in California and Texas as examples of the kinds of restrictions that states may constitutionally adopt. Those laws ban political apparel in polling places, but only political apparel that references a candidate, a ballot measure, or (in Texas) a political party. While the California and Texas bans are more clearly written than the Minnesota one, they still ban political speech. Thus, according to the public-forum doctrine discussed above, they would have to be reasonable and viewpoint neutral to be constitutional. Although the Court seemed inclined to view such statutes as constitutional, it should have been more cautious about saying so. Such bans are likely unreasonable, and therefore unconstitutional, because they restrict far more speech than necessary to serve the government’s interests in conducting elections and permitting voters to cast votes in an appropriately contemplative atmosphere.
Assessing reasonableness first requires one to identify with particularity the government interests a speech restriction is supposed to serve. Burson accepted that the government has a compelling interest in avoiding voter intimidation and electoral fraud. But there is no reason to think banning political apparel would have any effect on electoral fraud. And as to voter intimidation, the voter who is intimidated by another voter’s T-shirt is unusually susceptible to influence; the vast majority of voters would be no more intimidated by another voter’s clothing than by a neighbor’s yard sign. Viewed from the speaker’s perspective, on the other hand, political-apparel bans impose a burden on voting. If a voter wants to display his political apparel outside the polling place, he may have to bring a change of clothes for when he goes inside. If he forgets extra clothes, he may have to go home, return to the polling place, and wait in line again, before voting. If the polls have closed in the meantime, the ban could cost the voter the opportunity to cast a ballot. The political-apparel ban thereby imposes a non-trivial burden on speech and voting rights and barely, if at all, serves the compelling interest in preventing voter intimidation. The ban is therefore an unreasonable way of achieving those interests.
In nonpublic forums, however, the government has authority to pursue interests beyond those that are compelling. Indeed, in a nonpublic forum, the government may impose speech restrictions that serve no other interest than preserving the forum for its non-speech purpose. Simply stated, the government may restrict speech inside a polling place if the restriction is a reasonable way of preserving the polling place as a location for voting. While such a test is deferential to the government (and appropriately so), not all bans on speech in nonpublic forums are reasonable. Sometimes the restrictions sweep too broadly, and the First Amendment protects speakers whose expressive conduct presents little risk of interfering with the purpose of the forum—in this case, enabling voters to cast votes free of pressure or conflict.
Surely states may ban polling-place speech that actually makes it difficult to cast or record votes, that threatens other voters, or that “is intended to mislead voters about voting requirements and procedures.” Just as surely, the government may not ban certain speech simply because it would rather designate an area as off-limits to that kind of speech. Thus, it is crucial to specify what it means to preserve polling places for voting. Minnesota Voters Alliance phrased the government interest variously as “set[ting a polling place] aside as ‘an island of calm in which voters can peacefully contemplate their choices’”; “reflect[ing] th[e] distinction” between “choosing [and] campaigning”; “ensur[ing] that partisan discord not follow the voter up to the voting booth, and distract from a sense of shared civic obligation at the moment it counts the most”; and allowing voters to “focus on the important decisions immediately at hand.”
States may justifiably work to promote voters’ “peaceful contemplat[ion]” and “focus.” Any activity that disrupts an individual’s ability to cast a vote for a chosen candidate, or that prolongs the voting process by interfering with voters’ ability to concentrate, interferes with the voting process itself. It is by no means clear, however, that states may pass speech restrictions with the purpose of suppressing “partisan discord” or “campaigning” unless there is some other reason that discord or campaigning is harmful, such as interfering with voters’ free choices or, perhaps, “distract[ing] from a sense of shared civic obligation.” Discord itself, the Court has long recognized, is a natural by-product of the First Amendment, which expresses a “profound national commitment to the principle that debate on public issues should be uninhibited, robust, and wide-open, and that it may well include vehement, caustic, and sometimes unpleasantly sharp attacks on government and public officials.” Partisan discord, in other words, advances self-government unless there is something about the way the speech is expressed that interferes with the voting process. Some speech within the polling place surely can produce such interference, but not all speech will, and it is the interference with voting that might be caused by the speech, rather than the discord itself, which is the harm that states should try to prevent. States have no interest in avoiding public disagreement per se; such an interest seems to be nothing less than aversion to the clash of views that is an inherent (not to say beneficial) part of free debate protected by the First Amendment. As the Court eloquently said in Texas v. Johnson in reversing a conviction for flag-burning, “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.” Thus, states should not be able to create campaign-free zones if the only reason for doing so is that the government would prefer that campaign speech not occur in that place, or that some people encountering the speech will disagree with it; rather, the government should have to point to some reason that speech in that location would cause harm. It is particularly doubtful that a state could have a legitimate interest in suppressing “partisan discord,” given that (as the Court recently reaffirmed in Reed v. Town of Gilbert, Arizona) discrimination between different content-based categories of speech triggers strict scrutiny and political speech is the kind of speech most central to the purpose of the First Amendment. Thus, speech restrictions in nonpublic forums may be reasonable if they promote voters’ free choices and calm reflection, but not if they exclude political speech from polling places despite the lack of any threat to those interests.
Measured by this standard, most political-apparel bans are unconstitutional. States should be able to ban active political speech in polling places—for example, approaching or addressing other voters to persuade them to support a candidate or ballot measure. Burson recognized states’ interest in protecting voters from a barrage of campaigning immediately outside of the polling place, and that interest is even stronger inside the building. Within the polling place, states should be able to prohibit all loud communication, and perhaps all oral communication unconnected with voting, so as to preserve the peace and quiet that facilitate reflection by voters. Voters in the polling place are a captive audience, and states should be able to protect them from unwanted noise. States might even be able to discriminate by content and prohibit oral communication about campaigns for candidates and ballot measures, on the ground that such communication could present the greatest interference with other voters’ ability to decide on those very candidates and issues. Passive displays of clothing, however, would not interfere with other voters’ ability to reflect on their choices and make their selections. A voter concentrating on his ballot may be disrupted by a voice; he will not be disrupted by another voter’s T-shirt. One can turn away from a visual display and limit the distraction. Sounds, however, are far more invasive. So long as one is within earshot of a distracting sound, that sound can force its way into one’s consciousness and interfere with the ability to concentrate on other tasks. Once one looks away from a visual display, however, it is distracting only as a memory. The memory of seeing another voter’s T-shirt is thus little different from the memory of seeing a piece of political apparel—or a yard sign, or a billboard—outside the polling place, whereas distracting sounds present a continuing bombardment of our consciousness whether we direct our attention to them or not.
In the analogous context of government workplaces, the Court has protected expression so long as it does not obstruct government functions. And the Court has not simply rubber-stamped the government’s claims that speech would lead to obstruction of its functions. In Pickering v. Board of Education and Rankin v. McPherson, for example, the Court held that the government’s interests in the effective operation of schools and law-enforcement agencies did not require it to restrict employees’ speech. Pickering upheld a teacher’s right to publish a letter critical of his school board, and McPherson protected the right of an employee in a constable’s office to make a remark supporting the assassination of President Reagan. In both cases, the Court demanded that the government show that the speech would harm the functioning of the government office, and the Court analyzed the facts of each case before concluding that no such harm was likely. By analogy, the Court should not blindly defer to states’ claims that all political apparel presents a risk of interference with other voters’ ability to vote.
Schools are another nonpublic forum in which speech restrictions are often challenged, and they may be the most analogous context to polling places. In both contexts, the government must carry out a function (education/running an election) other than providing a forum for speech. In both schools and the polls, an excessive amount of speech (e.g., raucous chanting) could interfere with that government function. Neither schools nor polling places exist as forums for speech, yet both education and voting depend on the exchange of ideas that is protected by the First Amendment. Further, in both contexts, bans on political apparel would substantially limit speakers’ ability to reach their intended audiences (classmates/other voters). If schoolchildren were to be allowed to display political messages only outside of the school, and voters were allowed to display political messages only outside of polling places, each group of speakers would have less of an opportunity to communicate their messages to their intended audiences. Such a limitation should be permissible only where the communication of the messages has a realistic chance of producing harm.
Minnesota Voters Alliance seemed content to resolve this conflict by saying that the time for speech was before one arrived at the polls: “Casting a vote . . . is a time for choosing, not campaigning.” Yet the Court adopted a markedly less restrictive approach with respect to schools. In Tinker v. Des Moines Independent Community School District, the Court held that the school could not prohibit students from wearing arm bands that protested the Vietnam War, and it rejected Justice Hugo Black’s argument that school officials should be able to exclude political speech from classrooms so that students could “keep their minds on their own schoolwork.” Although the Court recognized that speech could be disruptive to schools’ ability to educate students, the Court was adamant that the mere possibility of such disruption was insufficient to justify a limitation on speech, even within schools: “[U]ndifferentiated fear or apprehension of disturbance is not enough to overcome the right to freedom of expression.” The Tinker Court demanded record facts to support the school’s contention that the arm bands would disrupt the functioning of the school, and it found no such support. The protesting students “neither interrupted school activities nor sought to intrude in the school affairs or the lives of others. They caused discussion outside of the classrooms, but no interference with work and no disorder.” The parallel with polling-place speech is apparent: Wearing “political” apparel—be it ideological, partisan, or candidate-specific—need not cause any interference with voting or disorder.
Minnesota Voters Alliance recognized the potential parallel with Tinker and further noted that in another case the Court had characterized the wearing of political apparel in an airport as “‘nondisruptive.’” The Court attempted to dismiss the relevance of these precedents by asserting that neither case involved the purportedly “unique context of a polling place on Election Day.” The Court did not, however, explain what was so different about polling places that a speech restriction could be justified there, but would be unconstitutional in schools. The two contexts seem remarkably similar; in both instances the government’s important function could suffer from noisy or distracting political speech, but not from mere displays of political ideology or affiliation on apparel or accessories. If schoolchildren can be trusted to ignore their classmates’ political apparel and concentrate on their lessons, we should be able to trust adults to ignore others’ political apparel and concentrate on voting for the few minutes that they are in the voting booth. In fact, speech by adult voters should be even more protected than speech by schoolchildren, as school authorities are permitted much greater control of students’ lives than election officials are able to exercise over voters. Further, children are more impressionable and distractible than are adults, so speech in schools is more likely to interfere with schools’ educational function than political apparel is to interfere with the ability to vote. Just as “[n]either students [n]or teachers shed their constitutional rights to freedom of speech or expression at the schoolhouse gate,” voters should not be deemed to shed their constitutional rights to freedom of speech or expression at the entrance to the polling place.
Minnesota’s political-apparel ban not only restricted political speech, but it was unclear about what speech would be treated as “political.” The state’s interpretations, meant to clarify the scope of the ban, created arbitrary distinctions and drew lines that were nearly impossible to administer. Accordingly, the Supreme Court easily concluded that the ban was unconstitutional without a clearer definition of the kind of “political” apparel that was barred from the polling place.
More significant for the future are the Court’s dicta suggesting that if states clearly identify the apparel that is prohibited, they may impose polling-place bans on clothing containing political messages or logos. These dicta go too far in approving restrictions of political speech. States need to ensure that polling places provide an environment that allows voters to think and concentrate on the choices they are making, and to make those choices without undue influence or intimidation. Wearing a “Make America Great Again” hat or a “Yes We Can” shirt, however, does nothing to undermine the purpose of a polling place because it does not intimidate voters or interfere with voters’ ability to contemplate the questions on the ballot.
The Court was correct to hold that polling places are nonpublic forums. Accordingly, the government may restrict speech within polling places so long as those speech restrictions are reasonable and viewpoint neutral. “Reasonable,” however, does not mean that the government has carte blanche to restrict expression whenever it would prefer not to see it. Rather, there must be some reason to think that the expression would interfere with the ability of people to cast their votes. The Court should recognize that one voter’s clothing does not interfere with another voter’s rights. Our democracy accepts—and in some ways depends on—our differences of opinion about politics. Far from being a threat to democracy, our political differences make democracy meaningful. So long as those differences are expressed in a way that does not intimidate others, we should be proud to live in a country that celebrates our ability to vote and to voice our opinions.
 138 S. Ct. 1876 (2018).
 Minn. Stat. § 211B.11(1). Despite the potential vagueness of the term “at or about,” the parties in Minnesota Voters Alliance agreed “that the political apparel ban applies only within the polling place.” 138 S. Ct. at 1883 (emphasis in original).
 The wearing of clothing is not literally “speech.” Nevertheless, it fits comfortably within the kind of “expressive conduct” that the Court has held to be protected by the First Amendment, such as displaying (or burning) a flag. See Texas v. Johnson, 491 U.S. 397, 402-06 (1989) (holding that burning the American flag was speech); Spence v. Washington, 418 U.S. 405, 409-11 (1974) (per curiam) (holding that the display of an upside-down American flag with a duct-taped peace sign was speech); Stromberg v. California, 283 U.S. 359 (1931) (holding that a law banning the display of a red flag violated the First and Fourteenth Amendments).
 See, e.g., Board of Airport Comm’rs of Los Angeles v. Jews for Jesus, Inc., 482 U.S. 569, 576 (1987); Heffron v. International Society for Krishna Consciousness, Inc., 452 U.S. 640, 649 (1981).
 See, e.g., Monitor Patriot Co. v. Roy, 401 U.S. 265, 272 (1971) (“[The First Amendment] has its fullest and most urgent application precisely to the conduct of campaigns for political office.”).
 See, e.g., Brown v. Hartlage, 465 U.S. 45, 53-54 (1982).
 See Boos v. Barry, 485 U.S. 312 (1988) (striking down a regulation of political speech near a foreign embassy); United States v. Grace, 461 U.S. 171 (1983) (striking down a ban on political displays on the grounds of the Supreme Court); Cohen v. California, 403 U.S. 15 (1971) (overturning the conviction of a draft protestor whose profane political message was emblazoned on his jacket); Mills v. Alabama, 384 U.S. 214 (1966) (striking down a ban on Election Day editorials).
 Cornelius v. NAACP Legal Defense and Educ. Fund, Inc., 473 U.S. 788, 806 (1985).
 504 U.S. 191 (1992).
 Id. at 196 (Blackmun, J., plurality opinion). By contrast, Justice Scalia, who concurred in the judgment, would have held that streets and sidewalks adjacent to polling places were nonpublic forums. See id. at 214-16 (Scalia, J., concurring in the judgment).
 Id. at 206 (plurality opinion).
 See id. at 210.
 See Minnesota Voters Alliance, 138 S. Ct. at 1887 (“The [Burson] plurality’s conclusion that the State was warranted in designating an area for the voters as ‘their own’ as they enter the polling place suggests an interest more significant, not less, within that place.”) (emphasis in original).
 It also prohibited the “display of campaign posters, signs or other campaign materials,” but the ban on displays was not specifically discussed by the Burson plurality.
 Minnesota Voters Alliance, 138 S. Ct. at 1887.
 Id. at 1886.
 See Minnesota Voters Alliance, 138 S. Ct. at 1886 (“A polling place in Minnesota qualifies as a nonpublic forum.”).
 Id. at 1891.
 Id. at 1890.
 Minn. Stat. § 211B.11(1).
 See Minnesota Voters Alliance, 138 S. Ct. at 1883.
 Id. at 1889.
 See id. at 1891 (quoting Tr. of Oral Arg. at 40).
 Id. at 1890.
 Id. at 1893 (Sotomayor, J. dissenting).
 See id. at 1891 (opinion of the Court) (offering the California and Texas statutes as examples or “more lucid” laws, but stating that “[w]e do not suggest that such provisions set the outer limit of what a State may proscribe, and do not pass on the constitutionality of laws that are not before us”).
 Cf. Doe v. Reed, 561 U.S. 186, 228 (Scalia, J., concurring in the judgment) (“There are laws against threats and intimidation; and harsh criticism, short of unlawful action, is a price our people have traditionally been willing to pay for self-governance. Requiring people to stand up in public for their political acts fosters civic courage, without which democracy is doomed.”).
 If the purpose of the forum were speech, the forum would not be a nonpublic forum. Rather, it would be a designated public forum, and speech restrictions would be evaluated under strict scrutiny. See Good News Club v. Milford Central School, 533 U.S. 98 (2001); Rosenberger v. Rector and Visitors of the University of Virginia, 515 U.S. 819 (1995).
 Minnesota Voters Alliance, 138 S. Ct. at 1889 n.4.
 Id. at 1887 (quoting respondents’ brief at 43).
 Id. at 1888.
 Id. It is difficult to know what this means, and even more difficult to understand how political apparel could provide such a distraction. The Court might mean that voting should be a solemn act and that voters’ clothing should reflect that attitude, but it is impossible to square that interest with a law that bans political apparel but permits people to vote in tank-tops and flip-flops. More likely, the Court means that the state should be able to promote a united front—the appearance that all voters are at the polling place “to reach considered decisions about their government and laws” in a non-partisan manner. Such an interest is farcical. Voters are there to choose one candidate over another, on (for most races, at least) partisan ballots. For the state positively to promote partisanship by printing partisan ballots and then to restrict speech based on the pretense that voters should not act in a partisan manner at the polling place is absurd.
In any event, political apparel does not distract from any realistic sense of civic obligation shared by voters. Everyone knows that voters support different candidates, and voting for one candidate over another—and advertising one’s intention to vote for one candidate over another—is fully consistent with the shared obligation to participate in self-government. Voters do not need to support the same candidate to support democracy, and people who are well aware that they support different candidates can still share the same commitment to self-government. It is also quite odd to justify a restriction on speech as reflecting the civic obligation to vote when, of course, there is no such obligation. Still less is there a shared obligation to vote without regard to party or ideology. Justifying a speech restriction as furthering a “sense” of an obligation, when that obligation does not exist, seems doubly odd.
 New York Times Co. v. Sullivan, 376 U.S. 254, 270 (1964).
 See, e.g., Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting) (classically expressing the marketplace-of-ideas theory behind the First Amendment). See also, e.g., Snyder v. Phelps, 562 U.S. 443, 461 (2011) (protecting the speech of funeral protestors, and noting that “even hurtful speech on public issues” must be protected “to ensure that we do not stifle public debate”).
 491 U.S. at 414.
 135 S. Ct. 2218 (2015).
 See Minnesota Voters Alliance, 138 S. Ct. at 1887.
 Cf. Lehman v. City of Shaker Heights, 418 U.S. 298 (1974).
 See Cohen, 403 U.S. at 21 (noting that persons offended by the words on Cohen’s jacket “could effectively avoid further bombardment of their sensibilities simply by averting their eyes”).
 See, e.g., Garcetti v. Ceballos, 547 U.S. 410 (2006); Rankin v. McPherson, 483 U.S. 378 (1987); Connick v. Myers, 461 U.S. 138 (1983); Pickering v. Board of Educ., 391 U.S. 563 (1968).
 See McPherson, 483 U.S. at 388-92; Pickering, 391 U.S. at 569-73.
 138 S. Ct. at 1887.
 393 U.S. 503 (1969).
 Id. at 518 (Black, J., dissenting).
 Id. at 508 (majority opinion).
 Id. at 514.
 138 S. Ct. at 1887 (quoting Board of Airport Comm’rs of Los Angeles, 482 U.S. at 576).
 Id. at 1887.
 The Court instead drew a parallel between polling places and courtrooms or legislatures, arguing that “[c]asting a vote is a weighty civic act, akin to a jury’s return of a verdict, or a representative’s vote on a piece of legislation.” Id. The comparison is peculiar if the Court wants to justify some political-apparel bans in polling places, because surely legislators cannot be prohibited from wearing political apparel or making political speeches in the legislature.
 Minnesota Voters Alliance’s suggestion that a content-based speech restriction at a polling place could enable voters to “focus on the important decisions immediately at hand,” id. at 1888, is in some tension with Tinker’s conclusion that the ban on arm bands was not necessary to enable students to focus on their studies. Even assuming that it is more important for voters than for students to focus, speech restrictions should be appropriate only when there is a significant risk that the speech will actually cause voters to lose focus.
 See, e.g., Morse v. Frederick, 551 U.S. 393 (2007) (permitting school officials to punish a student for expressing a pro-drug message at a school-sponsored event).
 Tinker, 393 U.S. at 506.