Jeffery Lichtenstein v. Tre Hargett

83 F.4th 575

• Court: Court of Appeals for the Sixth Circuit
• Decided: October 5, 2023
• Docket: 22-5028
• Status: Published

Opinion

RECOMMENDED FOR PUBLICATION Pursuant to Sixth Circuit I.O.P. 32.1(b) File Name: 23a0224p.06

UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT

┐ JEFFREY LICHTENSTEIN; MEMPHIS AND WEST │ TENNESSEE AFL-CIO CENTRAL LABOR COUNCIL; │ TENNESSEE STATE CONFERENCE OF THE NAACP; │ MEMPHIS A. PHILIP RANDOLPH INSTITUTE; FREE │ HEARTS, > No. 22-5028 Plaintiffs-Appellants, │ v. │ │ │ TRE HARGETT, in his official capacity as Tennessee │ Secretary of State; MARK GOINS, in his official │ capacity as Coordinator of Elections for the State of │ Tennessee; STEVEN JOHN MULROY, in his official │ capacity as District Attorney General for Shelby │ County, Tennessee, │ Defendants-Appellees. │ ┘

Appeal from the United States District Court for the Middle District of Tennessee at Nashville. No. 3:20-cv-00736—Eli J. Richardson, District Judge.

Argued: October 27, 2022

Decided and Filed: October 5, 2023

Before: McKEAGUE, WHITE, and MURPHY, Circuit Judges. _________________

COUNSEL

ARGUED: Danielle Lang, CAMPAIGN LEGAL CENTER, Washington, D.C., for Appellants. Clark Lassiter Hildabrand, OFFICE OF THE TENNESSEE ATTORNEY GENERAL, Nashville, Tennessee, for Appellees. ON BRIEF: Danielle Lang, Molly E. Danahy, Jonathan, Diaz, CAMPAIGN LEGAL CENTER, Washington, D.C., Jon M. Greenbaum, Ezra Rosenberg, Pooja Chaudhuri, LAWYERS’ COMMITTEE FOR CIVIL RIGHTS UNDER LAW, Washington, D.C., William L. Harbison, Lisa Katherine Helton, Christopher C. Sabis, SHERRARD ROE VOIGT & HARBISON PLC, Nashville, Tennessee, for Appellants. Janet M. Kleinfelter, Alexander S. Rieger, Matthew D. Cloutier, OFFICE OF THE TENNESSEE ATTORNEY GENERAL, Nashville, Tennessee, for Appellees. Kristin C. Cope, O’MELVENY No. 22-5028 Lichtenstein v. Hargett Page 2

& MYERS, LLP, Dallas, Texas, Charlotte Davis, PUBLIC INTEREST LEGAL FOUNDATION, Indianapolis, Indiana, for Amici Curiae.

MURPHY, J., delivered the opinion of the court in which McKEAGUE, J., joined. WHITE, J. (pp. 37–52), delivered a separate dissenting opinion.

_________________

OPINION _________________

MURPHY, Circuit Judge. Since 1979, Tennessee has made it a crime for anyone other than election officials to distribute the State’s official form for applying to vote absentee. During much of this time, Tennessee kept close guard of this form to deter fraud. But election officials now make the form widely available online so that eligible voters may more easily apply. According to the Plaintiffs, this change has rendered the ban on distributing the application form “outdated.” The Plaintiffs want to hand out this form while they encourage absentee voting at their get-out-the-vote drives. They allege that the First Amendment gives them the right to do so. Because they seek to distribute the form while expressing a political message, they argue, we must subject the ban to strict scrutiny. At the least, they say, we must evaluate the ban using the so-called “Anderson-Burdick” balancing test that applies to some election challenges.

We disagree on both fronts. Tennessee’s ban prohibits an act: distributing a government form. This act qualifies as conduct, not speech. Admittedly, the First Amendment provides some protection to “expressive conduct.” But strict scrutiny does not apply to Tennessee’s ban because it neutrally applies no matter the message that a person seeks to convey and because it burdens nobody’s ability to engage in actual speech. We have also never extended Anderson- Burdick’s balancing test to this sort of speech claim. At most, the Supreme Court’s lenient First Amendment test for neutral laws that regulate conduct applies here. And because the ban survives this nondemanding test, we affirm the district court’s dismissal of the Plaintiffs’ complaint. No. 22-5028 Lichtenstein v. Hargett Page 3

I.

A.

Tennessee permits all voters to vote early in person up to 20 days before most elections. Tenn. Code Ann. § 2-6-102(a)(1). But only a subset of voters may vote absentee through the mail. Id. § 2-6-201. The list of eligible absentee voters includes students, voters over 60, the hospitalized or disabled, and voters who will remain away from their home county during the voting period. Id. § 2-6-201(1)–(3)(A), (5). To vote absentee, a voter must “request an absentee ballot” from a county election commission within a certain time before the election. Id. § 2-6-202(a)(1).

Over the years, Tennessee has made it easier for eligible voters to vote absentee. Historically, if a voter sent a written request for an absentee ballot to a county election commission, the commission would treat this writing “as a request for an application for absentee ballot.” 1979 Tenn. Pub. Acts 687, 690. A state official created the official application “forms.” Id. To receive an absentee ballot, the voter would need to fill out this second document—the official application. Id. at 691. Tennessee limited access to this form. An election commission generally could send only one form to any voter. Id. at 690–91. The State also made it “a felony for any private person to supply an application for absentee ballot to any person by any means whatsoever.” Id. at 691.

Tennessee simplified the absentee-voting process in 1994. 1994 Tenn. Pub. Acts 633, 637–39 (creating Tenn. Code Ann. § 2-6-202). Today, a voter may submit both a written request for an application and the official application to an election commission “by mail, facsimile transmission or e-mail with an attached document that includes a scanned signature.” Tenn. Code Ann. § 2-6-202(a)(3). And if a voter’s mere “request” for an application contains several items—including the voter’s identifying information and the reason the voter wants to vote absentee—that “request serves as an application” itself. Id. § 2-6-202(a)(3)(A)–(G). This change eliminated the need for every voter to follow a two-step process by submitting a “request” for an application to vote absentee followed by the official “application.” If the voter’s request includes all required information, an election commission will now simply mail the voter No. 22-5028 Lichtenstein v. Hargett Page 4

the absentee ballot. Id. § 2-6-202(b). If not, the commission will send the voter the official application form. Id.

A state official still must provide each county election commission with the official “forms for applications for ballots or approve the usage of a county’s forms.” Id. § 2-6-202(c)(1). But Tennessee no longer keeps close guard of these government forms. State and local officials now post them “online,” so any person may freely “download and print” them. Compl., R.1, PageID 7. Tennessee also allows a voter to ask someone else to fill out a request for an application form or the form itself as long as the voter signs these documents. Tenn. Code Ann. § 2-6-203.

At the same time, a county election commission still may not provide more than one official absentee-ballot application form to a voter who requests one unless the voter states that a prior form has been “spoiled” or has not been received. Id. § 2-6-202(c)(2). And Tennessee law still prohibits everyone else from distributing these official forms: “A person who is not an employee of an election commission commits a Class E felony if such person gives an application for an absentee ballot to any person.” Id. § 2-6-202(c)(3).

Soon after the 1994 changes, an election official asked the Tennessee Attorney General for an opinion about the scope of this ban on distributing application forms. Tenn. Op. Att’y Gen. No. 95-003, 1995 WL 14087, at *1 (1995). A debate had arisen over whether the ban covered the distribution of privately made (and unofficial) requests for applications if those template documents included blank sections for the required information that would allow them to serve as unofficial applications under the recent amendments. See Tenn. Code Ann. § 2-6-202(a)(3). The Attorney General narrowly interpreted the ban in § 2-6-202(c)(3) to bar only the distribution of the official forms—not these “request” documents. 1995 WL 14087, at *3–4.

This narrow view led “various groups” to mass produce standard “request” documents that contained all required information to serve as applications and provide these “requests” to voters. Memphis A. Philip Randolph Inst. v. Hargett, 478 F. Supp. 3d 699, 704 (M.D. Tenn. 2020). That development concerned election administrators. Id. at 704–05. Some voters No. 22-5028 Lichtenstein v. Hargett Page 5

(including the elderly) who received these unofficial “request” documents thought they needed to fill them out even to vote at the polls. Id. at 705. But once they completed the request, these voters could not vote in person. Id. In 2002, Tennessee responded to these voter-confusion issues. Id. at 704–05. It added a separate ban that barred parties from distributing request forms for absentee-ballot applications: “A person who is not an employee of an election commission commits a Class A misdemeanor if such person gives an unsolicited request for application for absentee ballot to any person.” 2002 Tenn. Pub. Acts ch. 698, at 1 (adding Tenn. Code Ann. § 2-6-202(c)(4)). Unlike the ban on the distribution of the official application forms in § 2-6-202(c)(3), this ban more narrowly applies to the unsolicited distribution of privately created requests.

B.

Today’s suit involves only the ban on the distribution of the official application forms in § 2-6-202(c)(3) (what we will call “paragraph (c)(3)”). The five organizational Plaintiffs—the Memphis and West Tennessee AFL-CIO Central Labor Council (Labor Council), the Tennessee State Conference of the NAACP, the Equity Alliance, the Memphis A. Philip Randolph Institute (APRI), and Free Hearts—engage in voter-outreach efforts as part of their diverse missions. Compl., R.1, PageID 3–6. For instance, the Labor Council and the NAACP regularly educate their thousands of members about the voting process and encourage them to vote. Id., PageID 3–4. Similarly, APRI “sponsors voter education and Get-Out-The-Vote programs in the community.” Id., PageID 5. The individual Plaintiff, Jeffrey Lichtenstein, likewise undertakes “voter advocacy and engagement efforts” in his role as the Labor Council’s Executive Secretary. Id., PageID 2.

According to the Plaintiffs, their voter-outreach efforts sit “at the core of [their] political speech and advocacy activities.” Id., PageID 8. As a part of these efforts, the Plaintiffs encourage eligible Tennessee voters to vote absentee through the mail. Id., PageID 9. They regularly explain the “benefits of voting by mail” and the “submission deadlines and requirements” for absentee voting to their own members and to the public. Id. No. 22-5028 Lichtenstein v. Hargett Page 6

The Plaintiffs seek to distribute Tennessee’s absentee-ballot application form during these voter-outreach efforts. Id., PageID 2–6. They more often convince voters to cast a ballot when they can provide the voters with all the “requisite forms they might need[.]” Id., PageID 9. In other words, voters who may vote absentee will more likely apply if the Plaintiffs can hand them an application form than if the Plaintiffs must send them “to a website they may not be able to access, or to a form they may not be able to print.” Id., PageID 10. In fact, some voters have asked the Plaintiffs for the forms because they “lack reliable access to a computer, a printer, or the Internet.” Id. The Plaintiffs allege that unless they can hand voters these official forms, some will choose not to apply for an absentee ballot or cast a ballot at all. Id. They thus describe paragraph (c)(3)’s ban as an “extraordinarily burdensome” limit on their advocacy. Id., PageID 8.

To allow them to distribute the forms ahead of the 2020 election, the Plaintiffs brought two suits. In an initial suit, some Plaintiffs challenged the separate ban in § 2-6-202(c)(4) on distributing unsolicited requests for applications. The district court denied a preliminary injunction against this separate ban because paragraph (c)(3) (not (c)(4)) was the one that prevented them from engaging in their desired conduct of distributing the official forms. Memphis A. Philip Randolph Inst., 478 F. Supp. 3d at 707–10. The Plaintiffs responded by quickly filing this challenge to paragraph (c)(3) against three officials (collectively, “the State”). See Lichtenstein v. Hargett, 489 F. Supp. 3d 742, 748–50 (M.D. Tenn. 2020). Their complaint asserted two claims under the First Amendment. They first alleged that paragraph (c)(3)’s ban violated their freedom of speech by limiting their “core political speech and expressive conduct[.]” Compl., R.1, PageID 11–12. They next alleged that it violated their freedom of association by restricting their members’ advocacy toward voters. Id., PageID 12.

The district court dismissed the complaint. Lichtenstein v. Hargett, __ F. Supp. 3d __, 2021 WL 5826246, at *6–8 (M.D. Tenn. Dec. 7, 2021). To do so, it incorporated its analysis from an earlier opinion denying a preliminary injunction. See id. at *6–7 (citing Lichtenstein, 489 F. Supp. 3d 742). The court initially held that paragraph (c)(3)’s ban did not cover “expressive” conduct protected by the First Amendment. See id. at *6. Even if this ban did bar expressive conduct, the court next held that the ban would not trigger strict scrutiny because it No. 22-5028 Lichtenstein v. Hargett Page 7

did not limit “‘core’ political speech[.]” Id. Instead, the court applied a “rational-basis ‘plus’” test using the “Anderson-Burdick” framework from the election context. Id. at *7. The court lastly found that the law survived this test given the State’s interests in preventing voter confusion. See id.

The Plaintiffs appealed. They have renewed their freedom-of-speech and freedom-of- association claims. We will consider each claim in turn, reviewing the district court’s decision to dismiss their complaint de novo. See Daunt v. Benson, 999 F.3d 299, 307 (6th Cir. 2021).

II. Does Tennessee’s Ban Infringe the Plaintiffs’ Freedom of Speech?

As incorporated against the states by the Fourteenth Amendment, the First Amendment provides: “Congress shall make no law . . . abridging the freedom of speech, or of the press[.]” U.S. Const. amend. I; see Stromberg v. California, 283 U.S. 359, 368 (1931). This text poses an obvious obstacle for the Plaintiffs. It bars a state from “abridging” oral expression (the freedom of “speech”) or written expression (the freedom of the “press”); it does not bar the state from restricting conduct. See Rumsfeld v. Forum for Acad. & Institutional Rts., Inc., 547 U.S. 47, 62 (2006); 2 Noah Webster, An American Dictionary of the English Language (1828); 2 Samuel Johnson, A Dictionary of the English Language (4th ed. 1773). And the Plaintiffs do not dispute that a person “seek[s] to engage in” “conduct” (not speech) when the person hands an official form to another person. Appellants’ Br. 14.

But this fact does not end the inquiry. Even before the founding, Americans expressed political messages through symbolic conduct—whether by burning the king in effigy or by raising liberty poles. See Eugene Volokh, Symbolic Expression and the Original Meaning of the First Amendment, 97 Geo. L.J. 1057, 1061 & n.20 (2009). The Plaintiffs likewise argue that they want to distribute the absentee-ballot application in coordination with their message to vote absentee. So they allege that they seek to engage in “expressive conduct.” Appellants’ Br. 14. And because the message that they want to convey through this conduct qualifies as “core political speech,” they argue, we must review paragraph (c)(3)’s ban using the “exacting” (that is, strict) scrutiny from Meyer v. Grant, 486 U.S. 414, 420 (1988). Alternatively, they argue that we must evaluate it under the Anderson-Burdick balancing test that applies to election laws. No. 22-5028 Lichtenstein v. Hargett Page 8

These arguments misunderstand the Supreme Court’s First Amendment framework for expressive conduct. We will explain why in three parts. First, even if the First Amendment provides some protection for the Plaintiffs’ proposed actions, it does not require us to review the challenged ban using the strict scrutiny that they seek. See Part II.A. Second, Plaintiffs’ speech claim also does not trigger the Anderson-Burdick balancing test because we have never applied this test to their type of speech claim. See Part II.B. Third, at most, the Plaintiffs’ claim triggers the expressive-conduct test originating from United States v. O’Brien, 391 U.S. 367, 377 (1968). And for the reasons provided by the district court, the law passes muster under that “relatively lenient” test. Texas v. Johnson, 491 U.S. 397, 407 (1989). See Part II.C.

A. Does Tennessee’s Ban Trigger Strict Scrutiny?

1. The First Amendment’s protections reach their “zenith” for political speech. Meyer, 486 U.S. at 425. And paragraph (c)(3) bans conduct that the Plaintiffs wish to undertake to express a political message. Connecting these two dots, the Plaintiffs argue that we must evaluate this ban using Meyer’s “exacting scrutiny.” Id. at 420. But their logical chain confuses the deferential framework that governs regulations targeting conduct (even conduct meant to convey a political message) for the demanding framework that governs regulations targeting expression. The demanding framework governing political expression does not apply here because the act of distributing an official form qualifies as conduct rather than “core political speech,” id., and because paragraph (c)(3) does not target or impose costs on the Plaintiffs’ actual political speech.

Laws Targeting Conduct. The Supreme Court has never applied exacting free-speech scrutiny to laws that bar conduct based on the harm that the conduct causes apart from the message it conveys. See R.A.V. v. City of St. Paul, 505 U.S. 377, 385 (1992). That is true even if the ban on conduct imposes “incidental burdens on speech.” Sorrell v. IMS Health Inc., 564 U.S. 552, 567 (2011); Expressions Hair Design v. Schneiderman, 581 U.S. 37, 47 (2017). And it is true even if the person engaging in the conduct “intends thereby to express an idea.” O’Brien, 391 U.S. at 376. No. 22-5028 Lichtenstein v. Hargett Page 9

Title VII offers a classic example of such a conduct ban that incidentally burdens speech. The statute bars employers from discriminating on the basis of race in employment decisions. 42 U.S.C. § 2000e–2(a). By doing so, Title VII has the practical effect of stopping a store owner from posting a “White Applicants Only” sign in the storefront because that speech would provide smoking-gun evidence of the owner’s illegal intent. Rumsfeld, 547 U.S. at 62. But Title VII does not trigger rigorous free-speech scrutiny simply because it affects speech. Why? Congress can justify Title VII’s ban on “bias-inspired” employment decisions by the “harm” that these decisions cause employees “over and above” a disapproval of the message that the decisions send. Wisconsin v. Mitchell, 508 U.S. 476, 487–88 (1993); see Hishon v. King & Spalding, 467 U.S. 69, 78 (1984). Employees who lose out on a job suffer tangible injuries from an employer’s refusal to hire them distinct from the employer’s expression of a discriminatory point of view.

Many cases follow the same path. The Court has not applied rigorous scrutiny to a conduct ban on public nudity even as applied to those who sought to express an “erotic message” through “nude dancing.” Barnes v. Glen Theatre, Inc., 501 U.S. 560, 569–71 (1991) (plurality opinion); see City of Erie v. Pap’s A.M., 529 U.S. 277, 289–96 (2000) (plurality opinion). It has not applied rigorous scrutiny to a conduct ban on sleeping overnight in national parks even as applied to those who sought to sleep in a park to highlight “the plight of the homeless.” Clark v. Cmty. for Creative Non-Violence, 468 U.S. 288, 289, 293–99 (1984). And it has not applied rigorous scrutiny to a conduct ban on destroying draft cards even as applied to those who burned the cards to express “antiwar beliefs[.]” O’Brien, 391 U.S. at 369–70, 376–77.

If rigorous free-speech scrutiny does not apply, what does? As long as a ban on conduct is “unrelated to the suppression of free expression,” the ban will trigger, at most, the “relatively lenient” test that O’Brien adopted to uphold the law against destroying draft cards. Johnson, 491 U.S. at 407 (citation omitted). This test does not protect just any conduct; it protects only “inherently expressive” conduct. Rumsfeld, 547 U.S. at 66. And it requires that the government merely identify a “substantial” “interest” that would not be furthered as “effectively” without the ban. Id. at 67 (quoting United States v. Albertini, 472 U.S. 675, 689 (1985)). No. 22-5028 Lichtenstein v. Hargett Page 10

Laws Targeting Speech. The Court, by contrast, employs a far different test to evaluate laws that target expression. Some laws do so directly. Think of an “outright ban” on political speech. Citizens United v. FEC, 558 U.S. 310, 337 (2010). The Court has applied strict scrutiny to this type of law that prohibits certain disfavored messages. See id. at 340. The government must show both that it has a “compelling interest” for the ban and that the ban is the “least restrictive means” to achieve this interest. McCullen v. Coakley, 573 U.S. 464, 478 (2014).

But many laws burden expression only indirectly. In two circumstances, the Court has applied strict scrutiny to laws that targeted expression by regulating conduct. Sometimes, the government might bar conduct because of “the ideas it expresses” rather than the harms it causes. R.A.V., 505 U.S. at 385. The Supreme Court’s famous flag-burning decision exemplifies this type of speech-targeting law. See Johnson, 491 U.S. at 406–07. The Texas law there did not ban all flag burning because of the risks of “outdoor fires[.]” R.A.V., 505 U.S. at 385. It banned only flag burning done to convey a message “dishonoring” the flag. Id. In particular, the law applied to flag burning that would seriously offend its viewers. See Johnson, 491 U.S. at 400 n.1, 411–12. The Court applied strict scrutiny because a violation “depended” not on the conduct-related risks of burning things but on the “communicative impact” of the action. Id. at 411.

Other times, the government might restrict the “inputs” that speakers use to express a message. If, for example, the government banned the sale of ink for the use in political pamphlets, it ostensibly would be regulating conduct—the sale of a commodity. See Sorrell, 564 U.S. at 571. But the Court would rigorously review this law because it targets certain speakers by burdening the written words for which they will use the ink. See id.

The case on which the Plaintiffs rely—Meyer—addressed a law that burdened another speech “input”: money. Beginning with Buckley v. Valeo, 424 U.S. 1 (1976) (per curiam), the Court has repeatedly addressed whether bans on giving money to create speech trigger heightened scrutiny. Buckley considered, among other things, limits on the amounts that people could spend on speech promoting political candidates. Id. at 12–13. The Court subjected these limits to “the exacting scrutiny applicable to limitations on core First Amendment rights of political expression.” Id. at 44–45. It explained that the limits restricted the speech that the No. 22-5028 Lichtenstein v. Hargett Page 11

money paid for. Id. at 16. And this speech did not lose its protection just because it depended on money for its creation. Id.

Meyer fits Buckley’s mold. It considered a Colorado statute governing “initiative” petitions that sought to give voters the chance to vote on new laws or constitutional amendments. 486 U.S. at 415–16. Under Colorado’s process, petitioners had to collect a certain number of signatures before they could place a proposed law or amendment on the ballot. Id. at 416. But the challenged statute barred petitioners from paying circulators to gather these signatures. Id. at 416–17 & n.1. Citing Buckley, the Court subjected this payment limit to what it called “exacting scrutiny” because it restricted “political expression.” Id. at 420. The Court reasoned that the statute targeted expression (not just the conduct of paying circulators) because the circulators needed to speak to voters to convince them to sign the petition. Id. at 421. They engaged in “both the expression of a desire for political change and a discussion of the merits of the proposed change.” Id. And the law’s ban on paying for this “political speech” had “the inevitable effect of reducing the total quantum of” it. Id. at 422–23. If petitioners could not pay circulators, fewer “voices” could convey their message. See id. It was these “voices”—the “speech through petition circulators”—that garnered the First Amendment’s highest protections. Id. at 422, 424.

These types of burdens on speech “inputs” can take other forms. Such burdens can include, for example, restrictions on who may convey a message. After Meyer, therefore, the Court found unconstitutional a Colorado ban on the use of petition circulators who were not registered voters. See Buckley v. Am. Const. L. Found., Inc., 525 U.S. 182, 192–97 (1999). Because this ban reduced the number of people who could talk with potential petition signers, it too generated “a speech diminution of the very kind produced by the ban on paid circulators at issue in Meyer.” Id. at 194.

In sum, Meyer applied heightened scrutiny because the Colorado statute targeted speech by restricting the conduct that created the speech. And while Meyer used the phrase “exacting scrutiny” to describe the governing test, id. at 420, the Court applied standards that today go by “strict scrutiny.” It engaged in a least-restrictive-means analysis, finding that the ban was not “necessary” because Colorado could serve its interests by means that limited less speech. Id. at No. 22-5028 Lichtenstein v. Hargett Page 12

426–27; cf. Ams. for Prosperity Found. v. Bonta, 141 S. Ct. 2373, 2383 (2021) (plurality opinion).

2. How does this divide apply to paragraph (c)(3)’s ban on “giv[ing] an application for an absentee ballot to any person”? For two reasons, strict scrutiny does not apply to this conduct ban. The ban does not limit a speech “input” (like the laws in Buckley, Meyer, and American Constitutional Law Foundation). And it does not target the conduct because of the “idea” that the Plaintiffs want to express (like the law in Johnson).

Reason One: Paragraph (c)(3) does not regulate something that the Plaintiffs use to speak and thereby target or burden that speech. Unlike the ink that a party uses to create written speech, see Sorrell, 564 U.S. at 571, or the money or people that a party uses to create oral speech, see Am. Const. L. Found., 525 U.S. at 192–97; Meyer, 486 U.S. at 422–23, the distribution of official absentee-ballot application forms is not a speech “input.” To be sure, the Plaintiffs’ underlying get-out-the-vote activities—that is, their speech to convince voters to vote absentee—qualifies as “core political speech” entitled to rigorous First Amendment protection. Meyer, 486 U.S. at 422. But nothing in paragraph (c)(3) in any way restricts the Plaintiffs’ actual oral or written speech about the “benefits” of absentee voting. Compl., R.1, PageID 9. Nor does this statute make the creation of this speech “more costly” and thereby reduce its volume under the basic laws of supply and demand. Citizens for Tax Reform v. Deters, 518 F.3d 375, 388 (6th Cir. 2008). In fact, unlike the ban on paying petition circulators in Meyer, the Plaintiffs’ complaint does not even allege that paragraph (c)(3) will have any “effect” on the “quantum” of their oral or written speech encouraging absentee voting. 486 U.S. at 423.

Instead, the Plaintiffs argue (and our colleague’s thoughtful dissent agrees) that the ban on distributing the government form in this case triggers strict scrutiny because that distribution is “intertwine[d]” with or “involves” the actual political speech at their get-out-the-vote drives. Appellants’ Br. 30; Dissenting Op., at 47–48. We do not read Meyer this broadly. There, the Court held that the ban on paying petition circulators triggered strict scrutiny not because the payment prohibition was intertwined with speech but because the ban would have “the inevitable effect of reducing the total quantum of” it. Meyer, 486 U.S. at 423 (emphasis added). Why? Because a petition circulator’s efforts to convince the public to sign initiative petitions qualified No. 22-5028 Lichtenstein v. Hargett Page 13

as “both the expression of a desire for political change and a discussion of the merits of the proposed change.” Id. at 421. So the ban in Meyer limited the “direct one-on-one communication” that all agree is pure political expression. Id. at 424. Or, as a later case put it, the ban caused a “diminution” of this “speech” and so “impose[d] a burden” on it. Am. Const. L. Found., 525 U.S. at 194–95 (quoting Meyer, 486 U.S. at 428). We see nothing in Meyer that adopts the Plaintiffs’ broader reading, which asks whether the regulated conduct is “intertwined” with actual speech (rather than whether it burdens that speech). And we doubt the Supreme Court would accept this amorphous free-speech test. How much “intertwinedness” is necessary? How is it measured?

The Plaintiffs’ test would also call into doubt many of the Supreme Court’s expressive- conduct cases because conduct often accompanies speech. For example, David Paul O’Brien’s draft-card burning likewise could be said to “intertwine” with his speech conveying antiwar views. See O’Brien, 391 U.S. at 370. But the Court refused to transform this conduct into speech subject to strict scrutiny just because O’Brien conveyed a political message while burning the draft card. Id. at 376–77. Why was his conduct not intertwined enough with his speech to trigger strict scrutiny under the Plaintiffs’ view? See also Clark, 468 U.S. at 296.

To make their case fall within Meyer’s narrower holding, then, the Plaintiffs needed to show that paragraph (c)(3) burdened their actual oral or written speech by restricting conduct that helps produce it. If, for example, Tennessee barred the Plaintiffs from paying their employees to promote absentee voting, they may have a strong case for strict scrutiny. Cf. Emily’s List v. FEC, 581 F.3d 1, 8–11 (D.C. Cir. 2009). But again, the Plaintiffs have not alleged that the ban on distributing forms in any way restricts their ability to create or convey speech.

That said, the Plaintiffs do allege that paragraph (c)(3)’s ban on conduct makes it harder to achieve their bottom-line goal of increasing absentee voting. This outcome-focused effect, however, qualifies as the type of “incidental burden” on speech that does not trigger heightened scrutiny and instead triggers the more-lenient test for expressive conduct. Sorrell, 564 U.S. at 567. The Supreme Court’s precedent confirms this point. The ban on destroying draft cards might have reduced O’Brien’s ability to spotlight his “antiwar” views and made it harder to get those views enacted as our foreign policy. O’Brien, 391 U.S. at 370. Likewise, the ban on No. 22-5028 Lichtenstein v. Hargett Page 14

sleeping in parks might have made it tougher for the Community for Creative Non-Violence to convince others about “the plight of the poor and homeless” and so reduced its ability to enact laws consistent with its views. Clark, 468 U.S. at 296. But the Court did not apply strict scrutiny in either case. See id. at 293–99; O’Brien, 391 U.S. at 376–82; see also FTC v. Superior Ct. Trial Lawyers Ass’n, 493 U.S. 411, 429–32 (1990). Yet the plaintiffs in both cases presumably engaged in their chosen course of conduct because, like the Plaintiffs here, they thought it would be a “more effective way to encourage” their goals. Compl., R.1, PageID 9–10.

Indeed, the Plaintiffs’ argument—that a ban on conduct triggers strict scrutiny if the ban makes it harder for an entity to achieve the policies it promotes—has no stopping point. Ordinary speed limits might increase the time it takes speakers to travel in between venues and so reduce their speech’s reach and its chances of achieving a desired result. Does that outcome subject speed limits to strict speech scrutiny too? Of course not. The Plaintiffs’ theory that paragraph (c)(3)’s ban makes it more difficult to increase absentee voting likewise provides no basis to subject the ban to strict scrutiny. See Voting for Am., Inc. v. Steen, 732 F.3d 382, 391 n.5 (5th Cir. 2013); Initiative & Referendum Inst. v. Walker, 450 F.3d 1082, 1099–1101 (10th Cir. 2006) (en banc).

Reason Two: Paragraph (c)(3)’s ban does not target the distribution of application forms because of “the ideas” that the Plaintiffs “express[]” (instead of “the action [that] it entails”). R.A.V., 505 U.S. at 385. This fact distinguishes the law in Johnson, which targeted expression by barring only flag burning designed to offend others. See 491 U.S. at 411–12. Paragraph (c)(3)’s ban, by comparison, does not turn on the message conveyed when distributing a form. It bars this distribution when done for any reason. So it undoubtedly bans the distribution by those who seek to convince recipients to vote absentee. But it equally bans the distribution by those who, for example, seek to convince recipients that Tennessee makes it too easy to vote absentee and so increases the risks of fraud. And it bans the distribution by those who seek to wrongly convince voters who are ineligible to vote absentee that they may do so in an effort to deprive them of the vote. In short, a violation does not depend on the “message expressed” through a specific individual’s conduct. Reed v. Town of Gilbert, 576 U.S. 155, 163–64 (2015). The No. 22-5028 Lichtenstein v. Hargett Page 15

conduct ban is thus content neutral. See id.; cf. Nev. Comm’n on Ethics v. Carrigan, 564 U.S. 117, 125 (2011).

The Plaintiffs counter that the Tennessee law flunks the content-neutrality test because it bars the distribution of application forms but does not bar, say, the distribution of voter-registration forms. Reply Br. 14 n.6. But the law in O’Brien likewise did not prohibit people from tearing up their tax returns. See 391 U.S. at 375. Yet this differential treatment did not render the law content based. See id. at 376–82. That is because the First Amendment’s content-neutrality test asks whether a law treats different messages differently, not whether it treats different conduct differently. See Reed, 576 U.S. at 163. The Court evaluates conduct- based differences like the one that the Plaintiffs propose under the Equal Protection Clause, not the Free Speech Clause. See Armour v. City of Indianapolis, 566 U.S. 673, 680–81 (2012); City of New Orleans v. Dukes, 427 U.S. 297, 303–04 (1976) (per curiam). But the Plaintiffs do not assert an equal-protection claim.

The Plaintiffs also point to the Supreme Court’s recognition that a state may not avoid strict scrutiny for a speech restriction on the ground that it leaves open other ways to convey a message. See Cal. Democratic Party v. Jones, 530 U.S. 567, 581 (2000); Meyer, 486 U.S. at 424. The government thus cannot ban a book on the ground that it permits the author to give speeches about the book. But this rule applies only to laws that target speech. And the State does not defend paragraph (c)(3) on the ground that it limits some speech but leaves open other expressive avenues. The State defends paragraph (c)(3) on the ground that the law does not target speech at all.

That fact leads to a final point. Perhaps we need not ask whether the ban on distributing application forms burdens the Plaintiffs’ other speech at their get-out-the-vote drives. One might analogize the form’s distribution to that of a political pamphlet. And the Court has long treated the latter distribution (or more precisely, the distribution of a pamphlet’s expressive contents) as speech itself. See McCullen, 573 U.S. at 488–89; McIntyre v. Ohio Elections Comm’n, 514 U.S. 334, 341–45 (1995). Just as a person can express a message through oral communication in a “one-on-one” interaction, Meyer, 486 U.S. at 424, so too the person can express that message by No. 22-5028 Lichtenstein v. Hargett Page 16

handing out a written communication. Unlike our colleague in dissent, though, the Plaintiffs did not make this analogy.

Nor is it apparent to us that this unbriefed analogy to political pamphlets withstands scrutiny. For one thing, the application is a form that the State creates. Tenn. Code Ann. § 2-6-202(c)(1). It is safe to say that this government form does not resemble the ideological pamphlets that fueled the American Revolution. See Bernard Bailyn, The Ideological Origins of the American Revolution 1–21 (50th Anniversary ed. 2017). If the form’s contents were speech, the speech might well be the government’s. Cf. Shurtleff v. City of Boston, 142 S. Ct. 1583, 1589–90 (2022). For another thing, just as the challenged Tennessee law does not burden oral speech, it also does not burden written speech. Paragraph (c)(3) freely permits the Plaintiffs to distribute private pamphlets to express any message they desire about “the existence of absentee voting, the steps to apply, who is eligible to vote this way, and the deadline for submission.” Dissenting Op., at 45. So why are the Plaintiffs not content to circulate their own private “publication” “of information and opinion” about absentee voting? Lovell v. City of Griffin, 303 U.S. 444, 452 (1938). They wish to distribute the application for a separate functional (not expressive) reason: by giving voters the “requisite forms” needed to apply, the Plaintiffs make it easier for these voters to complete the application process. Compl., R.1, PageID 9. If they have a “speech” right to engage in that activity, do they also have a speech right to hand out return envelopes or postage stamps (an activity that can also make it easier for potential voters to apply)? A dubious claim at best. In any event, the Plaintiffs have argued that the banned conduct is entitled to rigorous protection only because it is intertwined with speech (under Meyer) or because the relevant ban is content based (under Johnson). They are mistaken on both points.

B. Does Tennessee’s Ban Trigger Anderson-Burdick “Balancing”?

Just because paragraph (c)(3)’s ban does not trigger strict scrutiny does not mean it falls outside the First Amendment altogether. The Plaintiffs at least ask us to evaluate their claim under the Anderson-Burdick balancing test used in the election context. See Anderson v. Celebrezze, 460 U.S. 780 (1983); Burdick v. Takushi, 504 U.S. 428 (1992). The test requires courts to balance an election law’s burdens on voters’ rights against the state’s interests in the No. 22-5028 Lichtenstein v. Hargett Page 17

law as a threshold inquiry to determine the level of scrutiny that courts should apply to the law. See Ohio Democratic Party v. Husted, 834 F.3d 620, 626–27 (6th Cir. 2016). Laws that impose minimal burdens need only satisfy something approaching rational-basis review; laws that impose severe burdens must satisfy something approaching strict scrutiny; and laws in between must satisfy a level of scrutiny commensurate with their burdens. See id. But we refuse to expand this test to cover a pure speech claim like the one that the Plaintiffs assert here.

1. The Anderson-Burdick balancing test has historically applied to claims that an election law interferes with the right of voters to vote or political parties to associate with voters—not the right of speakers to speak. In particular, we and the Supreme Court have applied this test to three types of claims: ballot-access claims, political-party associational claims, and voting-rights claims.

First, start with the cases from which this approach takes its name: Anderson and Burdick. Both cases addressed ballot-access challenges to laws that limited the candidates who could appear on a ballot. The Court explained that these laws implicate two “rights of voters”: the implied right to vote (under the substantive part of the Fourteenth Amendment’s Due Process Clause) and the implied right to associate with a candidate by voting for the candidate (under the First Amendment). Anderson, 460 U.S. at 786–88. To decide whether a ballot-access law violates these rights, the Court held, the judiciary must identify the “character and magnitude” of the harm to the rights and compare that harm to the “precise interests” that the state used to justify the law. Id. at 789. Applying this test, Anderson found unconstitutional an Ohio law that imposed a uniquely early deadline for independent candidates to apply to make the ballot. 460 U.S. at 790–806. But Burdick upheld a Hawaii law that banned voters from writing in candidates. 504 U.S. at 434–41.

Our cases have often applied Anderson-Burdick’s test to ballot-access challenges. The following cases have used this test to evaluate a claim that a law limiting a candidate’s ability to get on the ballot violated the rights of voters to vote and the rights of candidates and voters to associate: Graveline v. Benson, 992 F.3d 524, 534–46 (6th Cir. 2021); Kishore v. Whitmer, 972 F.3d 745, 749–51 (6th Cir. 2020); Hawkins v. DeWine, 968 F.3d 603, 605–07 (6th Cir. 2020); Esshaki v. Whitmer, 813 F. App’x 170, 171–72 (6th Cir. 2020) (order); Libertarian Party No. 22-5028 Lichtenstein v. Hargett Page 18

of Ky. v. Grimes, 835 F.3d 570, 574–78 (6th Cir. 2016); Libertarian Party of Ohio v. Husted, 831 F.3d 382, 399–405 (6th Cir. 2016); Green Party of Tenn. v. Hargett, 791 F.3d 684, 692–95 (6th Cir. 2015); Green Party of Tenn. v. Hargett, 767 F.3d 533, 545–49 (6th Cir. 2014); Jolivette v. Husted, 694 F.3d 760, 766–77 (6th Cir. 2012); Morrison v. Colley, 467 F.3d 503, 507–08 (6th Cir. 2006); Libertarian Party of Ohio v. Blackwell, 462 F.3d 579, 585–95 (6th Cir. 2006); Lawrence v. Blackwell, 430 F.3d 368, 372–75 (6th Cir. 2005); Gable v. Patton, 142 F.3d 940, 946–47 (6th Cir. 1998); Miller v. Lorain Cnty. Bd. of Elections, 141 F.3d 252, 256–59 (6th Cir. 1998); Corrigan v. City of Newaygo, 55 F.3d 1211, 1217–18 (6th Cir. 1995).

We also recently expanded the ballot-access claims that fall within this group of cases. Adding to a circuit conflict, our court now uses Anderson-Burdick’s test to evaluate challenges to the rules for getting initiatives—not just candidates—on the ballot. See Beiersdorfer v. LaRose, 2021 WL 3702211, at *9–12 (6th Cir. Aug. 20, 2021); Thompson v. DeWine, 976 F.3d 610, 615–19 (6th Cir. 2020) (per curiam); SawariMedia, LLC v. Whitmer, 963 F.3d 595, 596–98 (6th Cir. 2020) (order); Schmitt v. LaRose, 933 F.3d 628, 639–42 (6th Cir. 2019); Comm. to Impose Term Limits on Ohio Sup. Ct. v. Ohio Ballot Bd., 885 F.3d 443, 448 (6th Cir. 2018).

Second, the Supreme Court has applied Anderson-Burdick’s test to claims that election laws violate the rights of political parties to associate with candidates (or vice versa) under the First Amendment. The Court has used this test to resolve challenges to the information that a state puts on its ballot and to the way the state conducts primaries. So the Court relied on Anderson-Burdick to decide whether a state violated a political party’s associational rights by allowing candidates to identify the party as their preferred choice on the ballot. See Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 451–58 (2008). And it relied on Anderson-Burdick to decide whether a state violated a party’s rights by barring “fusion” candidates who sought to identify themselves as the nominee of two parties on the ballot. See Timmons v. Twin Cities Area New Party, 520 U.S. 351, 359 (1997). Similarly, the Court has used Anderson-Burdick to decide whether a state infringed a party’s rights by regulating the voters who could vote in its primary. Compare Clingman v. Beaver, 544 U.S. 581, 586–97 No. 22-5028 Lichtenstein v. Hargett Page 19

(2005), with Cal. Democratic Party, 530 U.S. at 572–86; Tashjian v. Republican Party of Conn., 479 U.S. 208, 213–29 (1986).

Some of our cases fit within this group. Like Timmons, we have invoked the Anderson- Burdick test to evaluate laws limiting a political party’s ability to identify its association with a candidate on a state’s official ballot. See Ohio Council 8 Am. Fed. of State v. Husted, 814 F.3d 329, 334–40 (6th Cir. 2016); Schrader v. Blackwell, 241 F.3d 783, 787–91 (6th Cir. 2001).

Third, the Supreme Court has applied the Anderson-Burdick test to claims that election laws violate a voter’s right to vote under the Fourteenth Amendment’s Due Process or Equal Protection Clauses. Most notably, in Crawford v. Marion County Election Board, 553 U.S. 181 (2008), the lead opinion (speaking for three Justices) invoked Anderson-Burdick to uphold a law requiring voters to present photo identification. Id. at 189–203 (lead opinion). Effectively applying substantive-due-process analysis, this opinion said that “even rational restrictions” on the “right to vote” will fail if they do not relate to “voter qualifications.” Id. at 189.

We too have often applied this test to resolve claims that laws infringed a voter’s due-process or equal-protection right to vote. We have balanced a voter’s right to vote against the state’s interests in the following voting rules: rules governing absentee-ballot drop boxes, see A. Philip Randolph Inst. of Ohio v. LaRose, 831 F. App’x 188, 191–92 (6th Cir. 2020) (order); rules governing signatures on ballots, see Memphis A. Philip Randolph Inst. v. Hargett, 978 F.3d 378, 390–91 (6th Cir. 2020); rules governing confined voters, see Mays v. LaRose, 951 F.3d 775, 783–93 (6th Cir. 2020); rules governing straight-ticket voting, see Mich. State A. Philip Randolph Inst. v. Johnson, 749 F. App’x 342, 349–50 (6th Cir. 2018); Mich. State A. Philip Randolph Inst. v. Johnson, 833 F.3d 656, 662–66 (6th Cir. 2016); rules governing vote counting, George v. Hargett, 879 F.3d 711, 724–28 (6th Cir. 2018); Stein v. Thomas, 672 F. App’x 565, 570 (6th Cir. 2016); rules governing early voting, see Ohio Democratic Party, 834 F.3d at 626–36; Obama for Am. v. Husted, 697 F.3d 423, 428–36 (6th Cir. 2012); and rules governing the completion of absentee and provisional ballots, see Ne. Ohio Coal. for the Homeless v. Husted, 837 F.3d 612, 631–35 (6th Cir. 2016); Ne. Ohio Coal. for the Homeless v. Husted, 696 F.3d 580, 591–97 (6th Cir. 2012). No. 22-5028 Lichtenstein v. Hargett Page 20

2. We would have to expand Anderson-Burdick’s balancing test into uncharted territory if we applied it here. The Plaintiffs’ speech claim looks nothing like the claims in these three groups of cases. For starters, paragraph (c)(3) imposes no conditions to get on Tennessee’s ballot. So the Plaintiffs do not argue that it infringes their due-process right to vote for (or First Amendment right to associate with) candidates on that ballot. See Anderson, 460 U.S. at 786–88. And they cannot pursue a ballot-access claim like those in Anderson, Burdick, or the first group of cases.

Next, the Plaintiffs are not political parties. Nor does paragraph (c)(3) regulate those parties. So the Plaintiffs do not argue that this law violates their First Amendment right of political association by barring them from associating with candidates on the ballot or from using their chosen method of picking nominees. Their claim thus does not resemble the freedom-of-association claims from Timmons, California Democratic Party, or the second group of cases.

Lastly, the Plaintiffs have not alleged that paragraph (c)(3)’s ban on distributing absentee-ballot application forms infringes anyone’s right to vote by making it too difficult to cast a ballot. See Crawford, 553 U.S. at 190–91 (lead opinion). So they do not argue that this law violates voters’ substantive-due-process or equal-protection rights. Their claim thus does not resemble the voting-rights claims from Crawford or the third group of cases.

The Plaintiffs instead raise a garden-variety freedom-of-speech claim. They say that paragraph (c)(3) restricts either pure “political expression” or “expressive conduct” by prohibiting them from handing out the official application forms to encourage absentee voting. These two theories implicate the well-developed speech framework from cases like Meyer, Johnson, Rumsfeld, and O’Brien. Neither theory triggers Anderson-Burdick balancing.

On the one hand, if the Plaintiffs were correct (contrary to what we have already held) that paragraph (c)(3) prohibits speech promoting absentee voting, no amount of “ad hoc balancing of relative social costs and benefits” could save such a content-based ban. United States v. Stevens, 559 U.S. 460, 470 (2010); see Brown v. Ent. Merchs. Ass’n, 564 U.S. 786, 792–93 (2011). Indeed, the Supreme Court has described that type of balancing test as “startling No. 22-5028 Lichtenstein v. Hargett Page 21

and dangerous” in the speech context. Stevens, 559 U.S. at 470. The “American people” already engaged in this balancing, deciding that the benefits of speech outweigh its costs by adopting the First and Fourteenth Amendments. Brown, 564 U.S. at 792 (quoting Stevens, 559 U.S. at 470). And courts have no authority to rebalance their choice. See id. So even if a ban on “pure speech” imposed minuscule burdens, we would subject it to strict scrutiny. McIntyre, 514 U.S. at 345–46, 346 n.10.

On the other hand, if we are correct that paragraph (c)(3) bars only the conduct of distributing a form, it would trigger the Supreme Court’s “relatively lenient” expressive-conduct test. Johnson, 491 U.S. at 407. Through this test, the Court has likewise already engaged in the relevant “balancing” by holding that the government may generally ban conduct even if it “incidentally burdens speech[.]” Albertini, 472 U.S. at 688. When a ban does not target conduct because of its message, this deferential review applies because the ban leaves open “ample alternative channels” to convey a message—by actually speaking. Clark, 468 U.S. at 293. And courts may not engage in a Lochner-style evaluation of the “wisdom” of the conduct ban under the guise of the First Amendment. Pap’s A.M., 529 U.S. at 301 (plurality opinion); cf. Lochner v. New York, 198 U.S. 45, 64 (1905). In short, either Meyer’s strict-scrutiny test or O’Brien’s expressive-conduct test applies to the Plaintiffs’ claim. The First Amendment leaves no room for balancing.

3. The Plaintiffs argue instead that we now use Anderson-Burdick to evaluate all “election law” challenges—whether the challenger raises a free-speech claim, a substantive-due- process claim, an equal-protection claim, or any other claim. Appellants’ Br. 36 (citation omitted). This view conflicts with both text and precedent. The Supreme Court has noted that all interpretation (including constitutional interpretation) begins with the text of the document. See Gamble v. United States, 139 S. Ct. 1960, 1965 (2019); TransUnion LLC v. Ramirez, 141 S. Ct. 2190, 2203 (2021). And the Constitution contains no universal “cost-benefit balancing” provision. Cf. Tiwari v. Friedlander, 26 F.4th 355, 365–66 (6th Cir. 2022). It contains many different provisions that have many different meanings and that are relevant to election laws in many different ways. So why would we automatically interpret an amendment that bars the government from “abridging the freedom of speech” to mean the exact same thing No. 22-5028 Lichtenstein v. Hargett Page 22

(and require the exact same scrutiny) as one that bars the government from, say, “deny[ing]” “the equal protection of the laws”?

Unsurprisingly, then, the Supreme Court has not applied the Anderson-Burdick test to all First Amendment claims in the election or voting context. For example, it has applied strict scrutiny—not Anderson-Burdick balancing—to many election laws, including those banning anonymous campaign leaflets and prohibiting judicial candidates from airing their views on legal topics. See McIntyre, 514 U.S. at 344–46; Republican Party of Minn. v. White, 536 U.S. 765, 774–75 (2002). It has also applied “public forum” analysis—not Anderson-Burdick balancing— to laws restricting speech in and around polling places. See Minn. Voters All. v. Mansky, 138 S. Ct. 1876, 1885–86 (2018); Burson v. Freeman, 504 U.S. 191, 196–98 (1992) (plurality opinion); id. at 215–16 (Scalia, J., concurring in the judgment). And it has applied a test just below strict scrutiny (what the Court now calls “exacting scrutiny”) to disclosure requirements for “election-related” speech. Ams. for Prosperity Found., 141 S. Ct. at 2383 (plurality opinion). Lastly, in an analogous setting, the Court has held that a state’s conflict-of-interest rules barring legislators from voting on bills do not violate the First Amendment because they prohibit conduct rather than expression. Carrigan, 564 U.S. at 125–28. It did not engage in “balancing” to reach this result.

We also have not applied the Anderson-Burdick test to all election-related challenges. For example, we chose rational-basis review over Anderson-Burdick balancing when evaluating a state constitutional provision imposing terms limits on state legislators. See Kowall v. Benson, 18 F.4th 542, 546–49 (6th Cir. 2021). Like the Supreme Court, we have also applied “exacting scrutiny” to disclosure requirements in the election setting. Libertarian Party of Ohio v. Husted, 751 F.3d 403, 413–14 (6th Cir. 2014). And we have left open whether Anderson-Burdick balancing should apply to other election-related challenges, such as those asserting procedural- due-process claims. See Memphis A. Philip Randolph Inst., 978 F.3d at 390–91; cf. Daunt, 999 F.3d at 314.

In sum, the Plaintiffs are wrong to suggest that the Anderson-Burdick balancing test applies to all election-related challenges. It does not. And this test does not apply to the Plaintiffs’ free-speech claim here; rather, traditional speech rules govern that claim. No. 22-5028 Lichtenstein v. Hargett Page 23

C. Does Tennessee’s Ban Satisfy the Expressive-Conduct Test?

The Plaintiffs thus have one final free-speech path to allege a viable claim: the Supreme Court’s expressive-conduct test. This test requires us to ask two more questions. Do the Plaintiffs seek to engage in “inherently expressive” conduct? Rumsfeld, 547 U.S. at 66. And if so, does the State have a “substantial” reason for banning the distribution of application forms “that would be achieved less effectively absent the” ban? Id. at 67 (quoting Albertini, 472 U.S. at 689).

1. Do the Plaintiffs Intend to Undertake “Inherently” Expressive Conduct?

Although O’Brien held that the First Amendment applies to some conduct in some settings, the Supreme Court has limited the conduct eligible for this protection. The First Amendment protects only what the Court has called “inherently expressive” conduct. Id. at 66. To qualify as inherently expressive, an action must possess two traits. See Johnson, 491 U.S. at 404. The actor must intend to express a “particularized message” by engaging in the action. Blau v. Fort Thomas Pub. Sch. Dist., 401 F.3d 381, 388 (6th Cir. 2005) (quoting Spence v. Washington, 418 U.S. 405, 411 (1974) (per curiam)). And a high “likelihood” must exist that the audience who sees the action will understand its message. Id. (quoting Spence, 418 U.S. at 411).

The first element—that the speaker intends to convey a particularized message—does not pose a high bar. See Condon v. Wolfe, 310 F. App’x 807, 819 (6th Cir. 2009). Even a parade that includes groups with “all sorts of messages” triggers this speech protection. Hurley v. Irish-Am. Gay, Lesbian & Bisexual Grp. of Boston, 515 U.S. 557, 569 (1995); cf. Castorina ex rel. Rewt v. Madison Cnty. Sch. Bd., 246 F.3d 536, 539–40 (6th Cir. 2001). Yet the bar does exist. So a student’s challenge to a school dress code flunked this requirement because she did not intend to express any message by wearing clothes that the dress code prohibited. Blau, 401 F.3d at 389.

The second element—that the audience will likely understand the message—has more bite. That is because a viewer must be able to understand the message from the conduct alone without any accompanying speech explaining the reasons behind it. Rumsfeld, 547 U.S. at 66. No. 22-5028 Lichtenstein v. Hargett Page 24

When a party must include “explanatory speech” for the audience to get the message, the conduct does not warrant protection. Id. So a party who refuses to pay taxes cannot invoke the First Amendment merely by proclaiming disdain for the tax laws when committing the crime. See id. Likewise, the Court in Rumsfeld held that the First Amendment did not protect a law school’s decision to bar military recruiters from campus as an act of protest against the military’s limits on gay and lesbian servicemembers. See id. Without speech explaining this restriction, the Court reasoned, nobody would understand that the school meant to convey disapproval of the military’s policy. See id.

Yet other cases do not permit us to take Rumsfeld’s principle too far. The Court has also held that we must place a party’s conduct in the “context” in which the party engaged in it. Johnson, 491 U.S. at 405. Federal law, for example, treats burning a flag as a “dignified way” to dispose of it. 4 U.S.C. § 8(k). Other than by considering the context, how else can one tell whether a flag burner means to disparage the flag or to respect it? When finding flag burning expressive, then, the Court in Johnson acknowledged that it occurred during “a political demonstration” against President Reagan’s policies. 491 U.S. at 405–06; see also Spence, 418 U.S. at 410–11.

How do these elements play out here? In a thoughtful opinion, the district court held that the Plaintiffs’ complaint failed to plausibly allege that their proposed conduct was inherently expressive. See Lichtenstein, 2021 WL 5826246, at *4–6 (relying on Lichtenstein, 489 F. Supp. 3d at 764–74). The court recognized that the Plaintiffs met the first element because their complaint asserted that they subjectively intended to express a message encouraging absentee voting by handing out the application forms. See Lichtenstein, 489 F. Supp. 3d at 767. But it then held that nobody who saw only this conduct would understand the message. See id. at 767–69.

We are not as confident in this conclusion because of tension in the Court’s cases. When deciding whether an audience member would understand an action’s message, the Court has told us both that we cannot consider speech that “accompanies” the action, Rumsfeld, 547 U.S. at 66, and that we must consider the “context” in which the party engages in it, Johnson, 491 U.S. at 405. But the context will often include speech, such as the “political slogans” that the No. 22-5028 Lichtenstein v. Hargett Page 25

Court accounted for in Johnson. Id. at 399. So what divides the prohibited use of “explanatory speech,” Rumsfeld, 547 U.S. at 66, from the required use of contextual speech, Johnson, 491 U.S. at 405?

This case provides a good example of the dilemma. Plaintiffs allege that they intend to distribute application forms during get-out-the-vote drives. Compl., R.1, PageID 5. These drives will have plenty of speech encouraging absentee voting (whether on banners or in oral advocacy). Can we consider this speech in the analysis? Perhaps we should narrowly read Rumsfeld to bar the use of only “explanatory” speech—that is, speech that literally explains the conduct. 547 U.S. at 66. The Plaintiffs nowhere allege that they intend to engage in such speech—for example, “I am handing you this form to express that you should vote absentee.” Rather, they allege that they intend more generally to discuss the “benefits of voting by mail[.]” Compl., R.1, PageID 9. One could view this surrounding speech at a get-out-the-vote drive as analogous to the surrounding speech at a “political demonstration[.]” Johnson, 491 U.S. at 406. And even if the distribution of a form conveys no message in the abstract, it might convey one in the context of these drives.

If, by contrast, we broadly read Rumsfeld’s ban on the use of “speech that accompanies” the conduct, perhaps we must ignore all the speech during the drives. 547 U.S. at 66. That view would leave the Plaintiffs with nothing to clarify what they intend to convey by the distribution of the form. And unlike flying a flag, burning a flag, or wearing an armband, handing out an official document is not the type of thing that someone typically does “as a form of symbolism[.]” Spence, 418 U.S. at 410; see Johnson, 491 U.S. at 404–05; Tinker v. Des Moines Indep. Cmty. Sch. Dist., 393 U.S. 503, 505 (1969). Even if it “discloses” that the Plaintiffs support absentee voting, it is not clear handing out a form “symbolizes” that support. Carrigan, 564 U.S. at 126. So it is unlikely that viewers would get the message from that conduct alone. In the end, though, we opt not to settle these legal questions. Even if the Plaintiffs’ conduct is “inherently expressive,” Tennessee’s ban survives the expressive-conduct test. No. 22-5028 Lichtenstein v. Hargett Page 26

2. Does Tennessee’s Ban Survive the Scrutiny Governing Expressive Conduct?

The First Amendment does not protect inherently expressive conduct with nearly the same rigor that it protects actual expression. As long as a state has not banned the conduct to suppress the message behind it (we have already found Tennessee has not here), the Supreme Court allows the government to bar the conduct if it can satisfy three other elements. O’Brien, 391 U.S. at 377. The government must have the “constitutional power” to enact the ban. Id. The ban must serve “an important or substantial governmental interest[.]” Id. And the limits on the expressive part of the conduct must be “no greater than is essential to the furtherance of that interest.” Id.

a. The Supreme Court has sometimes called its expressive-conduct test “intermediate” scrutiny. Nixon v. Shrink Mo. Gov’t PAC, 528 U.S. 377, 386 (2000). Since O’Brien, however, it has applied a test that falls much closer to the deferential rational-basis test than the demanding strict-scrutiny test. That conclusion follows from the ways the Court evaluates both whether the government has a “substantial interest” for a conduct ban and whether an adequate means-ends “fit” exists between the ban and this interest.

Substantial Interest. O’Brien noted that the government must have “important or substantial” interests for a regulation of conduct. 391 U.S. at 377. This requirement is not demanding. Consider how a court may identify the “interests” underlying a conduct regulation. A legislature need not identify this “governmental interest” in the law itself. Barnes, 501 U.S. at 567–68 (plurality opinion). Rather, a court may discern the justifications for a regulation by examining its “text and history” alone. Id. at 568 (plurality opinion); id. at 582–83 (Souter, J., concurring in the judgment). And as long as these court-identified reasons are substantial, they will suffice even if other evidence suggests that the government adopted the ban for an “illicit” reason. Pap’s A.M., 529 U.S. at 292 (plurality opinion). So the ban on destroying draft cards in O’Brien did not violate the First Amendment despite statements from some legislators implying that they voted for it to suppress dissident views. See 391 U.S. at 384–86. In these respects, the expressive-conduct test resembles rational-basis review (which also does not require the court- identified interests to “actually motivate[]” a law) more than strict scrutiny or other types of intermediate scrutiny (which bar a court from relying on “hypothesized” interests). Compare No. 22-5028 Lichtenstein v. Hargett Page 27

FCC v. Beach Commc’ns, Inc., 508 U.S. 307, 315 (1993), with Kennedy v. Bremerton Sch. Dist., 142 S. Ct. 2407, 2432 n.8 (2022), and United States v. Virginia, 518 U.S. 515, 533 (1996).

Or consider how the government can prove the importance of its interests. It need not make an “evidentiary showing” that the harm caused by the prohibited conduct is “real.” Pap’s A.M., 529 U.S. at 299 (plurality opinion) (discussing O’Brien, 391 U.S. at 378–80). Nor must it prove that the specific conduct of a “particular” challenger will cause this harm. Albertini, 472 U.S. at 688. It instead may rely on highly general interests that are obviously important, such as the “interest in raising and supporting the Armed Forces,” Rumsfeld, 547 U.S. at 67, in “conserving park property,” Clark, 468 U.S. at 299, or in “deter[ring] crime,” Pap’s A.M., 529 U.S. at 293 (plurality opinion). And it may rely on prior cases that have already treated the cited interests as important. See id. at 296–97 (plurality opinion). In these respects, too, the expressive-conduct test looks more like rational-basis review (which allows the government to invoke general interests without “courtroom fact-finding”) than strict scrutiny (which requires the government to prove in court the importance of its interest as applied to the specific challenger). Compare Beach Commc’ns, 508 U.S. at 315, with Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1881 (2021).

Required Fit. O’Brien also noted that a conduct ban’s “incidental restriction on alleged First Amendment freedoms [must be] no greater than is essential to the furtherance of that interest[.]” 391 U.S. at 377. Some of the Supreme Court’s cases (but not others) have described this part of the expressive-conduct test as a “narrow tailoring” element. Compare Bd. of Trs. of State Univ. of N.Y. v. Fox, 492 U.S. 469, 478 & n.3 (1989), with Rumsfeld, 547 U.S. at 67–68. That said, the Court has used the phrase “narrow tailoring” to describe several different “means- ends” tests. And the more demanding “narrow tailoring” that it requires for speech restrictions looks nothing like the lenient “narrow tailoring” (if any) that it requires for conduct bans.

The Court, for example, has held that the most rigorous test (strict scrutiny) requires a speech limit to be “narrowly tailored” to a compelling interest. Citizens United, 558 U.S. at 340 (citation omitted). To satisfy strict scrutiny’s version of narrow tailoring, the government must adopt the “least” speech “restrictive means” to achieve its interest. McCullen, 573 U.S. at 478. No. 22-5028 Lichtenstein v. Hargett Page 28

If the government could serve its interests through an alternative that limited less speech, it must pick “that alternative.” United States v. Playboy Ent. Grp., Inc., 529 U.S. 803, 813 (2000).

It is, by contrast, well-established that a ban on conduct will pass muster even if a court can identify “alternative methods” that would serve the government interest equally well while imposing less burdens on the expressive component of the conduct. Rumsfeld, 547 U.S. at 67; see Albertini, 472 U.S. at 689. That these “alternative methods” exist is, in the Court’s words, “beside the point.” Rumsfeld, 547 U.S. at 67. When a dispute involves conduct regulations, the First Amendment leaves the choice among alternatives to legislators, not courts. Id.

Apart from strict scrutiny, the Court has also held that a government must engage in narrow tailoring for content-neutral restrictions on the “time, place, and manner” of engaging in speech. See McCullen, 573 U.S. at 486; Ward v. Rock Against Racism, 491 U.S. 781, 799–800 (1989). And it has held that the “exacting scrutiny” test applicable to government-compelled disclosures requires similar “narrow tailoring.” Ams. for Prosperity Found., 141 S. Ct. at 2383– 84 (citation omitted). According to the Court, these versions of narrow tailoring require a “close fit” between the speech restriction and the government interest. McCullen, 573 U.S. at 486. That is, a regulation of speech cannot restrict “substantially more speech than is necessary to further the government’s legitimate interests.” Id. (quoting Ward, 491 U.S. at 799); Turner Broad. Sys., Inc. v. FCC, 512 U.S. 622, 662 (1994). And the regulation must “leave open ample alternative channels” for speaking. McCullen, 573 U.S. at 477 (quoting Ward, 491 U.S. at 791).

Although the Court has sometimes all but equated its test for conduct regulations and its test for laws restricting the time, place, and manner of speech, see Clark, 468 U.S. at 298–99, it has in practice imposed less rigorous “tailoring” requirements for the former regulations. Unlike cases involving speech, see Turner, 512 U.S. at 636–37, cases involving conduct have not asked whether a regulation covered “substantially more speech than is necessary to further the government’s legitimate interests,” id. at 662 (citation omitted); see Rumsfeld, 547 U.S. at 67. Also unlike cases involving speech, see Ward, 491 U.S. at 791, cases involving conduct have not asked whether the regulation left “open ample alternative channels of communication,” id. at 802; see Rumsfeld, 547 U.S. at 67. And although O’Brien seemingly suggested that any incidental restriction on First Amendment freedoms must be “no greater than is essential to the No. 22-5028 Lichtenstein v. Hargett Page 29

furtherance” of a substantial governmental interest, 391 U.S. at 377, the Court has since made clear that this test is not nearly as difficult to meet as it sounds. Instead, the Court has held that an adequate relationship exists between the means (the conduct ban) and the ends (the government interest) as long as the interest “would be achieved less effectively” without the ban. Rumsfeld, 547 U.S. at 67 (quoting Albertini, 472 U.S. at 689). That is, the ban need only “further” the interest. Pap’s A.M., 529 U.S. at 301 (plurality opinion); see also Arcara v. Cloud Books, Inc., 478 U.S. 697, 706–07, 707 n.4 (1986). Because, for example, the mandate that law schools allow military recruiters on campus would “add to the effectiveness of military recruitment,” the Court in Rumsfeld held that it satisfied this test without requiring more. 547 U.S. at 67.

An example further proves that the Court treats content-neutral speech regulations differently from conduct regulations. The Court has held that the government may not rely on administrative “convenience” or “efficiency” to adopt a content-neutral speech limit that sweeps up more speech than is required to achieve the restriction’s underlying objective. McCullen, 573 U.S. at 486, 495; cf. Ams. for Prosperity Found., 141 S. Ct. at 2387. But the Court has held the opposite for conduct bans. In that context, the government may adopt a “uniform” prohibition on “administrative efficiency” grounds even if the prohibition is overinclusive because it bars some expressive conduct that would “cause no serious damage” to the government’s underlying objective. Superior Ct. Trial Lawyers Ass’n, 493 U.S. at 430; see O’Brien, 391 U.S. at 378–79.

In the end, it should come as no surprise that the Court provides a softer First Amendment touch to limits on conduct rather than speech. Conduct restrictions will never “burden substantially more speech” than necessary. McCullen, 573 U.S. at 486 (citation omitted). By definition, they burden no actual speech. And these restrictions will always leave open ample avenues for speaking. See id. at 477. By definition, they leave open every avenue for actual speech.

Pap’s A.M. shows the lenient nature of both parts of this test. That case addressed a city’s public-nudity ban. 529 U.S. at 283–84. The ban required nude dancers at adult- entertainment venues to “wear, at a minimum ‘pasties’ and a ‘G-string.’” Id. at 284. The controlling plurality opinion suggested that “nude dancing” was inherently expressive. Id. at 289 No. 22-5028 Lichtenstein v. Hargett Page 30

(plurality opinion). But it then held that the ban satisfied the expressive-conduct test. It reasoned that the city sought to combat the “secondary effects”—namely, the increased crime— that these venues caused. Id. at 290. The plurality found this interest substantial without evidence that crime had increased near the city’s adult-entertainment venues. Id. at 296. It instead allowed the city to “rely” on the secondary effects that the Court had identified in other cases. Id. at 296–97. But how would the requirement that dancers wear barely any attire reduce crime? Some Justices ridiculed this claim as a “titanic surrender to the implausible.” Id. at 323–24 (Stevens, J., dissenting). Yet the plurality stated, without evidence, that the ban would “further” this crime-reduction interest (if not by much). Id. at 301 (plurality opinion). And it reiterated that the expressive-conduct test did not permit it to assess the “wisdom” of the city’s “chosen” method to serve that interest. Id.

b. Even at the pleading stage, we hold that paragraph (c)(3)’s ban on distributing the official absentee-ballot application forms also satisfies this test. Nobody disputes that Tennessee had the “constitutional power” to enact this ban. O’Brien, 391 U.S. at 377. The Constitution gives the states the authority to “prescribe[]” the “Times, Places and Manner of holding” federal elections. U.S. Const. art. I, § 4, cl. 1. And the states have equally broad authority to regulate their own elections. See Clingman, 544 U.S. at 586; Burdick, 504 U.S. at 433. The states thus retain the general power to identify the voters who may vote absentee, to regulate the process they must follow to vote this way, and to criminalize violations of election-related rules. See, e.g., Meyer, 486 U.S. at 427–28; McDonald v. Bd. of Election Comm’rs of Chi., 394 U.S. 802, 807–09 (1969); Mays, 951 F.3d at 792; Ohio Democratic Party, 834 F.3d at 626.

Paragraph (c)(3) also meets the expressive-conduct test’s substantial-interest and fit elements based on the same considerations that led the district court to find it met the Anderson-Burdick balancing test. See Lichtenstein, 2021 WL 5826246, at *7 (relying on Lichtenstein, 489 F. Supp. 3d at 780–86). To begin with, merely by examining paragraph (c)(3)’s “text” alone, we may reasonably identify at least one interest that the government seeks to serve: minimizing voter confusion. Barnes, 501 U.S. at 567 (plurality opinion). Tennessee strictly limits absentee voting by mail. Tenn. Code Ann. § 2-6-201. So only about 2.5% of Tennessee voters have historically voted this way. See Memphis A. Philip Randolph Inst., No. 22-5028 Lichtenstein v. Hargett Page 31

978 F.3d at 382; cf. Compl., R.1, PageID 8–9. If paragraph (c)(3)’s ban did not exist, however, “various groups” could send the official application form to all voters—even though only a fraction may vote absentee. Memphis A. Philip Randolph Inst., 478 F. Supp. 3d at 704. These types of mass mailings could cause mass confusion. That is especially true if a group’s commentary accompanying the official form is misleading or if the group has prefilled the part of the form listing the reason that the voter seeks to vote absentee. See Lichtenstein, 489 F. Supp. 3d at 783. On receiving this form, some voters might wrongly conclude that they are eligible to vote absentee. Others might wrongly believe that they must fill it out to vote at all. Cf. Memphis A. Philip Randolph Inst., 478 F. Supp. 3d at 704–05. Yet if these voters show up at the polls, they must cast a provisional ballot. Tenn. Code Ann. § 2-7-112(a)(3)(A). And if voters receive this form from multiple parties, they may submit several versions—increasing burdens on election officials. See Lichtenstein, 489 F. Supp. 3d at 783.

Under the expressive-conduct test, Tennessee need not introduce evidence to prove that its general concern with avoiding voter confusion qualifies as a substantial interest. Cf. Pap’s A.M., 529 U.S. at 299–300 (plurality opinion). In other contexts, the Supreme Court has described this interest as “important”—indeed, “compelling.” Munro v. Socialist Workers Party, 479 U.S. 189, 193–95 (1986) (citation omitted); Burson, 504 U.S. at 199 (plurality opinion). The Court, for example, has allowed states to rely on this rationale to limit the number of candidates on the ballot. See Munro, 479 U.S. at 194–95; Jenness v. Fortson, 403 U.S. 431, 442 (1971). It has relied on this rationale to justify a speech ban around the polling place. See Burson, 504 U.S. at 199–211 (plurality opinion). And it has relied on this rationale to discourage last-minute election changes. See Purcell v. Gonzalez, 549 U.S. 1, 4–5 (2006) (per curiam).

Next, paragraph (c)(3) would “further” (in the words of Pap’s A.M.) or “add to” (in the words of Rumsfeld) the goal of reducing confusion about whether voters may vote absentee. Pap’s A.M., 529 U.S. at 301 (plurality opinion); Rumsfeld, 547 U.S. at 67. If nobody can distribute the official applications to voters, no voters will get confused about their eligibility to vote absentee when they receive an unsolicited application. And no voters will get confused over whether they must fill out this form to vote at the polls. In short, because Tennessee would “be more exposed to [voter confusion] without [paragraph (c)(3)] than with it, the ban is safe No. 22-5028 Lichtenstein v. Hargett Page 32

from invalidation under the First Amendment[.]” Clark, 468 U.S. at 297. This means-ends relationship is much more “evident” to us than, say, the plurality’s conclusion in Pap’s A.M. that crime will go down near adult-entertainment venues if dancers wear next to nothing. 529 U.S. at 300–01 (plurality opinion). And this relationship is all that is required. See Rumsfeld, 547 U.S. at 67.

In response, the Plaintiffs do not dispute that Tennessee generally serves an important interest when it seeks to reduce voter confusion. Appellants’ Br. 40. But they argue that paragraph (c)(3) does so in a blunderbuss way by banning far more conduct than necessary to achieve this goal. They note, for example, that the ban applies even when eligible voters request the application form from them. Reply Br. 18. Why not simply bar the unsolicited distribution of the form? Or why not bar only the distribution of the form to ineligible voters? Or why not just bar mass mailings? All fair points. And these arguments might have merit if we had to subject paragraph (c)(3) to anything like strict scrutiny. But these “alternative methods . . . are beside the point” here. Rumsfeld, 547 U.S. at 67. Under the expressive-conduct test, “[i]t suffices that the means chosen by [Tennessee] add to the effectiveness of” Tennessee’s efforts to avoid voter confusion. Id. And we cannot upend the State’s choice about “the most appropriate method” to prevent this confusion simply because we might have advocated for a narrower ban if we sat in the Tennessee legislature. Albertini, 472 U.S. at 689. Nor can we upend the State’s choice simply because the Plaintiffs allege that they will not engage in any conduct that would cause confusion. See id. at 688–89. Even if making some exceptions would not significantly undermine Tennessee’s interests, the State can use a “uniform rule” to serve “administrative efficiency.” Superior Ct. Trial Lawyers Ass’n, 493 U.S. at 430. And for what it is worth, the Plaintiffs’ argument is also partially overbroad. After all, the State permits them to hand out a privately made “request” for an absentee-ballot application (which can serve as an application itself) if voters ask for one. See Tenn. Code Ann. § 2-6-202(a)(3), (c)(4).

The test may be deferential, the Plaintiffs counter, but we still cannot find it met at the pleading stage. Appellants’ Br. 13, 41 & n.17. At this stage, they remind us, we must accept their well-pleaded factual allegations as true and make all reasonable inferences in their favor. See Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). They raise another fair point. But we disagree No. 22-5028 Lichtenstein v. Hargett Page 33

with them given the nondemanding nature of the expressive-conduct test. Consider two analogies. When rational-basis review applies, it is downright routine for courts to dismiss constitutional challenges at the pleading stage. See, e.g., Gregory v. Ashcroft, 501 U.S. 452, 456, 470–73 (1991); Sanchez v. Off. of State Superintendent of Educ., 45 F.4th 388, 395–400 (D.C. Cir. 2022); Lyda v. City of Detroit (In re City of Detroit), 841 F.3d 684, 701–02 (6th Cir. 2016). Likewise, “we have not shied away from” granting motions to dismiss under the Anderson- Burdick test even though it can require a fact-intensive balance of a law’s burdens on voting rights against the state’s precise justifications for the law. Daunt, 999 F.3d at 312–13 (collecting cases).

The same result should apply here. We “fail to see how further factual development could change” our conclusion that paragraph (c)(3) furthers important state interests—all that is required. See id. at 313. The State need not make an “evidentiary showing” proving a sufficiently high risk of voter confusion. Pap’s A.M., 529 U.S. at 299 (plurality opinion). Nor must it prove that paragraph (c)(3) sufficiently furthers this interest in avoiding confusion, see id. at 300–01, or that the State could not achieve it through “alternative” means, Rumsfeld, 547 U.S. at 67. Indeed, the Plaintiffs direct their request for factual development primarily to the district court’s separate holding that they failed to plead that they seek to engage in inherently expressive conduct. Even if they could meet that element, however, their claim would still fail because paragraph (c)(3)’s nearly non-existent intrusion on their expression survives the expressive- conduct test.

III. Does Tennessee’s Ban Infringe the Plaintiffs’ Freedom of Association?

This conclusion leaves the Plaintiffs’ alternative theory that “exacting scrutiny” should apply because paragraph (c)(3) infringes their freedom of association. The Supreme Court has read the First Amendment to impliedly protect a right to associate with others for the purpose of exercising the “freedom of speech” that the amendment expressly protects. See Roberts v. U.S. Jaycees, 468 U.S. 609, 622 (1984). The Plaintiffs assert that the Tennessee law violates this right because it interferes with their ability to “engage and associate with their” members and with other voters “to advance their core belief in civic participation.” Appellants’ Br. 32. But No. 22-5028 Lichtenstein v. Hargett Page 34

this amorphous theory interprets the right of association more broadly than the Supreme Court’s cases permit.

The Court protects the free-speech right of expressive association (as compared to the substantive-due-process right of intimate association) because collaboration can help people spread a message. See Rumsfeld, 547 U.S. at 68; cf. Roberts, 468 U.S. at 618–22. Just as money or ink can be speech “inputs” that create protected expression, so too people can form groups to better convey a shared point of view. See Rumsfeld, 547 U.S. at 68. And a law regulating group activity can harm the group’s ability to produce its desired message—thereby abridging the speech of the group and its members. See Boy Scouts of Am. v. Dale, 530 U.S. 640, 648 (2000).

The Supreme Court’s precedent has led us to consider three things when confronted with an expressive-association claim. See Hamilton Cnty. Educ. Ass’n v. Hamilton Cnty. Bd. of Educ., 822 F.3d 831, 840 (6th Cir. 2016); Miller v. City of Cincinnati, 622 F.3d 524, 538 (6th Cir. 2010). First: Does the First Amendment apply to the group because it engages in speech? See Dale, 530 U.S. at 648. Second: If the group engages in speech, does a law “significantly burden” its ability to express its message? See id. at 653; Roberts, 468 U.S. at 622–23. Third: If this burden exists, can the government satisfy the scrutiny that applies? For some burdens, the Court has applied strict scrutiny. See Dale, 530 U.S. at 648; Roberts, 468 U.S. at 623. For others, it has applied its less rigorous “exacting scrutiny” test. Ams. for Prosperity Found., 141 S. Ct. at 2383 (plurality opinion) (quoting Buckley, 424 U.S. at 64); cf. id. at 2391–92 (Alito, J., concurring in part and concurring in the judgment). And for the ballot-access burdens that we have discussed, it has applied the Anderson-Burdick balancing test. See Wash. State Grange, 552 U.S. at 451–52.

The Plaintiffs’ complaint in this case alleges facts that easily make it past our first step. The Supreme Court’s expressive-association cases apply most obviously to “[p]olitical advocacy groups” whose raison d’être is speaking. Miller, 622 F.3d at 538. And here, most of the Plaintiffs allege that they qualify as these sorts of groups that exist to engage in political expression. See id. No. 22-5028 Lichtenstein v. Hargett Page 35

But the Plaintiffs cannot make it past step two, which asks whether paragraph (c)(3) “significantly burden[s]” their ability to associate with others to express a message. Dale, 530 U.S. at 653. A government can burden the right to associate in a “number” of ways. Ams. for Prosperity Found., 141 S. Ct. at 2382 (quoting Roberts, 468 U.S. at 622). It might compel a group (say, the Boy Scouts) to accept members with whom the group does not want to associate (say, openly gay individuals) because the group believes that this membership will dilute its message (say, disapproval of non-heterosexual conduct). See Dale, 530 U.S. at 653. The government also might compel a group (say, the NAACP) to disclose its list of members to a hostile audience (say, the segregated south) and so deter individuals from joining the group out of fear of harassment. See NAACP v. Alabama ex rel. Patterson, 357 U.S. 449, 461–63 (1958). Or the government might discriminate against a group (say, the Students for a Democratic Society) by refusing to give it a generally available benefit (say, access to a college’s facilities) that groups generally use to air their points of view. See Healy v. James, 408 U.S. 169, 181 (1972). Or it might prohibit a group (say, the NAACP) from soliciting individuals (prospective clients) to associate with the group’s lawyers for litigation purposes. See NAACP v. Button, 371 U.S. 415, 429–37 (1963); see also NAACP v. Claiborne Hardware Co., 458 U.S. 886, 911–12, 924–26 (1982).

Here, however, we do not see how paragraph (c)(3) affects the Plaintiffs’ ability to associate with others in order to create or convey a message, even after accepting their complaint’s allegations as true. Nothing in paragraph (c)(3) requires the Plaintiffs to accept certain members who would contradict their message. See Rumsfeld, 547 U.S. at 69–70; cf. Dale, 530 U.S. at 653. Nothing bars them from associating with anyone. See Button, 371 U.S. at 429–37. Nothing threatens legal punishment or practical harm for those who decide to join the Plaintiffs. Cf. Elrod v. Burns, 427 U.S. 347, 355–56 (1976) (plurality opinion); Patterson, 357 U.S. at 462. So nothing discourages individuals from becoming members. And nothing denies the Plaintiffs or their members a benefit that Tennessee broadly makes available to others. Cf. Healy, 408 U.S. at 181–82. To the contrary, the law neutrally applies to all groups and individuals. The Plaintiffs’ complaint thus asserts no facts that plausibly allege that this law “directly or indirectly” affects their “group membership” in a way that undermines their message. Miller, 622 F.3d at 538. No. 22-5028 Lichtenstein v. Hargett Page 36

In response, the Plaintiffs argue that they “associate” with members in the community through their “civic engagement activity” and that paragraph (c)(3) limits their ability to undertake this speech. Reply Br. 16. As far as we can tell, this argument merely repackages their claim that strict scrutiny should apply because the law interferes with expression. But that same alleged “interference” will occur whether an individual or group engages in the speech; the interference does not affect their ability to associate at all. Besides, this claim fails for the reasons that we have explained: nothing in paragraph (c)(3) affects their ability to create or convey speech.

The Plaintiffs also fault the district court for disregarding this distinct freedom-of- association claim when dismissing their complaint. But that court did not ignore this claim. It held that the Anderson-Burdick balancing test applied to the claim and rejected it on that ground. Lichtenstein, 489 F. Supp. 3d at 778; see Lichtenstein, 2021 WL 5826246, at *4, *6 & n.5, *7–8. Although this is not the type of associational claim to which we have applied Anderson-Burdick, we may affirm on alternative grounds. See McCormick v. Braverman, 451 F.3d 382, 396 (6th Cir. 2006); cf. Mays, 951 F.3d at 791–93. And, as a matter of law, the Plaintiffs have not adequately pleaded that paragraph (c)(3) burdens their right of expressive association in any way.

* * *

In the end, the Plaintiffs may well have articulated several good “policy” arguments about why Tennessee should reconsider paragraph (c)(3)’s scope now that its election officials have posted the official application forms online. Lichtenstein, 489 F. Supp. 3d at 786. Indeed, our colleague mentions several of those reasons in dissent. Dissenting Op., at 37–38 & n.2. But our job is not to decide whether the ban represents good or bad policy. That is the job of the Tennessee legislature. We may intervene to stop the enforcement of this democratically passed law only if it violates some federal standard, here the First Amendment. And under the deferential free-speech rules that, at most, apply to the Plaintiffs’ claims, the ban passes constitutional muster.

We affirm. No. 22-5028 Lichtenstein v. Hargett Page 37

DISSENT _________________

HELENE N. WHITE, Circuit Judge, dissenting. The majority upholds a Tennessee law that threatens to imprison persons who distribute publicly available absentee-ballot applications. See Tenn. Code. Ann. § 2-6-202(c)(3). Because Plaintiffs plausibly allege that the law burdens their “core political speech,” for which the First Amendment is “at its zenith,” I reject the majority’s reliance on the O’Brien test and find resolution of this case on the pleadings inappropriate. Buckley v. Am. Const. L. Found., 525 U.S. 182, 187 (1999) (quoting Meyer v. Grant, 486 U.S. 414, 422, 425 (1988)). I also disagree with the majority’s discussion of Anderson-Burdick and O’Brien and the majority’s rejection of Plaintiffs’ expressive-association claim. See Maj. Op. 16–22, 26–36. For these reasons, I respectfully dissent.

Like many states, Tennessee makes absentee-ballot applications available online for anyone to print and download. But under Tennessee law, “[a] person who is not an employee of an election commission commits a Class E felony if such person gives an application for an absentee ballot to any person.” Tenn. Code. Ann. § 2-6-202(c)(3). A Class E felony carries a prison sentence of one to six years and a fine up to $3,000. See id. § 40-35-111(b)(5). Thus, in Tennessee, a grandson risks years behind bars for encouraging his grandparents over age 60 to vote by mail and handing them publicly available forms.1 The same is true for a soldier sharing

1Recipients in this example and the following in-text examples are all eligible to vote absentee in Tennessee. See Tenn. Code Ann. §§ 2-6-201(5)(A), (C), -502(a). The “subset of voters” who may vote absentee in the state, Maj. Op. 3, is significant in number. In 2019, 1,577,807 Tennesseans were age 60 or older, comprising 23% of the population. See Population Counts by Age Group, Sex, Race and Ethnicity, Estimates 2019, https://www.tn.gov/content/dam/tn/health/documents/population/TN-Population-by- AgeGrp-Sex-Race-Ethnicity-2019.pdf (last visited Aug. 25, 2023). Added to this number are persons who are outside their counties of registration during in-person voting, see Tenn. Code Ann. § 2-6-201(1), (9), students who attend in-state higher-education institutions located outside their counties of registration and their spouses, see id. § 2-6-201(2), persons who are unable to vote in person due to sickness, hospitalization, or physical disability or who reside full-time in nursing homes or similar establishments located outside their counties of registration, see id. § 2-6-201(3), (5)(B)–(C), persons who are unable to vote in person due to jury service, see id. § 2-6-201(4), candidates for office, see id. § 2-6-201(6), election officials, see id. § 2-6-201(7), and persons who are unable to vote in person because of observance of religious holidays, see id. § 2-6-201(8). No. 22-5028 Lichtenstein v. Hargett Page 38

forms with other Tennesseans stationed overseas, or a neighbor delivering forms to those who cannot vote in person due to illness or disability.2

The same is also true for Plaintiffs, a labor organizer and several Tennessee-based civic and labor groups, all “committed to engaging and organizing Tennesseans around making their voices heard through voting.” R. 1, PID 8. These groups include the Memphis and West Tennessee ALF-CIO Central Labor Council (MCLC),3 the Tennessee State Conference of the NAACP,4 the Equity Alliance,5 the Memphis A. Philip Randolph Institute,6 and Free Hearts.7

2In this regard, I acknowledge that Plaintiffs do not challenge Tennessee’s law on due-process grounds; nor do they argue in their briefing that the law fails for overbreadth, see Wash. State Grange v. Wash. State Republican Party, 552 U.S. 442, 449 n.6 (2008), vagueness, see Nat’l Endowment for the Arts v. Finley, 524 U.S. 569, 588 (1998), or otherwise chilling protected speech that is not directly prohibited, see Laird v. Tatum, 408 U.S. 1, 11 (1972); and they do not argue the law violates their right to petition, as Amicus Cato Institute asserts. Br. of Amicus Cato Inst. 8–12. I thus do not address these arguments, notwithstanding any potential merit. 3“The Memphis and West Tennessee AFL-CIO Central Labor Council, also known as the Memphis Central Labor Council (‘MCLC’), is a Memphis, Tennessee-based union that acts as an umbrella organization for 41 affiliate unions based in western Tennessee. MCLC is dedicated to representing the interests of working people at the state and local levels by advocating for social and economic justice. In support of its advocacy agenda, MCLC routinely engages in voter outreach efforts, including of its approximately 20,000 members and their families, through voter identification, education, and mobilization drives.” R. 1, PID 3. 4“The Tennessee State Conference of the NAACP (‘Tennessee NAACP’) is a nonpartisan, multi-racial, non-profit membership organization headquartered in Jackson, Tennessee. Tennessee NAACP has three regional divisions—Eastern, Middle, and Western Tennessee—as well as the 33 local branch units and 22 college chapters and youth councils. Tennessee NAACP was founded in 1946 to serve as the Tennessee arm of the National Association for the Advancement of Colored People. Its mission is to eliminate race-based discrimination through securing political, educational, social, and economic equality rights and ensuring the health and well-being of all persons. Tennessee NAACP has more than 10,000 members across the State, primarily consisting of African Americans, other people of color, and allies. Tennessee NAACP and most local branch units are primarily volunteer-run, and all officers are volunteers.” R. 1, PID 3–4. 5“The Equity Alliance is a Nashville, Tennessee-based nonpartisan, non-profit organization that seeks to equip citizens with tools and strategies to engage in the civic process and empower them to take action on issues affecting their daily lives. The Equity Alliance is dedicated to expanding the electorate, educating communities of color about the political process, and engaging and empowering citizens to vote.” R. 1, PID 4. 6The “Memphis A. Philip Randolph Institute (‘APRI’) is a Memphis, Tennessee-based non-profit political advocacy and membership organization which works to strengthen ties between the labor movement and the community, increase the political impact of black voters, and implement structural changes through the American democratic process. In support of its advocacy and engagement efforts, APRI sponsors voter education and Get- Out-The-Vote programs in the community.” R. 1, PID 5. 7“Free Hearts is a nonpartisan, nonprofit organization based in Nashville, Tennessee. Since 2015, Free Hearts has educated, organized, advocated for, and supported the families of individuals impacted by the criminal punishment system. Free Hearts works across the State and in coalition with other nonprofit groups to support approximately 426 incarcerated and formerly incarcerated individuals and their families. It has 5 full-time staff and approximately 40 volunteers who engage with their community base on a regular basis. Since 2017, Free Hearts has also been working to expand the franchise to those impacted by the criminal justice system. To that end, Free No. 22-5028 Lichtenstein v. Hargett Page 39

They and Jeffrey Lichtenstein, who is Executive Secretary of the MCLC, seek to provide absentee-ballot applications to their members as part of get-out-the-vote campaigns, through which they educate and encourage voters to participate in the electoral process, including by voting absentee.

Plaintiffs brought this challenge before the 2020 Presidential Election and describe their intended activity in their complaint. “In election after election, Plaintiffs have run or participated in voter engagement programs involving voter registration activities, voter education, and voter turnout.” Id. at PID 8. They engage with a wide variety of voters, including union members and their families, see id. at PID 3, individuals and “families . . . impacted by the criminal punishment system,” id. at PID 5, “communities of color,” id. at PID 4, and “communities that have had historically low voter registration and turnout,” id.

Plaintiffs anticipate that “outreach to eligible absentee voters will play a central role in [their] voter engagement strategy,” noting that, “in the August 2020 election, more than 116,000 Tennessee voters cast absentee ballots, which [was] over five times more than had done so in the prior four August elections.” Id. at PID 8–9. Accordingly, “as a key part of [their] absentee voter engagement, Plaintiffs will, if permitted, provide potential absentee voters with the blank absentee ballot applications that are available online from the state and county election commissions.” Id. at PID 9. Distributing forms “is necessary,” they allege, “because Plaintiffs have found that their voter engagement efforts are significantly more effective when they are able to provide voters with all of the information and requisite forms they might need.” Id. Plaintiffs’ activity “will necessarily include discussing with voters the benefits of voting by mail, reminding eligible absentee voters about application and ballot submission deadlines and requirements, and following up with voters to ensure their ballots were received, cast and counted.” Id.

Hearts registers eligible individuals in jails, advocates for jails to become polling sites, helps restore voting rights to formerly incarcerated individuals, and advocates for reforms around automatic voter registration in Tennessee.” R. 1, PID 5–6. No. 22-5028 Lichtenstein v. Hargett Page 40

II.

Based on the pleadings, I would subject Tennessee’s law to the “exacting scrutiny” that the Supreme Court applies to burdensome restrictions on sharing other kinds of political documents, such as petitions and pamphlets. Like these protected activities, distributing absentee-ballot applications necessarily involves “core political speech.” Meyer, 486 U.S. at 420; see McIntyre v. Ohio Elections Comm’n, 514 U.S. 334, 344–47 (1995).

The Supreme Court has employed “exacting scrutiny,” or a similarly demanding standard, when evaluating restrictions on activities that involve sharing political documents, recognizing that these activities necessarily involve “core political speech,” and thus that restrictions on these activities restrict “core political speech.” Meyer, 486 U.S. at 420–22.8 Under this standard, a law is only sustainable “if it is narrowly tailored to serve an overriding state interest.” McIntyre, 514 U.S. at 347.

1.

The Court’s unanimous decision in Meyer v. Grant is instructive. There, Colorado enabled citizens to place a proposed state law or constitutional amendment on the general-election ballot if they gathered enough signatures on a petition. See 486 U.S. at 416. Plaintiffs in that case, whose petition concerned whether to deregulate the trucking industry, challenged a state law banning the payment of petition circulators. See id. at 417–21. The Court invalidated the law, applying “exacting scrutiny.” Id. at 420, 425–28.

The Court began its analysis by observing the speech interests at stake, noting that the subject of the petition was “a matter of societal concern” and that the plaintiffs “[u]nquestionably . . . [had] a right to discuss” this issue “publicly without risking criminal sanctions.” Id. at 421. As the Court explained, “[t]he freedom of speech and of the press guaranteed by the Constitution embraces at the least the liberty to discuss publicly and truthfully all matters of public concern

8The Court has also termed this standard “strict scrutiny.” McIntyre, 514 U.S. at 346 n.10 (employing Meyer’s standard and noting that the Court there “unanimously applied strict scrutiny”). No. 22-5028 Lichtenstein v. Hargett Page 41

without previous restraint or fear of subsequent punishment.” Id. (quoting Thornhill v. Alabama, 310 U.S. 88, 101–02 (1940)). This is because “the First Amendment reflects a ‘profound national commitment’ to the principle that ‘debate on public issues should be uninhibited, robust, and wide-open.’” Boos v. Barry, 485 U.S. 312, 318 (1988) (quoting N.Y. Times Co. v. Sullivan, 376 U.S. 254, 270 (1964)).

The Court then considered how these speech interests intersected with the activity at issue. Citing several paragraphs of testimony in the record, the Court found that petition circulation “will in almost every case involve an explanation of the nature of the proposal and why its advocates support it.” Id. at 421 & n.4. “Thus,” the Court concluded, “the circulation of a petition involves the type of interactive communication concerning political change that is appropriately described as ‘core political speech.’” Id. at 421–22. And based on that conclusion, the Court found that Colorado’s law restricted “political expression” by reducing the number of speakers, how much they would speak, the size of the audience reached, and the likelihood that the issue would get on the ballot. See id. at 422–23.

The Court determined that Colorado could not justify this restriction. The state argued that “the burden [was] permissible because other avenues of expression remain[ed] open.” Id. at 424. But this argument failed because the law “restrict[ed] access to the most effective, fundamental, and perhaps economical avenue of political discourse, direct one-on-one communication,” and “[t]he First Amendment protect[ed] [plaintiffs’] right not only to advocate their cause but also to select what they believe[d] to be the most effective means for so doing.” Id. This was an unconstitutional burden even though the plaintiffs could still spend heavily on direct communication by paying workers to discuss the petition door-to-door or elsewhere; what mattered was that they could not pay workers to gather signatures in this fashion. The state also argued that the burden was allowed “because [it] [had] the authority to impose limitations on the scope of the state-created right to legislate by initiative.” Id. But this argument failed too because the state did not have “the power to limit discussion of political issues raised in initiative petitions.” Id. at 425.

Accordingly, Meyer teaches two important lessons. First, restrictions on an activity trigger heightened scrutiny under the First Amendment if, as a factual matter, the activity No. 22-5028 Lichtenstein v. Hargett Page 42

necessarily involves core political speech. Second, exacting scrutiny applies even if speakers are free to express themselves in other ways and even if the speech arises in the context of a state- controlled mechanism.9

2.

Outside the realm of petition circulation, the Court has applied Meyer’s exacting scrutiny to restrictions on a similar activity—pamphlet distribution. See McIntyre, 514 U.S. at 346 (citing Meyer, 486 U.S. at 420); see also Buckley, 525 U.S. at 190 (noting that “[i]nitiative-petition circulators . . . resemble handbill distributors”). And in cases long preceding Meyer, the Court also employed demanding forms of scrutiny for restrictions on this activity. See, e.g., Jamison v. Texas, 318 U.S. 413, 416 (1943); Schneider v. State (Town of Irvington), 308 U.S. 147, 164 (1939); Lovell v. City of Griffin, 303 U.S. 444, 451–52 (1938).

This stringent approach is unsurprising, given that pamphlet distribution “is the essence of First Amendment expression.” McIntyre, 514 U.S. at 347. Few traditions are more embedded in this nation’s political history:

It was in this form—as pamphlets—that much of the most important and characteristic writing of the American Revolution appeared. For the Revolutionary generation, as for its predecessors back to the early sixteenth century, the pamphlet had peculiar virtues as a medium of communication. Then, as now, it was seen that the pamphlet allowed one to do things that were not possible in any other form.

McCullen v. Coakley, 573 U.S. 464, 489 n.5 (2014) (quoting B. Bailyn, The Ideological Origins of the American Revolution 2 (1967)). Pamphlets “have been historic weapons in the defense of liberty, as the pamphlets of Thomas Paine and others in our own history abundantly attest.” Lovell, 303 U.S. at 452.

These cases reinforce Meyer’s lessons. Like petition circulation, pamphlet distribution is an activity that inherently involves “core political speech,” and restrictions on this activity are

9I reject the majority’s cramped focus on speech “inputs” and suggestion that exacting scrutiny applied in Meyer only because Colorado restricted spending on speech. See Maj. Op. 10–11. Plainly, it mattered to the Court that the Meyer plaintiffs were restricted from spending their money in particular ways. That is, the plaintiffs were restricted from pursuing particular activities, the same as Plaintiffs here. No. 22-5028 Lichtenstein v. Hargett Page 43

thus subject to exacting scrutiny. McIntyre, 514 U.S. at 347. Parties have come before the Court and asserted that pamphlet distribution may be banned so long as publication of the underlying material is allowed. See, e.g., Lovell, 303 U.S. at 452. But this country’s free-speech traditions emphatically reject that notion: “Liberty of circulating is as essential to that freedom as liberty of publishing; indeed, without the circulation, the publication would be of little value.” Id. (quoting Ex parte Jackson, 96 U.S. 727, 733 (1877)).

Against this backdrop, I would employ “exacting scrutiny” to evaluate Tennessee’s law.

The first question is whether Tennessee’s law restricts Plaintiffs’ core political speech. As done in Meyer, I start by identifying the subject of Plaintiffs’ speech—whether and how to vote by absentee ballot in Tennessee—which is “[u]nquestionably . . . a matter of societal concern that [Plaintiffs] have a right to discuss publicly without risking criminal sanctions,” Meyer, 486 U.S. at 421. As the majority notes, “Plaintiffs’ underlying get-out-the-vote activities . . . [are] entitled to rigorous First Amendment protection.” Maj. Op. 12. And as this court noted in a case concerning Tennessee’s voting laws, the COVID-19 pandemic brought absentee voting “into the spotlight” and “led to increased interest in the policies and procedures governing how and when voters may vote absentee.” Memphis A. Philip Randolph Inst. v. Hargett, 978 F.3d 378, 381 (6th Cir. 2020).

I next look to how these speech interests intersect with the activity of distributing absentee-ballot applications, considering whether this activity necessarily involves core political speech. Given the factual nature of this inquiry, see Meyer, 486 U.S. at 421 n.4, and that we are reviewing the grant of a motion to dismiss, I accept Plaintiffs’ nonconclusory factual allegations as true and consider their import, see Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). Plaintiffs allege that they “engage[] in voter outreach efforts . . . through voter identification, education, and mobilization drives,” R. 1, PID 3, and that their efforts “will necessarily include discussing with voters the benefits of voting by mail, reminding eligible absentee voters about application No. 22-5028 Lichtenstein v. Hargett Page 44

and ballot submission deadlines and requirements, and following up with voters to ensure their ballots were received, cast and counted,” id. at PID 9.

Based on these allegations, Plaintiffs’ intended distribution activity “involves the type of interactive communication concerning political change that is appropriately described as ‘core political speech,’” Meyer, 486 U.S. at 421–22. The activity “of necessity involves both the expression of a desire for political change and a discussion of the merits of the proposed change,” id. at 421—that absentee voting is good for the electorate and that a person should vote absentee if eligible. “[I]n almost every case,” it appears, giving someone an absentee-ballot application “involve[s] an explanation of the nature of [voting absentee] and why its advocates support [doing so],” id.

Thus, Tennessee’s law “restricts political expression,” id. at 423. The law bans Plaintiffs from engaging in an activity that inherently involves core political speech. And this burden is even more significant because Plaintiffs allege that distributing absentee-voting applications is the most informative and effective means for them to convey their message. “Having the ability to provide voters with the absentee ballot application is necessary,” Plaintiffs allege, “because [they] have found that their voter engagement efforts are significantly more effective when they are able to provide voters with all of the information and requisite forms they might need to register to vote, or to request to vote absentee.” R. 1, PID 9; cf. Meyer, 486 U.S. at 424 (plaintiffs were permitted to pay workers to engage citizens on the issue but not to collect signatures)); McCullen, 573 U.S. at 489.10

Reinforcing this analysis, the Fifth Circuit would appear to take the same approach. In Voting for America, Inc. v. Steen, plaintiffs challenged restrictions on who could serve

10In McCullen, the Court addressed the burden placed on abortion protestors by a buffer zone: “Respondents also emphasize that the Act does not prevent petitioners from engaging in various forms of ‘protest’— such as chanting slogans and displaying signs—outside the buffer zones. That misses the point. Petitioners are not protestors. They seek not merely to express their opposition to abortion, but to inform women of various alternatives and to provide help in pursuing them. Petitioners believe that they can accomplish this objective only through personal, caring, consensual conversations. And for good reason: It is easier to ignore a strained voice or a waving hand than a direct greeting or an outstretched arm.” 573 U.S. at 489 (citation omitted). No. 22-5028 Lichtenstein v. Hargett Page 45

as volunteer deputy registrars in the context of voter-registration drives. 732 F.3d 382, 389 (5th Cir. 2013). The court concluded that the restrictions, which limited who could “collect, review for completeness, and deliver completed voter registration forms,” did not burden “core political speech,” noting that these activities were “non-expressive” and that there was no restriction on “who [could] advocate pro-voter-registration messages, the manner in which they [might] do so, or any communicative conduct.” Id. at 390–91 (citation omitted). However, the court assumed that the First Amendment protects the distribution of registration forms:

The state does not deny that some voter registration activities involve speech— “urging” citizens to register; “distributing” voter registration forms; “helping” voters to fill out their forms; and “asking” for information to verify that registrations were processed successfully. Texas neither regulates nor limits any of this constitutionally protected speech.

Id. at 389 (emphasis added). This reasoning supports that Tennessee’s law burdens Plaintiffs’ core political speech.

Additionally, I note the undeniable parallels between Plaintiffs’ intended activity and pamphlet distribution, a “classic form[] of speech that lie[s] at the heart of the First Amendment.” Schenck v. Pro-Choice Network of W.N.Y., 519 U.S. 357, 377 (1997). The majority appears to recognize these parallels as well, although it rejects the comparison. See Maj. Op. 15 (“One might analogize the form’s distribution to that of a political pamphlet.”). In Buckley, the Court noted that “[i]nitiative-petition circulators . . . resemble handbill distributors, in that both seek to promote public support for a particular issue or position.” 525 U.S. at 190–91. The same can be said of Plaintiffs, who wish to promote engagement in the democratic process through absentee voting by distributing absentee-ballot applications. And Plaintiffs “resemble handbill distributors” even more than the petition circulators in Buckley; like pamphlet distributors, Plaintiffs wish to distribute printed information about a pertinent political issue. By the nature of the application and what appears on Tennessee’s precise form, the state’s absentee-ballot applications inform voters directly about the existence of absentee voting, the steps to apply, who is eligible to vote this way, and the deadline for submission. Thus, Plaintiffs’ intended activity is substantially similar to pamphlet distribution, which the First Amendment fiercely protects. See Schenk, 519 U.S. at 377. No. 22-5028 Lichtenstein v. Hargett Page 46

Relatedly, as Amicus Cato Institute points out, this reasoning is consistent with broader traditions for protecting access to the law. See Br. of Amicus Cato Inst. 3–4, 9–12. In the copyright context, the Court has explained that “‘[e]very citizen is presumed to know the law,’ and ‘it needs no argument to show . . . that all should have free access’ to its contents.” Georgia v. Public.Resource.Org, Inc., 140 S. Ct. 1498, 1507 (2020) (second alteration in original) (quoting Nash v. Lathrop, 6 N.E. 559, 560 (Mass. 1886)); see also Veeck v. S. Bldg. Code Cong. Int’l, Inc., 293 F.3d 791, 799 (5th Cir. 2002) (“[P]ublic ownership of the law means precisely that ‘the law’ is in the ‘public domain’ for whatever use the citizens choose to make of it. Citizens may reproduce copies of the law for many purposes, not only to guide their actions but to influence future legislation, educate their neighborhood association, or simply to amuse.”). Similarly, access to government forms is intertwined with the ability “to petition the government for a redress of grievances,” as explicitly enshrined in the First Amendment. U.S. Const. amend. I. In Cato’s words, “[a]llowing the government to criminalize the distribution of publicly available government forms . . . set[s] a dangerous precedent,” which “should give th[is] [c]ourt pause.” Br. of Amicus Cato Inst. 1.

3.

For the reasons discussed, Plaintiffs’ allegations support the use of “exacting scrutiny” to evaluate Tennessee’s law, making resolution on the pleadings inappropriate. Tennessee “must come forward with compelling evidence” showing that the law is “narrowly tailored and advance[s] a compelling state interest.” Citizens for Tax Reform v. Deters, 518 F.3d 375, 387 (6th Cir. 2008). And as the majority concedes, Plaintiffs raise “fair points” concerning the statute’s tailoring, which “might have merit” under “anything like strict scrutiny.” Maj. Op. 32. Therefore, I would allow the case to proceed for resolution on a full record, as is fitting in light of the numerous issues subject to factual inquiry.11

11To the extent there is any question that Meyer as later interpreted in Buckley requires some inquiry into the degree of the burden before applying exacting scrutiny, see Citizens for Tax Reform, 518 F.3d at 382, exacting scrutiny remains appropriate because, as addressed, Tennessee’s law completely bars Plaintiffs from an activity that inherently involves core political speech and bars them from their most effective form of advocacy. See Buckley, 525 U.S. at 192 n.12 (“Our decision is entirely in keeping with the ‘now-settled approach’ that state regulations ‘impos[ing] “severe burdens” on speech . . . [must] be narrowly tailored to serve a compelling state interest.’” (alterations in original) (quoting id. at 206 (Thomas, J., concurring))). No. 22-5028 Lichtenstein v. Hargett Page 47

C.

I do not agree with the majority’s arguments regarding the appropriate level of scrutiny. According to the majority, “Tennessee’s ban prohibits an act: distributing a government form,” and “[t]his act qualifies as conduct, not speech.” Maj. Op. 2 (emphasis added). The majority also notes that “Plaintiffs’ underlying get-out-the-vote activities . . . qualif[y] as ‘core political speech,’” id. at 12 (quoting Meyer, 486 U.S. at 422), but concludes that there is no First Amendment problem because “nothing in [Tennessee’s statute] in any way restricts the Plaintiffs’ actual oral or written speech about the ‘benefits’ of absentee voting,” id. (emphasis added) (quoting R. 1, PID 9).

This reasoning contradicts Meyer because, applied to the facts of that case, it suggests that Colorado’s ban on paid circulators merely burdened conduct. After all, signature-gathering could be termed “conduct, not speech,” and Colorado’s law did not prevent the plaintiffs from spending money in other ways to support their “underlying . . . activities.” Id. at 2, 13. But this is not how the Court approached the case. The Court recognized that circulating petitions “of necessity involve[d] . . . interactive communication concerning political change,” and thus the Court treated burdens on this activity as burdens on core political speech. Meyer, 486 U.S. at 421–22. I follow that same reasoning here.

The majority too easily distinguishes Plaintiffs’ activity from pamphlet distribution. See Maj. Op. 15–16. First, the majority states that “the Court has long treated [pamphlet] distribution (or more precisely, the distribution of a pamphlet’s expressive contents) as speech itself,” not conduct, and “Plaintiffs did not make this analogy.” Id. at 15. This reasoning misses Plaintiffs’ point: Distribution involves conduct but because that conduct is necessarily intertwined with political speech and expressive intent, exacting scrutiny should apply to the entire activity. See Meyer, 486 U.S. at 421. The majority’s reasoning comes dangerously close to suggesting that distribution of political pamphlets could be restricted as mere conduct. Technically, the act of giving another person a piece of paper—whether a pamphlet or an absentee-ballot application—is conduct, not speech, and the majority seems to recognize this. See Maj. Op. 15. The Court actually refers briefly to pamphlet distribution as “conduct” in McIntyre. 514 U.S. at 339; cf. Org. for a Better Austin v. Keefe, 402 U.S. 415, 419 (1971) No. 22-5028 Lichtenstein v. Hargett Page 48

(referring to pamphlet distribution as an “activity” that “is a form of communication”). But sensibly, the McIntyre Court did not follow O’Brien; it followed Meyer.

Second, the majority notes that “the application is a form that the State creates” and “[i]f the form’s contents were speech, the speech might well be the government’s,” not Plaintiffs’. Maj. Op. 16. To be sure, that the application is a government form is likely important in considering the government’s interest in restricting distribution and the law’s tailoring. But Plaintiffs may still assert a speech interest in sharing the form and its contents. To illustrate, McIntyre applied exacting scrutiny to a ban on anonymous pamphlet distribution, but I doubt the Court’s analysis would have changed had the plaintiff, who distributed literature before a schoolboard meeting, see 514 U.S. at 337, instead distributed local government reports on school spending.

Third, the majority asserts that, “[i]n any event, the Plaintiffs have argued that the banned conduct is entitled to rigorous protection only because it is intertwined with speech . . . or because the relevant ban is content[-]based.” Maj. Op. 16. But this does not make caselaw on pamphlet distribution any less relevant. Pamphlet distribution stands as an example of an activity that involves aspects of conduct but is treated as speech for very good reason. Cf. Lovell, 303 U.S. at 452 (rejecting argument that a restriction on distributing pamphlets is allowed because publication still permitted).

D.

Because I conclude that Tennessee’s law restricts core political speech, I agree that Anderson-Burdick does not apply. See Mazo v. N.J. Sec’y of State, 54 F.4th 124, 142–43 (3d Cir. 2022). But the Sixth Circuit has not cabined Anderson-Burdick as tightly as the majority suggests, see, e.g., Daunt v. Benson, 956 F.3d 396, 407 (6th Cir. 2020) (reasoning that “the Anderson-Burdick framework is used for evaluating ‘state election law[s]’” and applying to challenge to Michigan’s redistricting commission (alteration in original) (quoting Burdick v. Takushi, 504 U.S. 428, 441 (1992)), nor has the Supreme Court done so, see McIntyre, 514 U.S. at 345 (finding case not suitable for Anderson-Burdick balancing because law did “not control the mechanics of the electoral process”). No. 22-5028 Lichtenstein v. Hargett Page 49

E.

Turning to the majority’s expressive-conduct analysis, I am less skeptical about whether distributing absentee-ballot applications qualifies as expressive conduct. It seems plain that giving someone an absentee-ballot application with even the smallest amount of context—verbal or not—conveys a pro-vote-by-mail message. Cf. Maj. Op. 23–25. I also believe that the majority goes too far in characterizing intermediate scrutiny for expressive conduct as nearly equivalent to rational-basis review, given that the Supreme Court has never instructed courts to apply rational-basis review to these challenges, despite rational-basis review being readily available as an alternative. See id. at 26–30.

Moreover, I do not agree that “[t]he Supreme Court has never applied exacting free-speech scrutiny to laws that bar conduct based on the harm that the conduct causes apart from the message it conveys,” id. at 8, and that Plaintiffs are not entitled to heightened scrutiny simply because Tennessee’s law “does not target the conduct because of the ‘idea’ that the Plaintiffs want to express,” id. at 12. Even when a law “does not focus on the ideas expressed by persons or groups subject to its regulations,” we apply heightened scrutiny if the government’s interests are based on “suppressing communication.” Buckley v. Valeo, 424 U.S. 1, 17 (1976). Nevertheless, Plaintiffs have not shown that Tennessee fails to assert “an interest . . . unrelated to the suppression of expression,” Texas v. Johnson, 491 U.S. 397, 407 (1989).

III.

Additionally, I would reverse the district court’s dismissal of Plaintiffs’ expressive-association claim. “It is beyond debate that freedom to engage in association for the advancement of beliefs and ideas is an inseparable aspect of the ‘liberty’ assured by the Due Process Clause of the Fourteenth Amendment, which embraces freedom of speech.” NAACP v. Alabama ex rel. Patterson, 357 U.S. 449, 460 (1958). “[T]he practice of persons sharing common views banding together to achieve a common end is deeply embedded in the American political process. . . . Its value is that by collective effort individuals can make their views known, when, individually, their voices would be faint or lost.” Citizens Against Rent Control/Coal. for Fair Hous. v. City of Berkeley, 454 U.S. 290, 294 (1981). No. 22-5028 Lichtenstein v. Hargett Page 50

We evaluate expressive-association claims using a three-part process:

The first determination is whether a group is entitled to protection. A group need not associate “for the ‘purpose’ of disseminating a certain message” to be protected; it is enough that a group “engage[s] in expressive activity that could be impaired.” . . . Second, courts ask whether the government action in question “significantly burden[s]” the group’s expression, affording deference “to an association’s view of what would impair its expression.” Lastly, the government’s interest in any restriction must be weighed against plaintiff’s right of expressive association.

Miller v. City of Cincinnati, 622 F.3d 524, 538 (6th Cir. 2010) (citations omitted) (quoting Boy Scouts of Am. v. Dale, 530 U.S. 640, 653, 655 (2000)).

I agree with the majority that Plaintiffs “easily make it past our first step,” Maj. Op. 34, but diverge at step two and conclude that Plaintiffs have sufficiently pleaded that Tennessee’s law “significantly burden[s]” their expression, id. at 35 (alteration in original) (quoting Dale, 530 U.S. at 653). The majority “do[es] not see how” the law “affects the Plaintiffs’ ability to associate with others in order to create or convey a message,” citing examples where the Court has found unconstitutional burdens and reasoning that Tennessee’s law does not compel Plaintiffs to accept members, discourage membership, or deny a benefit broadly available to others. Id. The majority also notes that “Plaintiffs argue that they ‘associate’ with members in the community through their ‘civic engagement activity’ and that” the law “limits their ability to undertake this speech,” but reasons that the alleged interference occurs regardless whether people act individually or as a group, and so concludes that “the interference does not affect [Plaintiffs’] ability to associate at all.” Id. at 36 (quoting Appellants’ Reply Br. 16).

This reasoning is too narrow in its application of the Supreme Court’s cases. The majority correctly acknowledges that “[a] government can burden the right to associate in a ‘number’ of ways,” id. at 35 (quoting Ams. for Prosperity Found. v. Bonta, 141 S. Ct. 2373, 2382 (2021) (plurality opinion)), but fails to consider the diversity of possible burdens, instead only comparing Plaintiffs’ claim with the three scenarios mentioned above. Cf. Dale, 530 U.S. at 648 (explaining that unconstitutional burdens on association “may take many forms”). Further, the majority fails to acknowledge that “we must . . . give deference to an association’s view of what would impair its expression.” Id. at 653 (emphasis added). No. 22-5028 Lichtenstein v. Hargett Page 51

The Court has recognized a wider variety of associational burdens than the majority suggests. “In the domain of these indispensable liberties, whether of speech, press, or association, . . . abridgement of such rights, even though unintended, may inevitably follow from varied forms of governmental action.” Patterson, 357 U.S. at 460. For example, the Court held that a restriction on soliciting legal clients unconstitutionally interfered with the NAACP’s associational rights because “the litigation [the NAACP] assists, while serving to vindicate the legal rights of members of the American Negro community, at the same time and perhaps more importantly, makes possible the distinctive contribution of a minority group to the ideas and beliefs of our society.” NAACP v. Button, 371 U.S. 415, 431 (1963). Similarly, the Court determined that a restriction on boycotts, asserted to be an economic protection measure, unconstitutionally interfered with the NAACP’s associational rights because the organization’s boycott “clearly involved constitutionally protected activity,” noting that “[t]hrough speech, assembly, and petition—rather than through riot or revolution—petitioners sought to change a social order that had consistently treated them as second-class citizens.” NAACP v. Claiborne Hardware Co., 458 U.S. 886, 911 (1982).

With this background, Plaintiffs have pleaded enough to show a “significant burden” on their expressive association. In addition to what they describe in their Reply Brief, as recounted by the majority, Plaintiffs allege that Tennessee’s law “prohibits them from fully engaging their members and other eligible absentee voters,” and from “facilitating their ability to obtain an absentee ballot application and vote by mail.” R. 1, PID 11. They also allege that the law “hampers [their] ability to plan and execute voter engagement strategies.” Id. Giving these allegations the “deference” the Supreme Court tells me I must, Dale, 530 U.S. at 653, it seems to me “hardly a novel perception that” restricting the ability to share voting forms “may constitute . . . a restraint on freedom of association,” Patterson, 357 U.S. at 462 (concluding that law compelling disclosure of NAACP’s membership lists restricted the freedom of association because it discouraged membership).

As alleged, Tennessee’s law hinders the ability of Plaintiffs to share information with their members. This ability is essential to any associational enterprise, especially one engaged in expression. And importantly, we must consider Tennessee’s law “[i]n the context of [Plaintiffs’] No. 22-5028 Lichtenstein v. Hargett Page 52

objectives.” Button, 371 U.S. at 429. In Button, the Court recognized that, for the NAACP, “litigation is not a technique of resolving private differences; it is a means for achieving the lawful objectives of equality of treatment by all government, federal, state and local, for the members of the Negro community in this country,” and thus it was “a form of political expression.” 371 U.S. at 429. So too here—Plaintiffs’ intended distribution of absentee-ballot applications is more than a simple act; it provides the means for engaging members as part of Plaintiffs’ mission to promote a message that supports participation in our democracy through voting absentee. Tennessee’s law directly blocks Plaintiffs from sharing a form with members who need it to vote—even if they have asked Plaintiffs for help—or with members as a pure educational exercise to help them understand their rights and how to assist others in voting absentee. Plaintiffs thus satisfy the first two steps of our expressive-association inquiry.

The third step weighs Plaintiffs’ rights against the government’s interests. See Miller, 622 F.3d at 538. The freedom of expressive association can “be overridden ‘by regulations adopted to serve compelling state interests, unrelated to the suppression of ideas, that cannot be achieved through means significantly less restrictive of associational freedoms.’” Dale, 530 U.S. at 648 (quoting Roberts v. U.S. Jaycees, 468 U.S. 609, 623 (1984)). However, Tennessee has made no showing at this stage that its law satisfies this tailoring. Accordingly, I find resolution of Plaintiffs’ expressive-association claims on the pleadings inappropriate, and I would reverse the district court’s dismissal of this claim.

IV.

For the foregoing reasons, I dissent.