Emory Law Journal

State Constitutions as a Check on the New Governors: Using State Free Speech Clauses to Protect Social Media Users from Arbitrary Political Censorship by Social Media Platforms *This Comment received the Myron Penn Laughlin Award for Excellence in Legal Research and Writing.
Elijah O’Kelley *J.D. Candidate, Class of 2020.

Abstract

Social media platforms like Facebook are increasingly the arbiters of what political viewpoints get the light of day. As more people become active on social media, including using it as a primary news source, and as political speech increasingly moves onto these platforms, they will continue amassing the power to control news narratives, the size of speakers’ bullhorns, and the ideas that get discussed. Moreover, these platforms will likely face increasing external pressures to silence certain viewpoints deemed objectionable or offensive, whatever those views may be. The United States Constitution provides no check on this power. As private actors, social media platforms can ban whatever speech they want, and for whatever reason, subject only to market pressures. Simply, the First Amendment, based on its text and longstanding precedent, does not apply.

There may be another way to safeguard speech online, however. This Comment proposes looking to a different source of law, one all too often overlooked in litigation and scholarship: state constitutions. Nearly all state constitutions have free speech clauses that are textually different from the First Amendment in ways suggesting state action requirements can be softened or even jettisoned. Moreover, a handful of states have interpreted their free speech clauses as being broader than the First Amendment and applying to certain private actors—an approach with express approval by a unanimous United States Supreme Court. This case law provides reasoning applicable to a new and modern dilemma: the gravitation of political speech to social media platforms.

This Comment reviews those state free speech clauses and that case law to develop an approach state courts can adopt and apply to social media platforms. Using Facebook as an example, it pulls from the main themes of the relevant state cases to establish that the reasoning is extendable to Facebook and is even a better fit for it than for the private actors in the original cases. It then explores the nuances of what power Facebook should retain to moderate content, concluding that, although the underlying viewpoints of political speech are maximally protected, the platform can continue to regulate the specific way they are expressed if those expressions involve concrete language like slurs and other epithets unnecessary for the expression of the viewpoint. This approach allows Facebook to maintain much of its autonomy while creating an incentive for it to err on the side of caution before deleting political speech. And, importantly, it defines a way for courts to act as a check on Facebook’s otherwise unchecked power to control much of what political speech is seen online.

Introduction

Social media platforms are giants, and Facebook 1 Most Famous Social Network Sites Worldwide as of October 2018, Ranked by Number of Active Users, Statista, https://www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/ (last visited Jan. 14, 2019). See generally Social Media Fact Sheet, Pew Res. Ctr., http://www.pewinternet.org/fact-sheet/social-media/ (last accessed Jan. 14, 2019). Facebook recently broke two billion users after a decade of rapid growth, see Felix Richter, 3 in 10 World Citizens Are on Facebook, Statista (Apr. 25, 2019), https://www.statista.com/chart/10047/facebooks-monthly-active-users/, though its growth has plateaued recently, see Kurt Wagner & Rani Molla, Facebook’s User Growth Has Hit a Wall, Vox (July 25, 2018), https://www.vox.com/2018/7/25/17614426/facebook-fb-earnings-q2-2018-user-growth-troubles. This reflects a broader trend of plateaued social media growth in the last two years. Paul Hitlin, Internet, Social Media Use and Device Ownership in U.S. Have Plateaued After Years of Growth, Pew Res. Ctr. (Sept. 28, 2018), http://www.pewresearch.org/fact-tank/2018/09/28/internet-social-media-use-and-device-ownership-in-u-s-have-plateaued-after-years-of-growth/. This is not necessarily a sign of a decline in the power of social media, as it may be attributable to the fact that the platforms have reached near-saturation levels. Id.; see also infra note 233 (discussing inherent tendency toward scale of social media platforms). is the chief behemoth. 2Cf. Mikhail Bulgakov, The Master and Margarita 69 (Diana Burgin & Katherine Tiernan O’Connor trans., Ardis Publishers 1995) (1967) (“But there were worse things to be seen in the bedroom: sprawled in a relaxed pose on the pouffe that had once belonged to the jeweler’s wife was a third creature, namely, a black cat of horrific proportions with a glass of vodka in one paw and in the other a fork on which he had speared a pickled mushroom.”) (description of the character “Behemoth”). Around sixty-eight percent of all adults in the United States use Facebook, most of them at least daily and half of them many times a day. 3 Aaron Smith & Monica Anderson, Social Media Use in 2018, Pew Res. Ctr. (Mar. 1, 2018), http://www.pewinternet.org/2018/03/01/social-media-use-in-2018/. Other than adults older than sixty-five, majorities use Facebook across various demographic groups, including income, education, geographic location, sex, race, and age. 4Id. Younger people tend to use Facebook at even higher rates—around eighty percent of all adults aged eighteen to fifty use Facebook. 5Id. at Appendix A.

It is no surprise, then, that political speech is gravitating toward social media. Sixty-eight percent of adults now get their news from some social media platform, and forty-three percent get news from Facebook. 6See Katerina Eva Matsa & Elisa Shearer, News Use Across Social Media Platforms 2018, Pew Res. Ctr. (Sept. 10, 2018), http://www.journalism.org/2018/09/10/news-use-across-social-media-platforms-2018/. Not only is social media increasing in prevalence, but broadly speaking it has surpassed declining traditional news sources like print, 7 Elisa Shearer, Social Media Outpaces Print Newspapers in the U.S. as a News Source, Pew Res. Ctr. (Dec. 10, 2018), http://www.pewresearch.org/fact-tank/2018/12/10/social-media-outpaces-print-newspapers-in-the-u-s-as-a-news-source/. and online news generally is on track to surpass television news. 8See Jeffrey Gottfried & Elisa Shearer, Americans’ Online News Use Is Closing in on TV News Use, Pew Res. Ctr. (Sept. 7, 2017), http://www.pewresearch.org/fact-tank/2017/09/07/americans-online-news-use-vs-tv-news-use/. Social media and online advertising are also increasingly important for political activism 9See Benjamin F. Jackson, Censorship and Freedom of Expression in the Age of Facebook, 44 N.M. L. Rev. 121, 125–26 (2014); Loren F. Selznick & Carolyn LaMacchia, #Mall Ruckus Tonight: Should Mall Owners be Forced to Provide a Stage for Expression in the Virtual Age?, 53 Willamette L. Rev. 239, 271–73 (2017); Monica Anderson et al., Activism in the Social Media Age, Pew Res. Ctr. (July 11, 2018), http://www.pewinternet.org/2018/07/11/activism-in-the-social-media-age/. and campaigns. 10See Jackson, supra note 9, at 124. For example, the Obama presidential campaigns in 2008 and 2012 used social media, and the internet generally, to great effect. See, e.g., Ed Pilkington & Amanda Michel, Obama, Facebook and the Power of Friendship: The 2012 Data Election, Guardian (Feb. 17, 2012), https://www.theguardian.com/world/2012/feb/17/obama-digital-data-machine-facebook-election.

The United States Supreme Court recently recognized the importance of social media for speech: “[T]he most important place[] ... for the exchange of views, today … is cyberspace … and social media in particular.” 11 Packingham v. North Carolina, 137 S. Ct. 1730, 1735 (2017). The Court went on to praise Facebook’s capacity for unlimited, low-cost communication, including political speech. Id. at 1735–36; see also Reno v. ACLU, 521 U.S. 844, 870 (1997) (“Through the use of chat rooms, any person with a phone line can become a town crier with a voice that resonates farther than it could from any soapbox.”). In the same case, the Court struck down a state law prohibiting sex offenders from accessing social media platforms like Facebook because these platforms are so important for free speech that a law blocking access to them violates the First Amendment. 12Packingham, 137 S. Ct. at 1737 (“North Carolina with one broad stroke bars access to what for many are the principal sources for knowing current events, checking ads for employment, speaking and listening in the modern public square, and otherwise exploring the vast realms of human thought and knowledge…. [T]o foreclose access to social media altogether is to prevent the user from engaging in the legitimate exercise of First Amendment rights.”). Interestingly, some courts have found that government social media pages are public forums for First Amendment purposes, at least when they have some interactive component. See Davison v. Randall, 912 F.3d 666, 682, 688 (4th Cir. 2019); Knight First Amendment Inst. v. Trump, 302 F. Supp. 3d 541, 574 (S.D.N.Y. 2018). But see Morgan v. Bevin, 298 F. Supp. 3d 1003, 1010–11 (E.D. Ky. 2018).

The Court did not address, however, whether any actions the platforms take to curtail political speech violate the First Amendment. And it is unlikely to do so. Based on longstanding precedent, Facebook is not bound by the First Amendment because it is a private actor and the Amendment applies only to state action. 13See, e.g., Hudgens v. NLRB, 424 U.S. 507, 513 (1976) (“It is, of course, a commonplace that the constitutional guarantee of free speech is a guarantee only against abridgment by government, federal or state.”); Shulman v. Facebook.com, No. 17-764 (JMV), 2017 WL 5129885, at *4 (D.N.J. Nov. 6, 2017) (collecting cases of failed attempts to subject Facebook to First Amendment); Kate Klonick, The New Governors: The People, Rules, and Processes Governing Online Speech, 131 Harv. L. Rev. 1598, 1609–13 (2018) (discussing several current arguments regarding state action and the First Amendment). In recent litigation involving alleged censorship by YouTube, a federal district court dismissed based in part on YouTube not being a state actor and thus not subject to the First Amendment. Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 WL 1471939, at *8–9 (N.D. Cal. Mar. 26, 2018). In doing so, the district court provided a helpful summary of the current state of First Amendment jurisprudence as it relates to online platforms. See id. at *5–8. In a recent case involving alleged mismanagement of public access channels by a private nonprofit corporation, the Supreme Court held that simply operating public access channels did not turn the corporation into a “state actor,” and the corporation was thus not subject to the First Amendment. Manhattan Cmty. Access Corp. v. Halleck, 139 S. Ct. 1921, 1934 (2019). This suggests the Court is unlikely to expand the state action doctrine in the near future. Therefore, there seems to be little to nothing in federal law preventing social media platforms from banning all users who voice an unfavorable opinion, 14See Jonathan Peters, The “Sovereigns of Cyberspace” and State Action: The First Amendment’s Application—or Lack Thereof—to Third-Party Platforms, 32 Berkeley Tech. L.J. 989, 1013–18 (2017). Senator Josh Hawley of Missouri has proposed legislation that would subject social media platforms to audits and strip certain legal protections from them if they censor viewpoints. See Senator Hawley Introduces Legislation to Amend Section 230 Immunity for Big Tech Companies, Josh Hawley (June 19, 2019), https://www.hawley.senate.gov/senator-hawley-introduces-legislation-amend-section-230-immunity-big-tech-companies. It is unclear whether this kind of legislation has any chance of becoming law, and this Comment’s focus is on case law and judiciaries, not statutes and legislatures. whether that opinion is that immigration should be limited or that taxes should be increased on the wealthy.

It is easy to imagine scenarios where platforms may engage in this kind of content moderation, whether because of external pressures 15See Klonick, supra note 13, at 1616–29. or internal biases. 16See Jackson, supra note 9, at 129–31. It is no secret that Silicon Valley employees, for better or worse, skew heavily toward one end of the political spectrum. See, e.g., Dan Boylan, Silicon Valley an “Extremely Left-Leaning Place,” Admits Zuckerberg, Wash. Times (Apr. 10, 2018), https://www.washingtontimes.com/news/2018/apr/10/zuckerberg-admits-silicon-valley-extremely-left-le/; Kate Conger & Sheera Frenkel, Dozens at Facebook United to Challenge its “Intolerant” Liberal Culture, N.Y. Times (Aug. 28, 2018), https://www.nytimes.com/2018/08/28/technology/inside-facebook-employees-political-bias.html; Kirsten Grind & Keach Hagey, Why Did Facebook Fire a Top Executive? Hint: It Had Something To Do with Trump, Wall Street J. (Nov. 11, 2018), https://www.wsj.com/articles/why-did-facebook-fire-a-top-executive-hint-it-had-something-to-do-with-trump-1541965245; Farhad Manjoo, Silicon Valley’s Politics: Liberal, with One Big Exception, N.Y. Times (Sept. 6, 2017), https://www.nytimes.com/2017/09/06/technology/silicon-valley-politics.html. Given their continually growing power to shape what news and ideas appear before users, and given their massive number of users, it is not unfathomable that these platforms could use their positions to effectively delete certain speakers, and possibly even certain ideas, from public discourse. 17 A recent example of something approaching this was the collective action to delete Alex Jones’s accounts across platforms, including Facebook, Apple products, Spotify, YouTube, and, eventually, Twitter. See Kate Conger & Jack Nicas, Twitter Bars Alex Jones and Infowars, Citing Harassing Messages, N.Y. Times (Sept. 6, 2018), https://www.nytimes.com/2018/09/06/technology/twitter-alex-jones-infowars.html; Jack Nicas, Alex Jones and Infowars Content Is Removed From Apple, Facebook and YouTube, N.Y. Times (Aug. 6, 2018), https://www.nytimes.com/2018/08/06/technology/infowars-alex-jones-apple-facebook-spotify.html. Because Jones trafficked in ridiculous and demonstrably false conspiracy theories, some of which included the potential defamation of Sandy Hook parents, he falls outside any protection this Comment proposes. See also Rachel Kraus, 2018 Was the Year We (Sort of) Cleaned Up the Internet, Mashable (Dec. 26, 2018), https://mashable.com/article/deplatforming-alex-jones-2018/#GD8mPzDxkqq4 (documenting unsavory figures deplatformed in 2018). Indeed, social media platforms have deleted accounts 18Id.; cf. Disinvitation Database, Fire, https://www.thefire.org/resources/disinvitation-database/ (last visited Aug. 9, 2019) (compiling speakers prevented from speaking on college campuses and showing recent increase). —which is called “deplatforming”—and it appears effective at limiting the reach of speakers and their ideas. 19 Jason Koebler, Deplatforming Works, Vice (Aug. 10, 2018), https://vice.com/en_us/article/bjbp9d/do-social-media-bans-work. This is no small power, and it is one that is nearly absolute—leading scholars to refer to the platforms as the “New Governors” and “Sovereigns of Cyberspace.” 20See Klonick, supra note 13, at 1662–70; Peters, supra note 14, at 993. Though it may be a technical logical fallacy to call this situation a slippery slope, the road certainly has a few banana peels on it.

Because social media platforms are increasingly where speech occurs, and because they have the power to prohibit speech arbitrarily, and because free speech is crucial for public discourse and society, 21 There is no shortage of ideologically diverse writing praising the value of free speech. See John Stuart Mill, All Minus One: John Stuart Mill’s Ideas on Free Speech Illustrated (Richard V. Reeves & Jonathan Haidt eds., 2018); Freedom of Expression, ACLU, https://www.aclu.org/other/freedom-expression (last visited Aug. 9, 2019); Harvey C. Mansfield, The Value of Free Speech, Nat’l Aff. (Fall 2018), https://www.nationalaffairs.com/publications/detail/the-value-of-free-speech. There is, of course, a range of views that spawn from those who are more absolutist to those who see more room for government regulation of things like “hate speech.” See generally Michael Kent Curtis, Critics of “Free Speech” and the Uses of the Past, 12 Const. Comment. 29, 42–65 (1995) (discussing various criticisms of free speech and responding to those criticisms). This Comment will not relitigate battles over whether free speech is a virtue and will simply assume it is. If the reader does not buy that free speech is a good, then they are beyond hope and this Comment will not save them. it is worth asking whether anything within the law can check this power. Multiple ideas have been proposed, including fitting social media platforms into exceptions to the state action doctrine, 22 Jackson, supra note 9, at 146–54; Trevor Puetz, Note, Facebook: The New Town Square, 44 Sw. L. Rev. 385, 396–409 (2014). or revising the state action doctrine itself. 23 Peters, supra note 14, at 1022–26. Rather than looking to federal law, this Comment will look to another source of law: state constitutionalism. 24 At least two articles have mentioned state constitutionalism as an option. See Jackson, supra note 9, at 157–59; Dawn C. Nunziato, The Death of the Public Forum in Cyberspace, 20 Berkeley Tech. L.J. 1115, 1164–67 (2005). However, both included them as briefly mentioned alternatives to their main arguments, and neither fully explored the relevant state case law or developed a fleshed-out approach based on it. Additionally, at least one plaintiff has raised a state constitutional argument against YouTube (in addition to First Amendment arguments), though the federal court declined to exercise supplemental jurisdiction over the issue. See Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 WL 1471939, at *13–14 (N.D. Cal. Mar. 26, 2018).

The basic idea of state constitutionalism is that state courts should interpret state constitutional provisions independently of similar United States constitutional provisions 25See Jeffrey S. Sutton, 51 Imperfect Solutions: States and the Making of American Constitutional Law (2018); William J. Brennan, Jr., State Constitutions and the Protection of Individual Rights, 90 Harv. L. Rev. 489, 502 (1977). instead of “lockstepping” 26See Sutton, supra note 25, at 174–75. with federal courts’ interpretations. This is especially the case when state courts interpret state constitutions to provide greater protections of individual rights than the United States Constitution. 27See id. at 19 (“[I]t often will be easier to vindicate a state constitutional right in a State’s high court than a federal constitutional right in the U.S. Supreme Court.”); Brennan, supra note 25, at 495. In other words, when interpretations of the United States Supreme Court come up short, state courts should step up to “fill the constitutional gaps.” 28 William J. Brennan, Jr., The Bill of Rights and the States: The Revival of State Constitutions as Guardians of Individual Rights, 61 N.Y.U. L. Rev. 535, 548–49 (1986); see also Sutton, supra note 25, at 22–172 (detailing four instances, including compelled sterilization, in which state courts provided greater protection than federal courts).

The merits of state constitutionalism are far from an underdiscussed topic. 29See, e.g., Sutton, supra note 25; Brennan, supra note 25, at 502; Brennan, supra note 28, 548–52; Paul E. McGreal, Alaska Equal Protection: Constitutional Law or Common Law?, 15 Alaska L. Rev. 209, 213–14 nn.11–12 (1998) (gathering various articles); Jeffrey S. Sutton, State Constitutions in the United States Federal System, 77 Ohio St. L.J. 195, 197-99 (2016) (symposium introduction). Judge Jeffrey S. Sutton of the United States Court of Appeals for the Sixth Circuit recently distilled the strongest arguments for state constitutionalism in a widely praised 30See William H. Pryor, Jr., The Importance of State Constitutions, Nat’l Rev. (June 7, 2018), https://www.nationalreview.com/2018/06/state-constitutions-important-components-of-federalism/ (Judge William H. Pryor, Jr., Circuit Judge for the Eleventh Circuit of the United States Court of Appeals); John Paul Stevens, The Other Constitutions, N.Y. Rev. Books (Dec. 2018), https://www.nybooks.com/articles/2018/12/06/other-constitutions/ (former Associate Justice of the Supreme Court John Paul Stevens). The book also features blurbs from such diverse legal thinkers as Laurence Tribe and William Baude. book. 31Sutton, supra note 25. First, state courts do not face the same practical constraints as the United States Supreme Court. 32Id. at 16. Unlike the United States Supreme Court, which sets rules applicable to all fifty states and the over 300 million people in them, state courts’ decisions have less reach. 33Id. at 16. While the United States Supreme Court is constrained by these considerations, and therefore less likely to adopt innovative and dynamic constitutional holdings, state courts are in an opposite position. 34Id. at 16–17. Sutton also notes that the United States Supreme Court may face a “federalism discount,” or the underenforcement of federal constitutional law, because it must be applied to all fifty states; this is a concern not present for states. Id. at 17; see also Brennan, supra note 28, at 549. Second, flawed state court decisions are easier to correct because of judicial elections and simpler constitutional amendment processes. 35Sutton, supra note 25, at 18. This also is the heart of the book’s title. As Judge Sutton says, “it may be more appropriate to tolerate fifty-one imperfect solutions rather than to impose one imperfect solution on the country as a whole.” Id. at 19; see also Brennan, supra note 28, at 551. State courts are also in a position to allow local conditions and traditions, including culture, geography, and history, to influence their holdings, allowing them to be more tailored to the relevant state. 36Sutton, supra note 25, at 17–18. Third, various state courts taking different approaches to legal issues produces multiple solutions to similar problems, creating a market from which federal courts can draw on after seeing the strengths and weaknesses of state experiments. 37Id. at 19–20, 178. Finally, state constitutionalism is agreeable to proponents of all schools of judicial thought, as the concept itself can result in the protection of rights favored by both. 38Id. at 211–12. Sutton maintains that state constitutionalism is compatible with all forms of judicial interpretation but criticizes Justice Brennan’s argument because it encouraged state courts to diverge from federal courts “so long as there is a progressive will.” Id. at 177. In Sutton’s view, this, combined with Brennan’s liberalism, may have alienated conservative judges and commentators at the time. Id. at 176–77. Regardless of the historical veracity of this, state constitutionalism is compatible with the textual originalism that is typically associated with conservative legal thinkers. See Jeffrey S. Sutton, 51 Imperfect Solutions: States and the Making of American Constitutional Law, 132 Harv. L. Rev. 811, 815–18 (2018) (book review); see also Brennan, supra note 28, at 550.

This Comment is informed by those arguments and seeks to find a role state courts can play in checking social media platforms’ unbridled power to limit its users’ political speech. 39 A state-based approach is particularly timely as former Attorney General Jeff Sessions recently met with a number of state attorneys general to discuss the topic. See Jessica Guynn, Sessions to Meet With State AGs on Charges Social Media Companies Stifle Free Speech, Competition, USA Today (Sept. 5, 2018, 10:51 PM), https://usatoday.com/story/tech/news/2018/09/05/sessions-meet-state-attorneys-general-charges-social-media-companies-hurt-free-speech/1204915002/. To do so, Part I will explore why state constitutionalism is a worthwhile approach for this issue specifically. It will first look to the text of the fifty state constitutional free speech clauses, nearly all of which have significant differences from the First Amendment. 40 Although state courts are free to interpret textually identical clauses differently from the corresponding United States Constitution clause, see Sutton, supra note 25, at 174–75, textual differences provide a strong reason to do so. Then, it will review a United States Supreme Court case that expressly approved state courts interpreting state constitution free speech clauses as broader than the First Amendment. 41 Bock v. Westminster Mall Co., 819 P.2d 55, 58 (Colo. 1991) (en banc). Part II will review case law from states that have adopted and applied these broad readings and applied the clauses to certain private actors. Part III will combine themes from this case law to develop an approach state courts can apply to social media platforms to protect political speech online from arbitrary viewpoint censorship. Finally, Part IV will briefly touch on practical matters like personal jurisdiction and Section 230 of the Communications Decency Act, noting that further scholarship and litigation advancing this Comment’s approach are worthwhile.

I. Why State Constitutions Are Worthy Vehicles for Protecting Free Speech Online

While the prenominate arguments for state constitutionalism are persuasive generally, they are relevant here specifically for two reasons: First, nearly all state constitution free speech clauses have significant textual differences from the First Amendment, and these differences provide a basis for going beyond a strict state action requirement—or, at least, they are not a bar to it. Second, the United States Supreme Court has expressly endorsed state courts interpreting state free speech clauses as broader than the First Amendment, including by applying those clauses to private actors.

A. The Text of the Fifty State Free Speech Clauses

It might be dubious to argue that state courts interpreting state free speech clauses should abandon First Amendment jurisprudence if the two were textually identical, or even substantially similar. 42 Though, of course, a state court would be perfectly within its right to do this. See Sutton, supra note 25, at 174–75. That is not, however, the case, as the following survey of state free speech clauses shows. In almost all state constitutions, there are textual differences from the First Amendment that provide a sound starting point for state courts to reconsider the extent of their state free speech protections. 43 Of course, each state supreme court must still engage in whatever brand of constitutional analysis required by its precedent. For examples of how this might look, see Maxim Cabaret, Inc. v. City of Sandy Springs, 816 S.E.2d 31, 39–40 (Ga. 2018) (Peterson, J., concurring), which addresses Georgia’s free speech clause specifically, and Elliott v. State, 824 S.E.2d 265, 267–70, 273 (Ga. 2019) and Olevik v. State, 806 S.E.2d 505, 514–15 (Ga. 2017), which address state constitutional interpretation generally and at length.

The First Amendment states: “Congress shall make no law … abridging the freedom of speech, or of the press, or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.” 44 U.S. Const. amend. I. State constitutions, by and large, take a different tack. Only two track the First Amendment, 45Haw. Const. art. I, § 4; S.C. Const. art. I, § 2. and only five others have text that similarly implicate only state action without separately enumerating an affirmative free speech right. 46Ind. Const. art. I, § 9; Or. Const. art. I, § 8; R.I. Const. art. I, § 21; Utah Const. art. I, § 15; W. Va. Const. art. III, § 7. The remaining forty-three state free speech clauses are dramatically different. With grammatical and slight wording variations, thirty-two free speech clauses follow language along these lines: “Every person may freely speak, write, and publish sentiments on all subjects, being responsible for the abuse of this right.” 47Ala. Const. art. I, § 4; Alaska Const. art. I, § 5; Ariz. Const. art. II, § 6; Cal. Const. art. I, § 2, cl. a; Colo. Const. art. II, § 10; Conn. Const. art. I, §§ 4–5; Fla. Const. art. I, § 4; Ga. Const. art. I, § 1, para. 5; Idaho Const. art. I, § 9; Ill. Const. art. I, § 4; Iowa Const. art. I, § 7; Kan. Const. Bill of Rights § 11; La. Const. art. I, § 7; Me. Const. art. I, § 4; Mich. Const. art. I, § 5; Minn. Const. art. I, § 3; Mo. Const. art. I, § 8; Mont. Const. art. II, § 7; Neb. Const. art. I, § 5; Nev. Const. art. I, § 9; N.J. Const. art. I, para. 6; N.M. Const. art. II, § 17; N.Y. Const. art. I, § 8; N.D. Const. art. I, § 4; Ohio Const. art. I, § 11; Okla. Const. art. II, § 22; S.D. Const. art. VI, § 5; Tex. Const. art. I, § 8; Va. Const. art. I, § 12; Wash. Const. art. I, § 5; Wis. Const. art. I, § 3; Wyo. Const. art. I, § 20. Along these lines, many state constitutions include language that expressly protects defendants in libel prosecutions—both private and public—when they can prove that their statements were true. For purposes of this analysis, this aspect of these provisions does not require detailed accounting. Twenty of those thirty-two states also include a phrase similar to: “No law shall be passed impairing 48 Some constitutions use words other than “impair” like “abridge” or “curtail” or “restrain” or “curtail or restrain.” For purposes of this analysis, these variations are not important. the freedom of speech.” 49Ala. Const. art. I, § 4; Cal. Const. art. I, § 2, cl. a; Colo. Const. art. II, § 10; Conn. Const. art. I, §§ 4–5; Fla. Const. art. I, § 4; Ga. Const. art. I, § 1, para. 5; Iowa Const. art. I, § 7; La. Const. art. I, § 7; Mich. Const. art. I, § 5; Mo. Const. art. I, § 8; Mont. Const. art. II, § 7; Nev. Const. art. I, § 9; N.J. Const. art. I, para. 6; N.M. Const. art. II, § 17; N.Y. Const. art. I, § 8; Ohio Const. art. I, § 11; Okla. Const. art. II, § 22; Tex. Const. art. I, § 8; Va. Const. art. I, § 12; Wis. Const. art. I, § 3. Another six states include the “every person” phrase, but have notable differences from this template. 50Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Ky. Const. Bill of Rights § 8; Md. Const. Declaration of Rights art. 40; Pa. Const. art. I, § 7; Tenn. Const. art. I, § 19. Additionally, five states do not include the “every person” sentence, but nevertheless differ from the First Amendment. 51Mass. Const. pt. I, art. 16; Miss. Const. art. III, § 13; N.H. Const. pt. I, art. 22; N.C. Const. art. I, § 14; Vt. Const. ch. I, art. XIII.

These textual differences are important because, on their face, they do not necessitate a state action requirement, much less a strict one. Moreover, they do not necessitate that state courts follow First Amendment jurisprudence—and indeed they may mandate that state courts deviate from such case law. 52See Maxim Cabaret, Inc. v. City of Sandy Springs, 816 S.E.2d 31, 39 (Ga. 2018) (Peterson, J., concurring). To begin with, the template found in thirty-two state constitutions includes two separate clauses. One states a broad right: “Every person may freely speak, write, and publish sentiments on all subjects, being responsible for the abuse of this right.” 53See supra note 47. The other imposes a restriction specifically on the government: “No law shall be passed impairing the freedom of speech.” 54See supra note 49. Unlike the First Amendment, which includes only the negative “no law” command, this approach also includes an affirmative declaration that every person has a free speech right. 55See Curious Theatre Co. v. Colo. Dep’t of Pub. Health & Env’t, 220 P.3d 544, 551 (Colo. 2009); Bock v. Westminster Mall Co., 819 P.2d 55, 57–58 (Colo. 1991) (en banc). This affirmative clause can be naturally read as expanding the free speech right beyond that in the First Amendment and potentially to private actors. 56See Golden Gateway Ctr. v. Golden Gateway Tenants Ass’n, 29 P.3d 797, 817–19 (Cal. 2001) (Werdegar, J., dissenting) (“[T]he latter proviso neither grammatically nor legally qualifies the simple and sweeping free speech guarantee with which [the] section … begins.”).

On the other hand, it can be argued that the negative clause defines the affirmative clause, creating a state action requirement, or that it otherwise implicitly suggests framer intent to reach only state action. 57See id. at 804 (plurality opinion). It is worth noting that questions of framer intent present a complicated issue of constitutional interpretation. It may be that such intent, unless unequivocally clear, should not override the plain text of these state free speech clauses. See Cologne v. Westfarms Assocs., 469 A.2d 1201, 1215–16 (Conn. 1984) (Peters, J., dissenting) (“These principles of construction would require … the intent of the draftsmen would not supersede the clear language of the constitution itself unless the intent of the framers were indisputably clear with respect to the particular provision arguably subject to implied limitation.”). It may also be that what the framers intended the words of a clause to mean is only a data point considered in a larger effort to discern the objective public meaning at the time of ratification. See Olevik v. State, 806 S.E.2d 505, 514–15 (Ga. 2017). And, such extratextual intent, if relied upon, would appropriately be discerned by reference to the relevant constitutional convention. See Golden Gateway Ctr., 29 P.3d at 804–06 (plurality opinion). This Comment takes no position on these questions of constitutional interpretation, but merely flags the issue. But, even assuming the validity of this argument, grammatical variations among state constitutions limit its applicability. While thirty-two state constitutions follow the prenominate template, eight include the word “and” between the affirmative and negative clauses 58Mich. Const. art. I, § 5; Nev. Const. art. I, § 9; N.M. Const. art. II, § 17; N.Y. Const. art. I, § 8; Ohio Const. art. I, § 11; Okla. Const. art. II, § 22; Tex. Const. art. I, § 8; Wis. Const. art. I, § 3. Connecticut’s Constitution separates the two clauses entirely, putting “No law shall ever be passed to curtail or restrain the liberty of speech or of the press” in a separate section, Conn. Const. art. I, § 5, and the two have been interpreted distinctly. See Cologne, 469 A.2d at 1208–09. Therefore, this survey does not include Connecticut’s free speech clause in this category. and twelve have only the affirmative clause with no reference to the state. 59Alaska Const. art. I, § 5; Ariz. Const. art. II, § 6; Idaho Const. art. I, § 9; Ill. Const. art. I, § 4; Kan. Const. Bill of Rights § 11; Me. Const. art. I, § 4; Minn. Const. art. I, § 3; Neb. Const. art. I, § 5; N.D. Const. art. I, § 4; S.D. Const. art. VI, § 5; Wash. Const. art. I, § 5; Wyo. Const. art. I, § 20. The section of Maine’s Constitution that includes the free speech clause does have negative “no laws” language, but it is specific only to the separate “freedom of the press” clause in that section. Me. Const. art. I, § 4; cf. City of Portland v. Jacobsky, 496 A.2d 646, 648 (Me. 1985) (quoting only the free speech clause without the negative language in an obscenity case).

The “and” is important: It separates, and thus distinguishes, the two clauses. This confers a general, affirmative free speech right on top of an additional right against government infringement. 60See supra note 56. Moreover, those twelve state free speech clauses making no reference to the state or law seem, on their face, to have no state action requirement. The plain effect of including only an affirmative free speech right is the removal of the state action requirement. 61See Southcenter Joint Venture v. Nat’l Democratic Policy Comm., 780 P.2d 1282, 1296 (Wash. 1989) (en banc) (Utter, J., concurring in the result) (“[F]rom a purely linguistic point of view, removing … language referring specifically to acts of state has great effect. Quite simply, it changes the facial meaning of the provision to state it in the absolute rather than in terms of state action.”); cf. Sharrock v. Dell Buick-Cadillac, Inc., 379 N.E.2d 1169, 1173–74 (N.Y. 1978) (“In contrast to the due process clause of the Fourteenth Amendment … Conspicuously absent from the [New York] State Constitution is any language requiring State action before an individual may find refuge in its protections. That is not to say, of course, that the due process clause of the State Constitution eliminates the necessity of any State involvement …. Rather, the absence … simply provides a basis to apply a more flexible State involvement requirement than … the Federal provision.”).

Similarly, six states include the “every person” language and deviate from the template in a way suggesting no state action requirement. 62Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Ky. Const. Bill of Rights § 8; Md. Const. Declaration of Rights art. 40; Pa. Const. art. I, § 7; Tenn. Const. art. I, § 19. For example, Pennsylvania’s free speech provision states:

The printing press shall be free to every person who may undertake to examine the proceedings of the Legislature or any branch of government, and no law shall ever be made to restrain the right thereof. The free communication of thoughts and opinions is one of the invaluable rights of man, and every citizen may freely speak, write and print on any subject, being responsible for the abuse of that liberty. 63Pa. Const. art. I, § 7.

Thus, the negative command is connected directly only to the printing press provision. 64See id. Tennessee’s and Kentucky’s free speech clauses also follow this language. 65Ky. Const. Bill of Rights § 8; Tenn. Const. art. I, § 19. Arkansas’s, Delaware’s, and Maryland’s follow similar language, but differ by making no reference to state action regarding the printing press. 66Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Md. Const. Declaration of Rights art. 40. Therefore, of these six states’ free speech clauses, three include only an affirmative right 67See supra note 66. and the other three have a state action reference only regarding free printing presses. 68See supra notes 63 and 65.

Five states have free speech provisions that differ entirely from the above structure, but nevertheless lack a facial state action requirement. New Hampshire’s simply states: “Free speech and Liberty of the press are essential to the security of Freedom in a State: They ought, therefore, to be inviolably preserved.” 69N.H. Const. pt. I, art. XXII. Massachusetts’s is similar: “The liberty of the press is essential to the security of freedom in a state: it ought not, therefore, to be restrained in this commonwealth. The right of free speech shall not be abridged.” 70Mass. Const. pt. I, art. XVI, amended by Mass. Const. art. LXXVII. North Carolina’s states: “Freedom of speech and of the press are two of the great bulwarks of liberty and therefore shall never be restrained, but every person shall be held responsible for their abuse.” 71N.C. Const. art. I, § 14. Mississippi’s likewise makes no reference to laws: “The freedom of speech and of the press shall be held sacred ….” 72Miss. Const. art. III, § 13. Finally, Vermont’s makes no express reference to state action, but appears limited to speech regarding government: “That the people have a right to freedom of speech, and of writing and publishing their sentiments, concerning the transactions of government, and therefore the freedom of the press ought not to be restrained.” 73Vt. Const. ch. I, art. 13. None of these clauses make reference to state action, and to interpret them as doing so requires either legislative history (in this sense, from the relevant constitutional convention(s)) showing that the framers intended verbs like “restrained” to be applicable implicitly only to the government, or otherwise an assumption thereof. 74See Golden Gateway Ctr. v. Golden Gateway Tenants Ass’n, 29 P.3d 797, 804–06 (Cal. 2001) (plurality opinion).

Although each state constitution free speech clause requires its own tailored analysis, as a general matter, the differences noted above show that many states—at least twenty-three 75Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Ky. Const. Bill of Rights § 8; Md. Const. Declaration of Rights art. 40; Pa. Const. art. I, § 7; Tenn. Const. art. I, § 19.; N.H. Const. pt. I, art. XXII; Mass. Const. pt. I, art. XVI, amended by Mass. Const. art. LXXVII; N.C. Const. art. I, § 14; Miss. Const. art. III, § 13; Vt. Const. ch. I, art. 13; Alaska Const. art. I, § 5; Ariz. Const. art. II, § 6; Idaho Const. art. I, § 9; Ill. Const. art. I, § 4; Kan. Const. Bill of Rights § 11; Me. Const. art. I, § 4; Minn. Const. art. I, § 3; Neb. Const. art. I, § 5; N.D. Const. art. I, § 4; S.D. Const. art. VI, § 5; Wash. Const. art. I, § 5; Wyo. Const. art. I, § 20. and up to forty-three 76Ala. Const. art. I, § 4; Alaska Const. art. I, § 5; Ariz. Const. art. II, § 6; Cal. Const. art. I, § 2, cl. a; Colo. Const. art. II, § 10; Conn. Const. art. I, §§ 4–5; Fla. Const. art. I, § 4; Ga. Const. art. I, § 1, para. 5; Idaho Const. art. I, § 9; Ill. Const. art. I, § 4; Iowa Const. art. I, § 7; Kan. Const. Bill of Rights § 11; La. Const. art. I, § 7; Me. Const. art. I, § 4; Mich. Const. art. I, § 5; Minn. Const. art. I, § 3; Mo. Const. art. I, § 8; Mont. Const. art. II, § 7; Neb. Const. art. I, § 5; Nev. Const. art. I, § 9; N.J. Const. art. I, para. 6; N.M. Const. art. II, § 17; N.Y. Const. art. I, § 8; N.D. Const. art. I, § 4; Ohio Const. art. I, § 11; Okla. Const. art. II, § 22; S.D. Const. art. VI, § 5; Tex. Const. art. I, § 8; Va. Const. art. I, § 12; Wash. Const. art. I, § 5; Wis. Const. art. I, § 3; Wyo. Const. art. I, § 20; Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Ky. Const. Bill of Rights § 8; Md. Const. Declaration of Rights art. 40; Pa. Const. art. I, § 7; Tenn. Const. art. I, § 19; Mass. Const. pt. I, art. 16; Miss. Const. art. III, § 13; N.H. Const. pt. I, art. 22; N.C. Const. art. I, § 14; Vt. Const. ch. I, art. XIII. —have a sound textual 77 Whether states with free speech clauses textually similar to the First Amendment may nevertheless choose to expand those clauses—based on evolving modern standards or otherwise—to private actors is another question, and the arguments for state constitutionalism in the introduction are applicable. basis to interpret their free speech clauses not just differently than the First Amendment but as broader than applying only to state actors.

B. The U.S. Supreme Court has Approved State Courts Applying State Constitution Free Speech Clauses to Private Actors

The United States Supreme Court has been consistent—with the exception of narrow and rare situations 78See Jackson, supra note 9, at 142–53 for a discussion of the exceptions to the state action doctrine.—in requiring state action before the First Amendment is invoked. 79See Hudgens v. NLRB, 424 U.S. 507, 513, 520–21 (1976) (“It is, of course, a commonplace that the constitutional guarantee of free speech is a guarantee only against abridgment by government, federal or state.”) (holding that the First Amendment does not apply to private shopping centers); Tsao v. Desert Palace, Inc., 698 F.3d 1128, 1140 (9th Cir. 2012) (noting the four narrow exceptions as the public function test, joint action test, state compulsion test, and governmental nexus test). See generally David M. Howard, Rethinking State Inaction: An In-Depth Look at the State Action Doctrine in State and Lower Federal Courts, 16 Conn. Pub. Int. L.J. 221, 221–40 (2017) (discussing state action doctrine and situations when private action becomes state action). But in Pruneyard Shopping Center v. Robins80 447 U.S. 74 (1980). all nine Justices joined in core reasoning expressly holding state courts can interpret state constitution free speech clauses as broader than the First Amendment, including applying to private actors, without violating the United States Constitution. 81Id. at 88. Also, the Court has denied petitions for certiorari in a number of subsequent cases raising the same or similar issues. See Ralphs Grocery Co. v. United Food & Commercial Workers Union Local 8, 290 P.3d 1116 (Cal. 2012), cert. denied, 133 S. Ct. 2799 (2013); Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742 (Cal. 2007), cert. denied, 555 U.S. 819 (2008); N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757 (N.J. 1994), cert. denied sub nom. Short Hills Assocs. v. N.J. Coal. Against War in the Middle E., 516 U.S. 812 (1995).

Pruneyard involved a privately owned large shopping center 82 For consistency, when this Comment discusses cases involving shopping centers or otherwise references shopping centers, it is referring to centers with size comparable to the one involved in Pruneyard, unless expressly noted. in California. 83Pruneyard Shopping Ctr., 447 U.S. at 77. No negligible mom-and-pop strip mall, it covered around twenty-one acres; had numerous stores, restaurants, and a movie theatre; and attracted 25,000 people daily. 84Id. at 77–78. Its policy regarding publicly expressive activity was a blanket prohibition, enforced in a nondiscriminatory way. 85Id. at 77. When a group of high school students set up a table in the shopping center’s central courtyard to distribute pamphlets, they were asked to leave despite being orderly and peaceful and garnering no complaints from patrons. 86Id. They complied, and promptly filed a lawsuit to enjoin the shopping center from denying them access to the property. 87Id.

The Supreme Court of California held 88 This decision’s reasoning is discussed below. See infra Part II and accompanying notes. that the California free speech clause included protection for people reasonably exercising speech in shopping centers, even when those centers are privately owned and the owners object. 89Pruneyard Shopping Ctr., 447 U.S. at 78 (citing Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979)). The shopping center appealed, arguing that the California court’s ruling violated its federal constitutional rights against takings under the Fifth Amendment, deprivation of property without due process under the Fourteenth Amendment, and free speech under the First Amendment. 90Id. at 82, 85. The United States Supreme Court rejected each of these arguments.

First, there was no taking because there was nothing to suggest the speech activity at issue would unreasonably impair the value of the shopping center’s property. 91Id. at 83. Moreover, the shopping center retained the ability to impose reasonable time, place, and manner restrictions to minimize any interference with its commercial functions. 92Id. Ultimately, the “right to exclude” was not “so essential to the use or economic value” of the shopping center’s property that the California court’s ruling was a taking. 93Id. at 84. Second, the shopping center’s due process argument failed because the California court’s ruling was not unreasonable, arbitrary, or capricious, and the state’s interest in promoting expansive free speech protections had a substantial relation to the means selected. 94Id. at 84–85.

Third, the shopping center’s First Amendment argument failed. Among multiple reasons for this failure, the most important was that its business was open to the public, making it unlikely that the views of people distributing literature would be identified as those of the shopping center. 95Id. at 87. Moreover, the government was not dictating that the shopping center display any particular message. 96Id. And, in any event, the center could easily disavow any message being expressed. 97Id. The Court also distinguished Pruneyard from a prior case, Miami Herald Publishing Co. v. Tornillo98 418 U.S. 241 (1974). which involved a state statute requiring a newspaper to publish a political candidate’s response to criticisms in that same newspaper. 99Pruneyard Shopping Ctr., 447 U.S. at 88. Whereas Tornillo involved an intrusion into the function of editors and deterred them from publishing controversial political statements, those concerns were not present in Pruneyard100Id.

The big takeaway from Pruneyard is that state courts are free under the U.S. Constitution to expand their constitutions’ free speech clauses to private actors, so long as they do not run afoul of any federal constitutional rights. 101 It is worth noting that Pruneyard was the culmination of a line of U.S. Supreme Court cases that first adopted reasoning similar to the reasoning of state courts discussed in Part II, see Amalgamated Food Emps. Union Local 590 v. Logan Valley Plaza, Inc., 391 U.S. 308, 319–20 (1968), then narrowed it, see Lloyd Corp. v. Tanner, 407 U.S. 551, 561–62 (1972), then overruled it, see Hudgens v. NLRB, 424 U.S. 507, 518 (1976). Because this Comment is focused on state constitutions, which need not be interpreted similarly to the First Amendments, a detailed discussion of these cases and the First Amendment would muddy the waters. For such a discussion, see Selznick & LaMacchia, supra note 9, at 252–54. As the Colorado Supreme Court has said, First Amendment jurisprudence after Pruneyard is “not the United States Supreme Court’s last word on free speech in the several states. The definitive word [is] left to the state courts to write.” 102 Bock v. Westminster Mall Co., 819 P.2d 55, 58 (Colo. 1991) (en banc).

 

II. Four Main Themes Across the Relevant State Case Law

This Part will explore the state case law emanating from Pruneyard by extracting the recurring themes across states that have applied their free speech clauses to private actors. Most states have declined to do so, but New Jersey, 103 Dublirer v. 2000 Linwood Ave. Owners, Inc., 103 A.3d 249 (N.J. 2014); Green Party of N.J. v. Hartz Mountain Indus., Inc., 752 A.2d 315 (N.J. 2000); N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757 (N.J. 1994), cert. denied sub nom. Short Hills Assocs. v. N.J. Coal. Against War in the Middle E., 516 U.S. 812 (1995); State v. Schmid, 423 A.2d 615 (N.J. 1980), appeal dismissed sub nom. Princeton Univ. v. Schmid, 455 U.S. 100 (1982). California, 104 Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742 (Cal. 2007), cert. denied, 555 U.S. 819 (2008); Robins v. Pruneyard Shopping Ctr., 592 P.2d 341 (Cal. 1979). Washington, 105 Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 533 (Wash. 1999) (en banc) (relying on state constitution’s initiative clause); Southcenter Joint Venture v. Nat’l Democratic Policy Comm., 780 P.2d 1282, 1289, 1291 (Wash. 1989) (en banc) (rejecting Alderwood plurality); Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108 (Wash. 1981) (en banc) (plurality opinion). Colorado, 106Bock, 819 P.2d at 58. and Pennsylvania, 107 W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331 (Pa. 1986); Commonwealth v. Tate, 432 A.2d 1382 (Pa. 1981). at some point, have. 108 At least two states have, at some point, found protections similar to those in Pruneyard, except under state constitution clauses other than free speech. See Glovsky v. Roche Bros. Supermarkets, Inc., 17 N.E.3d 1026 (Mass. 2014); Batchelder v. Allied Stores Int’l, Inc., 445 N.E.2d 590 (Mass. 1983); Lloyd Corp. v. Whiffen, 849 P.2d 446 (Or. 1993), abrogated by Stranahan v. Fred Meyer, Inc., 11 P.3d 228, 243 (Or. 2000). Washington has done this also and is addressed in the body. Additionally, Massachusetts has intimated as much, but not reached the issue. 109See Roman v. Trs. of Tufts Coll., 964 N.E.2d 331, 338 (Mass. 2012) (noting that state constitutions can provide more expansive protections of rights than the federal constitution and that the Massachusetts Supreme Court has approvingly cited cases from other jurisdictions protecting free speech rights on private colleges).

Case law from these states has four main themes: (1) courts focus on protecting political speech, treating it differently than commercial speech; (2) courts often rely on a “functional equivalent of a public forum” argument to justify their holdings by emphasizing changing empirical trends in where speech occurs; (3) courts balance interests to weigh free speech rights against the property rights of the private actor; and (4) to maintain an appropriate balance, courts emphasize the right of the private actor to implement reasonable time, place, and manner restrictions on the protected speech. This Part addresses each in turn.

A. Political Speech Is Protected

The relevant state cases have invariably involved political speech. The form of this speech has typically come either in distributing literature 110See, e.g., Pruneyard Shopping Ctr. v. Robins, 447 U.S. 74, 77 (1980); Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 744 (Cal. 2007); Bock, 819 P.2d at 56; Green Party of N.J. v. Hartz Mountain Indus., Inc., 752 A.2d 315, 319 (N.J. 2000); N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 775 (N.J. 1994), cert. denied sub nom. Short Hills Assocs. v. N.J. Coal. Against War in the Middle E., 516 U.S. 812 (1995); State v. Schmid, 423 A.2d 615, 616 (N.J. 1980), appeal dismissed sub nom. Princeton Univ. v. Schmid, 455 U.S. 100 (1982); Tate, 432 A.2d at 1385. or collecting signatures for petitions or initiatives. 111See, e.g., W. Pa. Socialist Workers, 515 A.2d at 1333; Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 525 (Wash. 1999) (en banc); Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108, 110 (Wash. 1981) (plurality opinion). The subject matter has ranged from opposition to a United Nations resolution against Zionism 112 Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 342 (Cal. 1979). to protesting war 113N.J. Coal., 650 A.2d at 762; Tate, 432 A.2d at 1384; see also Bock, 819 P.2d at 56 (“[P]ledging non-violent dissent from the federal government’s foreign policy toward Central America.”). to boycotting 114Fashion Valley Mall, 172 P.3d at 743. to electoral campaigning. 115Green Party of N.J., 752 A.2d at 319; Schmid, 423 A.2d at 616; see also Dublirer v. 2000 Linwood Ave. Owners, Inc., 103 A.3d 249, 250 (N.J. 2014) (campaigning for the board of a private cooperative apartment). These holdings seem limited to political speech, both because of the fact patterns and the language used in the reasoning. Indeed, the Supreme Court of New Jersey explicitly said so when applying its free speech clause to a large shopping center: “Our holding is limited to … associated speech in support of, or in opposition to, causes, candidates, and parties—political and societal free speech.” 116N.J. Coal., 650 A.2d at 781. In the same case, the court emphasized that the speech interest involved was “the most substantial in our constitutional scheme.” Id. at 776. Also, in a previous case applying its free speech clause to a private college, the Supreme Court of New Jersey noted that political speech is particularly important and worthy of protection. See Schmid, 423 A.2d at 627. The Supreme Court of New Jersey reiterated this sentiment in a subsequent case. See Green Party of N.J., 752 A.2d at 328. Similarly, in applying its free speech clause to a private college, the Supreme Court of Pennsylvania specified that it was the “absolutely fundamental rights of the public to freedom of political speech” that were implicated in the case. 117Tate, 432 A.2d at 1390 (emphasis added). The Colorado Supreme Court, in applying its free speech clause to a shopping center, also defined the speech interests around political speech: “Free political speech, such as that involved in this case, occupies a preferred position in this country and this state.” 118Bock, 819 P.2d at 57. Moreover, this free speech right holds “high rank … in the constellation of freedoms guaranteed by both the United States Constitution and [Colorado’s] state constitution.” Id. The Supreme Court of California, when first applying its free speech clause to a shopping center, coupled that clause with its Constitution’s right to petition clause, implying that the political nature of the involved speech was crucial. 119 Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979). The Supreme Court of California later relied on only the state’s free speech clause to hold a shopping center could not prohibit union related boycotting of its stores. Fashion Valley Mall, 172 P.3d at 754. The Washington Supreme Court followed similar reasoning in applying its state free speech clause to a shopping center, 120See Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108, 117 (Wash. 1981) (plurality opinion). and when that court later moved away from this holding, it still reached a similar conclusion under the state’s inherently political right to initiative clause. 121See Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 529–30 (Wash. 1999) (en banc); Southcenter Joint Venture v. Nat’l Democratic Policy Comm., 780 P.2d 1282, 1290 (Wash. 1989) (en banc).

This is the correct position, and political speech can be contrasted with commercial speech. 122 Generally speaking, commercial speech holds a position inferior to political speech under the First Amendment. See e.g., Hoffman v. Capital Cities/ABC, Inc., 255 F.3d 1180, 1184–85 (9th Cir. 2001) (“[Commercial speech] is entitled to a measure of First Amendment protection … however, [it] do[es] not receive the same level of constitutional protection as other types of protected expression.”). Notably, although content-based restrictions are more likely to survive the First Amendment when they affect commercial speech, the Supreme Court has been skeptical of viewpoint-based restrictions and has “specifically declined to recognize a distinction between commercial and noncommercial speech that would render” the “offensiveness” of speech a government interest sufficient to justify prohibiting it. Bolger v. Youngs Drug Prods. Corp., 463 U.S. 60, 71–72 (1983). See also Iancu v. Brunetti, 139 S.Ct. 2294 (2019); Matal v. Tam, 137 S.Ct. 1744 (2017). Forcing business property owners to allow commercial speech on their property would directly contravene the purpose of that property, likely failing any balancing test between the speech interest and the owners’ rights. 123See N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 781 (N.J. 1994) (“Commercial free speech at regional and community shopping centers is fundamentally so discordant with the purposes and uses of those centers as to disqualify it from constitutional protection.”). The specifics of a balancing test are discussed infra Part II.C. This becomes especially evident by imagining a speaker who enters one shopping center and encourages patrons to visit another. 124See N.J. Coal., 650 A.2d at 782 (“[C]ommercial free speech could obviously be directly in conflict with the centers’ activities, uses, and success, the most obvious example being leafletting seeking to persuade shoppers and non-shoppers to go elsewhere.”). In a more general sense, commercial speech can also disrupt the carefully planned advertising strategies of the property owner, even if the commercial speech is not intended to lure patrons away. 125See id. at 781 (“[Commercial speech] is generally discordant: the owners … carefully plan their merchandising strategy, their advertising programs, and are entitled to reap the rewards of their efforts without commercial interference, even well-intentioned commercial interference, from others.”). Even speech otherwise political but involving solicitation of funds or the sale of literature has been held unprotected by state courts otherwise willing to protect political speech in shopping centers. 126See id. at 782 (“[T]he sale of literature and the solicitation of funds on the spot (as distinguished from appeals found in the leaflets themselves) are not covered by the protection.”); see also Southcenter Joint Venture, 780 P.2d at 1306 (Utter, J., concurring in the result) (noting that “[t]he nature of this speech activity competed directly with the property interests of the mall owners” because it solicited contributions and literature sales, which distinguished it from the solicitation of signatures in Alderwood and made it incompatible with the uses of the mall and thus unprotected).

Political speech, on the other hand, generally does not raise these concerns. To the extent that it might disrupt ordinary business, this can be appropriately managed through time, place, and manner restrictions, as numerous court opinions have held. 127See, e.g., Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979); Bock v. Westminster Mall Co., 819 P.2d 55, 62–63 (Colo. 1991) (en banc); N.J. Coal., 650 A.2d at 783; State v. Schmid, 423 A.2d 615, 630 (N.J. 1980); Commonwealth v. Tate, 432 A.2d 1382, 1390 (Pa. 1981); Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108, 116–17 (Wash. 1981) (plurality opinion). Political speech also typically does not target directly the commercial purposes of the property. Fashion Valley Mall, LLC v. NLRB 128 172 P.3d 742 (Cal. 2007). involves a fine distinction between commercial and political speech and shows the importance of the categorization. There, the Supreme Court of California noted that a shopping center could prohibit conduct calculated to disrupt or unduly interfere with its normal business operations but held the center could not prohibit speech advocating boycott of its stores. 129Id. at 750. A boycott, of course, directly targets the property’s commercial purpose. But it also often targets some noncommercial aspect of the business and is a tactic of a cause connected to something political like labor unions—which was the case in Fashion Valley130See id. at 743–44. This fundamentally differs from speech that simply promotes a different business for commercially driven purposes.

B. Shopping Centers Are Defined as Functional Equivalents of Traditional Public Forums

Pervasive in the state case law is reasoning that the shopping centers at issue have become the functional equivalents of public forums. 131 This “functional equivalent” argument has also been central in cases using clauses other than free speech to similar effect. See Glovsky v. Roche Bros. Supermarkets, Inc., 17 N.E.3d 1026, 1031 (Mass. 2014); Batchelder v. Allied Stores Int’l, Inc., 445 N.E.2d 590, 597 (Mass. 1983); Lloyd Corp. v. Whiffen, 849 P.2d 446, 465 (Or. 1993), abrogated by Stranahan v. Fred Meyer, Inc., 11 P.3d 228 (Or. 2000). This reasoning stems from basic facts of economic and social change: As suburbs grew, so did shopping centers. As shopping centers grew, they attracted greater numbers of people, becoming increasingly important hubs in communities. And, importantly, this growth displaced traditional “downtown business districts” where much speech activity once took place and was constitutionally protected. 132See Alderwood Assocs., 635 P.2d at 114 (“Then, most public forums were located in downtown business districts, whereas now people spend more time in shopping centers than anywhere else, excepting their home and job.”). For a fun portrayal of this cultural phenomenon, see its depiction in season three of the mega-hit Netflix show Stranger Things, which take place in the 1980s. See Stranger Things (Netflix 2019). As the Supreme Court of New Jersey succinctly put it, “malls are where the people can be found today.” 133 N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 767 (N.J. 1994). Numerous opinions document this growth, relying on empirical data and common sense observations. 134See id. at 766–68 (detailing extensive empirical evidence and citing academic support); Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 344–45 (Cal. 1979); Bock v. Westminster Mall Co., 819 P.2d 55, 61 (Colo. 1991); W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331, 1336–37 (Pa. 1986); Alderwood Assocs., 635 P.2d at 114; see also James R. Spady, Free Speech, Initiative and Property Rights in Conflict—Four Alternatives to the State Action Requirement in Washington, 58 Wash. L. Rev. 587, 614–15 (1983). From this data, the courts figure, it follows that speech protections ought to—and perhaps must—extend to shopping centers for free speech to flourish, or perhaps even to meaningfully exist. 135See N.J. Coal., 650 A.2d at 777 (“[W]here private ownership of property that is the functional counterpart of the downtown business district has effectively monopolized significant opportunities for free speech, the owners cannot eradicate those opportunities by prohibiting it.”); Alderwood Assocs., 635 P.2d at 114 (“Because the nature of the public forum has changed, it is necessary to reassess the underpinnings for the existing free speech principles.”). Otherwise, constitutional free speech protection would lose much of its power as speech activity transitions from government-owned locations to private ones that nevertheless function like public forums. 136See N.J. Coal., 650 A.2d at 777.

Some courts have found the shopping centers at issue do not necessarily have a primarily commercial function, bolstering the “functional equivalent” reasoning. Instead, the centers open themselves up for people to visit and congregate for a variety of purposes, merely hoping that they also buy something, but not requiring it. This dichotomy between the commercial purpose being primary versus only ancillary is best shown by contrasting cases from New Jersey and Pennsylvania. In New Jersey Coalition Against War in the Middle East v. J.M.B. Realty Corp.137 650 A.2d 757 (N.J. 1994). the Supreme Court of New Jersey explicitly found the commercial purpose of the shopping center ancillary: “The invitation to the public is simple: ‘Come here, that’s all we ask. We hope you will buy, but you do not have to, and you need not intend to.’” 138Id. at 773. The construction design and uses of shopping centers made them “attractive to everyone, for all purposes, to make them a magnet for all people, not just shoppers” and the centers “carr[ied] the message that this is the place to be—this is your community, where you can rest, relax, talk, listen, be entertained and be educated.” 139Id. In other words, whether patrons actually engaged in any commercial activity was only secondary to the center’s primary goal of just having their presence. 140 Similar to the Supreme Court of New Jersey, the Colorado Supreme Court also found the commercial purposes of the shopping centers to not override its functioning as an equivalent of a public forum. See Bock, 819 P.2d at 62 (“[T]he historical connection between the marketplace of ideas and the market for goods and services is not severed because goods and services today are bought and sold within the confines of a modern mall …. The range of activities permitted in the common areas of the Mall also indicates the extent to which the Mall effectively functions as the latter-day public forum.”); see also Lewis v. Colo. Rockies Baseball Club, Ltd., 941 P.2d 266, 271 n.8 (Colo. 1997) (en banc) (“We found the concept of ‘public forum’ to be more expansive under the Colorado Constitution and held that the private shopping mall at issue was a public forum for free speech purposes.”). When California and Washington first addressed shopping centers, they did not clearly engage in comparing the commercial to non-commercial purposes of the centers but did rely on the functional equivalent argument. See Robins, 592 P.2d at 345–47; Alderwood Assocs., 635 P.2d at 116. And, though Washington has moved away from Alderwood’s analysis of the state free speech clause, it has relied on functional equivalent reasoning to reach similar results under its initiative clause. See Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 529–33 (Wash. 1999) (en banc).

The Supreme Court of Pennsylvania, however, rejected this argument in Western Pennsylvania Socialist Workers 1982 Campaign v. Connecticut General Life Insurance Co. 141 515 A.2d 1331 (Pa. 1986). Based on prior precedent 142See Commonwealth v. Tate, 432 A.2d 1382 (Pa. 1981). the court first held that the state’s free speech clause only applies when the property in question becomes the functional equivalent of a public forum. 143 W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331, 1336 (Pa. 1986). A shopping center, though, was not such a forum. 144Id. at 1337–38. Even though shopping centers had come to serve many functions downtown business districts once had, these social benefits were ancillary and their commercial purpose was primary and exclusive. 145Id. at 1336. As the Supreme Court of Pennsylvania put it, “[the shopping center] has invited the public at large into the mall only for commercial purposes …. [It] is operated as a market place for the exchange of goods and services but not as a market place for the exchange of ideas.” 146Id. at 1337. This holding was further based in the shopping center’s policy to uniformly prohibit all political speech, thus not giving unfair advantage to one view over others. 147See id. at 1333. For these reasons, the court rejected the applicability of the functional equivalent argument it had previously used to apply its free speech clause to a private college. 148Id. at 1337–39. Washington, relying on a functional equivalent of a public forum analysis under its initiative clause, has also used commercial purpose reasoning to distinguish a large grocery store from a shopping center and to hold the grocery store was not a functional equivalent. See Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 530 (Wash. 1999) (en banc).

A limitation flows from this functional equivalent reasoning: The state free speech clauses are less likely to apply to shopping centers that are not of enormous size. The Supreme Court of California has explicitly said this: “‘It bears repeated emphasis that we do not have under consideration the property or privacy rights of an individual homeowner or the proprietor of a modest retail establishment.’” 149 Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979). The Colorado Supreme Court subsequently cited this quote when applying its free speech clause to a shopping center. 150 Bock v. Westminster Mall Co., 819 P.2d 55, 62–63 (Colo. 1991). The Supreme Court of New Jersey has also explicitly limited a similar holding: “No highway strip mall, no football stadium, no theater, no single huge suburban store, no stand-alone use, and no small to medium shopping center” satisfies the conditions necessary to warrant application of New Jersey’s free speech clause. 151 N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 781 (N.J. 1994). But, that court also left the door open for applications to smaller shopping centers or types of property that “ha[ve] clearly and consistently invited or permitted issue-oriented groups, candidates, and others, to leaflet.” 152Id.

C. Courts Rely on Interest Balancing Because of Competing Property and Free Speech Rights

Given that the expansion of free speech rights in this context conflicts with the property rights of the owners of the property where the speech is occurring, courts resort to a balancing test to weigh the two. 153 The Supreme Court of Pennsylvania concisely captured the relevant balancing inquiry in holding: [I]n certain circumstances, the state may reasonably restrict the right to possess and use property in the interests of freedom of speech, assembly, and petition …. To determine whether the appropriate circumstances must exist here, we must balance the college’s right to possess and protect its property against appellants’ rights of expression in light of the compatibility of that expression with the “activity of (the) particular place at (the) particular time.” Commonwealth v. Tate, 432 A.2d 1382, 1390 (Pa. 1981) (citations omitted). Whether because the relevant state constitution clause lacks a state action requirement or because the private property’s nature makes it the functional equivalent of a public forum, and thus within state action, it becomes necessary to balance interests out of due process concerns. New Jersey and Washington have both articulated multi-factor balancing tests, making the case law from each state worth a close look.

New Jersey first expanded its free speech clause to a private college in State v. Schmid154 423 A.2d 615 (N.J. 1980). A fundamental precept was crucial for its holding. The text of New Jersey’s free speech clause is “more sweeping in scope than the language of the First Amendment,” 155Id. at 626–27. leading the Supreme Court of New Jersey to expressly hold its free speech clause has no state action requirement. 156Id. at 628 (“They are also available against unreasonably restrictive or oppressive conduct on the part of private entities that have otherwise assumed a constitutional obligation not to abridge the individual exercise of such freedoms [of speech and assembly] because of the public use of their property.”); see also N.J. Coal., 650 A.2d at 779 (“In New Jersey, we have an affirmative right of free speech, and neither government nor private entities can unreasonably restrict it.”).

In light of the lack of a state action requirement, the court devised a multi-factor test to determine whether its free speech clause applies to a given state actor: (1) the nature, purposes, and primary use of the property, or generally, its “normal” use; (2) the extent and nature of the public’s invitation to use the property; and (3) the purpose of the expressional activity in relation to both the private and public use of the property. 157Schmid, 423 A.2d at 630. The court then applied this test to the facts before it, which involved a political activist distributing literature on a private college campus. 158Id. at 617–18. First, the primary use of the property, being a college, was to pursue truth and foster learning. 159Id. at 630. Second, the public’s invitation was extensive, seeking to encourage a wide and continuous exchange of opinion, and this invitation was “entirely consonant” with the college’s education mission. 160Id. at 631. Third, there was no evidence of the involved speech activity disturbing that mission. 161Id. Therefore, the college had to allow the speech activity and violated the activist’s constitutional rights by preventing it. 162Id. at 632–33.

After Schmid, the New Jersey Supreme Court addressed whether its free speech clause extends to shopping centers. 163 N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757 (N.J. 1994). The court applied the three Schmid factors and added a general balancing test between the speech and private property rights at issue. 164Id. at 772 (“The test to determine the existence of the constitutional obligation is multi-faceted; the outcome depends on a consideration of all three [Schmid] factors … and ultimately on a balancing ….”). This application was simple enough. First, the normal use of the property showed its all-inclusiveness, and as discussed above, it was the functional equivalent of a public forum. 165Id. at 772–74. Second, the public’s invitation was extensive, essentially being to come to the premises to do whatever one wants, and to hopefully purchase something; and there was a further implicit invitation to distribute controversial political literature. 166Id. at 774 (“The almost limitless public use of defendant’s property … gives rise to an implied invitation of constitutional dimensions … that includes leafletting on controversial issues.”). Third, the speech at issue—distributing literature—was compatible with the property’s use, and the record showed no negative impact on the center’s finances. 167Id. at 775. Finally, a general balancing test favored the speech rights: While the shopping center’s private property interests were greatly diminished by how extensively it opened itself to the public, 168Id. at 776. The Supreme Court of New Jersey adopted a sliding scale approach, saying “as private property becomes, on a sliding scale, committed either more or less to public use and enjoyment, there is actuated, in effect, a counterbalancing between expressional and property rights.” Id. at 775 (quoting State v. Schmid, 423 A.2d 615 (N.J. 1980)). The court also quoted the famous passage from Marsh v. Alabama that “[t]he more an owner, for his advantage, opens up his property for use by the public in general, the more do his rights become circumscribed by the statutory and constitutional rights of those who use it.” Id. at 775–76 (quoting Marsh v. Alabama, 326 U.S. 501, 506 (1946)). the public’s speech interests were “the most substantial in our constitutional scheme.” 169Id. at 776. For these reasons, the Supreme Court of New Jersey extended its free speech clause to large regional shopping centers. 170Id.

A plurality of the Washington Supreme Court articulated a similar balancing test under its state free speech clause, though the full court has since rejected that approach while adopting the same reasoning under the state’s initiative clause initiative clause. 171Wash. Const. art. II, § 1(a) (detailing the initiative petition process by which the people can “reserve to themselves the power to propose bills, laws, and to enact or reject the same…”). Although this Comment does not ultimately pull directly from the case law around Washington’s initiative clause, that clause is noted here for two reasons. First, to provide a complete and accurate representation of the development of Washington’s case law, specifically its shift away from using the state’s free speech clause. Second, because it uses the same reasoning as the free speech cases, providing another data point for comparison to this Comment’s ultimate approach. In Alderwood Associates v. Washington Environmental Council172 635 P.2d 108 (Wash. 1981) (plurality opinion). a shopping center case, a plurality of Washington’s supreme court said its free speech clause does not require the same state action as the First Amendment because its text contains only an affirmative right and makes no reference to state action. 173Id. at 114–16. The Washington free speech clause reads: “Every person may freely speak, write and publish on all subjects, being responsible for the abuse of that right.” Id. at 114 (quoting Wash. Const. art. I, § 5). In place of a state action requirement was a balancing test. 174Id. at 116 (“[B]eing sensitive to the competing speech and property rights, we conclude [the speech and petition clauses] are applicable when, after balancing all the interests, the balance favors the speech and initiative activity.”). The court used four factors: (1) the use and nature of the private property; (2) the nature of the speech activity; (3) the potential for reasonable regulation—through time, place, and manner restrictions—of the speech; and (4) whether barring the speech activity would significantly undermine free speech and the effectiveness of the initiative process. 175Id. at 116–17. These factors tipped toward allowing the speech on the private property, especially in light of the shopping center having become the functional equivalent of a town center or public forum, which made suppression of the speech particularly harmful and made the property owner’s private autonomy interests minimal. 176Id. at 117. The plurality also noted that the state’s initiative clause bolstered its opinion because shopping centers had become a crucial location for gathering signatures. 177Id.

The Washington Supreme Court quickly moved away from the Alderwood plurality, however, plainly holding that the state’s free speech clause protected only against state action. 178See Southcenter Joint Venture v. Nat’l Democratic Policy Comm., 780 P.2d 1282, 1289–90 (Wash. 1989) (en banc) (“A 4-member plurality of this court, i.e., less than a majority of the court, maintained that there was no ‘state action’ requirement under the free speech and initiative provisions of the state constitution …. Thus, in Alderwood, a 5-member majority of this court rejected the argument … that the free speech provision of our state constitution does not require ‘state action.’”). But Alderwood’s holding has been retained to the extent it was based on Washington’s initiative clause and not its free speech clause. 179Id. at 1288–89. Accordingly, Alderwood’s reasoning and multi-factor test has continued to be applied, albeit slightly altered and under the initiative clause. 180 Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 529–30 (Wash. 1999) (en banc) (“[W]e concluded in Alderwood that [the initiative clause] specifically guarantees the right to gather signatures for initiatives at large shopping malls.”). The court also expressly refused to overturn Alderwood. Id. at 530 (“Alderwood … effectuates and complies with the express mandate of this court to preserve the people’s right to engage in the initiative process. To overturn Alderwood would substantially diminish the ability of Washington’s citizens to participate in the initiative process in a meaningful way.”). The Washington Supreme Court has eliminated the second and third Alderwood factors (nature of the speech activity and potential for regulation) as unhelpful in the initiative context. 181Id. At the same time, the court added a factor: The scope of the invitation the property owner has extended to the public. 182Id. Ultimately, the court concluded that the property in question was not affected by the initiative clause because it was not large enough; thus, its interests won the balancing test. Id. at 533.

Combining the New Jersey and Washington case law produces a synthesized set of factors: (1) the primary use and nature of the property, (2) the extent of the invitation to the public to use the property, 183 This factor is closely related with the functional equivalent of a public forum concept, as the latter helps define the former. (3) the compatibility of the involved speech with the property’s use, 184 This represents a combination of a Schmid and Alderwood factor. Alderwood terms it “the nature of the speech activity” and cites Schmid, but otherwise gives the factor scant discussion. See Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108, 116 (Wash. 1981) (plurality opinion). Schmid terms it “the purpose of the expressional activity undertaken upon such property in relation to both the private and public use of the property.” State v. Schmid, 423 A.2d 615, 630 (N.J. 1980). A later New Jersey case further articulates this factor, stating that it “examines the compatibility of the free speech sought to be exercised with the uses of the property.” N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 775 (N.J. 1994). and (4) the extent to which speech would be undermined if it were prohibited by the private property owner. 185See Schmid, 423 A.2d at 630; Waremart, Inc., 989 P.2d at 530; Alderwood Assocs., 635 P.2d at 116–17.

D. Courts Allow Property Owners to Enact Reasonable Time, Place, and Manner Restrictions

Related to the balancing of rights, each of the relevant cases also holds that private property owners retain the right to impose reasonable time, place, and manner restrictions 186See Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979); Bock v. Westminster Mall Co., 819 P.2d 55, 63 (Colo. 1991); N.J. Coal., 650 A.2d at 783; Schmid, 423 A.2d at 630; Commonwealth v. Tate, 432 A.2d 1382, 1390 (Pa. 1981) (quoting State v. Schmid, 423 A.2d 615, 630 (N.J. 1980)); Alderwood Assocs., 635 P.2d at 116 (citing Cox v. Louisiana, 379 U.S. 536, 544-55 (1965)).—and this is likely required by United States Supreme Court precedent. 187See Pruneyard Shopping Ctr. v. Robins, 447 U.S. 74, 83–84 (1980). It is clear that such restrictions are allowed so the property owner can prevent its business from being disrupted and suffering financial loss. 188See, e.g., id. at 83 (reasoning that there was no takings violation because “[the shopping center] may restrict expressive activity by adopting time, place, and manner regulations that will minimize any interference with its commercial functions.” (emphasis added)); Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 751 (Cal. 2007) (“[A] shopping center may prohibit conduct ‘calculated to disrupt normal business operations’ or that would result in ‘obstruction of or undue interference with normal business operations.’” (quoting Diamond v. Bland, 477 P.2d 733, 741–42 (Cal. 1970))); Green Party of N.J. v. Hartz Mountain Indus., Inc., 752 A.2d 315, 327–28 (N.J. 2000) (“… designed to ensure that time, place, and manner regulation should minimize any interference with the mall’s commercial function without denying the counterbalancing interest of leafletters in expressive speech.”). Less clear is what exactly these restrictions can require.

California and New Jersey present contrasting approaches—strict scrutiny versus multi-factor balancing—and comparing the two shows New Jersey’s approach is preferable. California’s supreme court has used strict scrutiny to strike down a shopping center’s content-based restrictions. 189See Fashion Valley Mall, 172 P.3d at 751, 754. When a restriction of speech activity is not content-neutral, the court said, it applies strict scrutiny to determine if the restriction is necessary to serve a compelling interest and is narrowly drawn to achieve it. 190Id. When the restriction is content-neutral, it applies intermediate scrutiny to determine if it is (1) narrowly tailored, (2) serves a significant government interest, and (3) leaves ample alternative avenues of communication. 191Id. at 751. In Fashion Valley Mall, LLC v. NLRB, a shopping center prohibited speech that advocated boycotts of the center’s stores, so the restrictions were not content-neutral, even if they were viewpoint-neutral. 192Id. at 751–52. The restriction was therefore subject to strict scrutiny, and failed to withstand the California Constitution’s protection of free speech in large shopping centers. 193Id. at 754. Simply, the center’s interest in maximizing profits was not compelling when weighed against the free speech rights involved. 194Id.

New Jersey’s supreme court, on the other hand, rejected a strict scrutiny framework and adopted a multi-factor balancing test. 195 Green Party of New Jersey v. Hartz Mountain Industries, Inc., 752 A.2d 315, 325–26 (N.J. 2000). Whereas that court has used strict scrutiny when government restricts free speech to protect private property interests, the situation in these cases is the opposite: government restricting private property interests to protect free speech. 196Id. at 326. As such, the Supreme Court of New Jersey has held a strict scrutiny framework does not “fit.” 197Id. at 325–26. Another possible approach—applying the business judgment rule 198Id. at 326–27.—has also been rejected by the court because this context does not involve shareholder-board relations. 199Id. at 326 (“Under the rule, when business judgments are made in good faith based on reasonable business knowledge, the decision makers are immune from liability from actions brought by others who have an interest in the business entity.”). Instead, a balancing of interests is appropriate, and the court articulated three factors: (1) the nature of the affected right; (2) the extent to which the restriction intrudes upon it; and (3) the center’s need for the restriction. 200Id. at 327. The court further articulated that “the more important the constitutional right sought to be exercised, the greater the [center’s] need must be to justify interference with the exercise of that right.” 201Id. Moreover, the restriction “should minimize any interference with the [center’s] commercial functions without denying the counterbalancing interest of leafleteers in expressive speech” and “be designed to achieve the [center’s] legitimate purposes while preserving the leafleteer’s expressive rights.” 202Id. at 327–28. With this approach, the court struck down a large shopping center’s regulations that required would-be leafleteers to have a million-dollar insurance policy and sign a hold harmless, and restricted access to only a few days a year. 203Id. at 317. The speech right prevailed over the regulations because of the importance of the political speech, the minimal or nonexistent risk to the center, and the lack of any indication that the center suffered economic loss because of the speech activity. 204Id. at 328–32.

 

The Supreme Court of New Jersey’s balancing approach is superior to the Supreme Court of California’s strict scrutiny framework. 205 A factual difference between the two cases is that the restriction the California court addressed was content-based, whereas the New Jersey court addressed a content-neutral restriction. The New Jersey court, however, made no mention of whether the restriction was content-based or content-neutral, and this consideration seems irrelevant to its reasoning. Strict scrutiny is out of place because the property owner’s interests are likely always limited to profit related concepts, whereas the government can have numerous and widespread policy goals. 206 For example, in Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 754 (Cal. 2007), the court says that the shopping center’s goal of maximizing profits is not compelling compared to a labor union’s right to free expression. It is difficult to imagine when or how a business property owner might have an interest other than maximizing profit, and harder still to imagine when this interest would prevail over political speech. As such, it seems too rigid a standard. New Jersey’s factors, though, allow for the constitutional rights of both parties to be considered when evaluating a given restriction by encompassing the relevant considerations.

Beyond surviving some sort of balancing test, time, place, and manner restrictions are more likely to face constitutional problems if they favor one viewpoint over another or lack objective standards. First, it seems unlikely that a viewpoint-based restriction would survive any of the cases discussed above, whether under California’s strict scrutiny approach or New Jersey’s balancing test. 207See, e.g., Fashion Valley Mall, 172 P.3d at 751; Golden Gateway Ctr. v. Golden Gateway Tenants Ass’n, 29 P.3d 797, 821–22 (Cal. 2001) (Werdegar, J., dissenting); Bock v. Westminster Mall Co., 819 P.2d 55, 62–63 (Colo. 1991). Even when a court held that shopping centers were not subject to the state’s free speech clause, the holding was premised on the center banning all political speech, and the court suggested that the center giving one viewpoint an unfair advantage would be unconstitutional. 208 W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331, 1336 (Pa. 1986) (“Nor does it yet require [shopping centers] to provide a political forum for persons or groups with views on public issues, so long as the owner does not grant unfair advantages to particular interests or groups by making his premises arbitrarily available to those he favors while excluding others.”). Second, courts have consistently been tougher on restrictions lacking objective standards because they are vulnerable to arbitrary application. 209Green Party of N.J., 752 A.2d at 329 (regulations must focus on objective criteria and avoid “general and subjective standard[s] vulnerable to arbitrary or content related determination.”); State v. Schmid, 423 A.2d 615, 632 (N.J. 1980) (“[P]rivate colleges … must be accorded a generous measure of autonomy and self-governance … In this case, however, the University regulations that were applied to Schmid … contained no standards ….”); Commonwealth v. Tate, 432 A.2d 1382, 1387 (Pa. 1981) (“The record does not reveal what standards, if any, were applicable to the granting or denial of such permission. It appears, however … the college believed itself entitled to exercise its discretion arbitrarily.”). A clear example of this is in Schmid, where the lack of objective standards was fatal to the private college, and the court noted that restrictions adopted after the incident giving rise to that case would withstand constitutional challenge. 210Schmid, 423 A.2d at 633 (“These current amended regulations exemplify the approaches open to private educational entities seeking to protect their institutional integrity while at the same time recognizing individual rights of speech ….”). The revised restrictions were content-neutral and provided a process to get permission to distribute political literature on campus along with rules for times when and locations where this was allowed. Id. at 617 n.2.

III. Applying This Law to Social Media Platforms

This Part first establishes why state courts can use the precedent and reasoning discussed above to extend state free speech clauses to social media platforms. It then draws from that precedent to explore what kind of content moderation 211 The term “regulation” may seem equally appropriate here as a matter of dictionary definition. However, because that smacks of direct state action, this Comment is of the mind that “moderation” is the better word choice. would remain permissible even if a state free speech clause were applied. To simplify and concretize the discussion, the following analysis uses Facebook as an exemplary social media platform. 212 This does not mean, however, that the approach below is limited to Facebook, and in each instance that “Facebook” is used, it can be substituted for any given social media platform. For example, Twitter almost certainly can be analyzed identically to Facebook. On the other hand, LinkedIn may present a different analysis and outcome, given its more niche nature and limited invitation to the public.

A. Existing State Case Law Naturally Extends to Facebook

The threshold consideration for this discussion is whether the case law discussed above fits social media platforms—in other words, whether it can be extended from shopping centers and private colleges. The answer is yes, and there are two routes for getting there. First, the “functional equivalent of a public forum” 213See supra Part II.B and accompanying notes. reasoning is applicable to Facebook, even more so than it is to shopping centers. Second, applying the New Jersey and Washington balancing factors 214See supra Part II.C and accompanying notes. as they were originally used, which was to determine whether a private actor is subject to the state free speech clause, shows Facebook fits neatly into each.

1. Facebook Is the Functional Equivalent of a Traditional Public Forum

Just as the replacement of the downtown district by the shopping center led some state courts to apply state free speech clauses to shopping centers, social media’s increasing replacement of physical public forums should inspire the same. Empirical data detail this story. 215See supra notes 2–9 and accompanying text. While the internet has increased dramatically in prevalence, brick-and-mortar retail stores have declined. 216See Nikaela Jacko Redd & Lutisha S. Vickerie, The Rise and Fall of Brick and Mortar Retail: The Impact of Emerging Technologies and Executive Choices on Business Failure, 17 J. Int’l Bus. & L. 127, 127–28 (2017); Wolf Richter, Here’s Which Brick-and-Mortar Retailers Are Getting Hit the Hardest, Bus. Insider (May 19, 2018, 11:02 AM), https://www.businessinsider.com/brick-and-mortar-retailers-getting-hit-the-hardest-2018-5. Moreover, the shopping centers at the very heart of the cases above have been in decline. 217 Josh Sanburn, Why the Death of Malls is About More than Shopping, Time (July 20, 2017), https://time.com/4865957/death-and-life-shopping-mall/. Some malls have even become so empty and run down that they are now popular spots for filming horror movies. See J.D. Capelouto, NEW FINDINGS: North Dekalb Mall is so Empty, it’s Becoming a Hot Spot for Filming, AJC (May 3, 2019), https://www.ajc.com/news/local/north-dekalb-mall-empty-become-hot-spot-for-filming/kTWe3QStwwyAoPaALmY2QO/. See also Liz Stinson, Starcourt Mall From ‘Stranger Things’ is Actually a Real Dead Mall, Curbed (July 12, 2019, 1:13 PM), https://www.curbed.com/2019/7/12/20691222/stranger-things-starcourt-mall-real-mall-georgia. In fact, the availability of the internet as an alternative route for free speech has even been used by some to argue that shopping centers should not be subjected to free speech clauses, and that the case law discussed above rests on outdated empirical premises, and so should be overruled. 218 Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 760 (Cal. 2007) (Chin, J., dissenting). Moreover, this dissent was written in 2007. It’s safe to say the argument has only grown stronger empirically. See also Selznick & LaMacchia, supra note 9, at 267–74. This argument may be valid and it may be appropriate, as an empirical matter, to reconsider that core reasoning as it applies to shopping centers. But assuming that this argument is true, it actually bolsters the case for an application to social media platforms because it relies on the internet replacing shopping centers as prominent locations of free speech.

Social media, specifically, is an increasingly important source of political information. 219 So much so that extensive investigation has been dedicated to the idea that manipulation of political speech on social media effectively swayed a presidential election. See, e.g., Jane Mayer, How Russia Helped Swing the Election for Trump, New Yorker (Sept. 24, 2018), https://www.newyorker.com/magazine/2018/10/01/how-russia-helped-to-swing-the-election-for-trump. An ever-increasing number of people turn to such platforms for news and other political information. 220See supra notes 3–9 and accompanying text. Indeed, news outlets have even encouraged its journalists to build their social media presences as a way to drive traffic to stories. 221See Jane Elizabeth, After a Decade, It’s Time to Reinvent Social Media in Newsrooms, Am. Press Inst. (Nov. 14, 2017), https://www.americanpressinstitute.org/publications/reports/strategy-studies/reinventing-social-media/single-page/. At the same time, recognizing the importance of journalists’ social media presence, some news outlets have sought to limit the kind of content their employees post. See Mathew Ingram, Social Media Crackdowns at the Times and Journal Will Backfire, Colum. Journalism Rev. (Oct. 20, 2017), https://www.cjr.org/criticism/social-media-twitter-times-journal.php. Political campaign spending on social media is also rising, and consultants are increasingly likely to recommend putting more money into online advertising and less into traditional television advertising. 222See, e.g., Bridget Bowman, The Future of Ads Is Digital — But Not Quite the Present, Roll Call (Dec. 11, 2018, 5:03 AM), https://www.rollcall.com/news/politics/whats-next-digital-ads-spending-increases-still-slow-change; Lindsey Erdody, Political Campaigns Boost Investment in Social Media Ads, Indianapolis Bus. J. (Sept. 21, 2018), https://www.ibj.com/articles/70545-political-campaigns-boost-investment-in-social-media-ads; Nihal Krishan, Dems Ramp Up Digital Advertising Effort for Midterms. Trump’s Doing the Same for 2020, OpenSecrets.org (Sept. 11, 2018), https://www.opensecrets.org/news/2018/09/dems-ramp-up-digital-ads-for-midterms-trumps-doing-2020/. Finally, the last two successful presidential candidates have used social media presence to great effect, 223See supra note 10; Election 2016: Campaigns as a Direct Source of News, Pew Res. Ctr. (July 18, 2016), http://www.journalism.org/wp-content/uploads/sites/8/2016/07/PJ_2016.07.18_election-2016_FINAL.pdf. and candidate Donald Trump controlled entire news cycles with social media postings. 224See Andrew Buncombe, Donald Trump One Year on: How the Twitter President Changed Social Media and the Country’s Top Office, Indep. (Jan. 17, 2018, 9:43 PM), https://www.independent.co.uk/news/world/americas/us-politics/the-twitter-president-how-potus-changed-social-media-and-the-presidency-a8164161.html. All of this only scratches the surface, and the data discussed above put hard numbers on the story. Moreover, the trends seem unlikely to reverse, as teenagers and young adults are especially likely to spend significant amounts of time on social media platforms. 225See supra note 5 and accompanying text. This evidence of where people and speech “are” tracks the evidence used by courts in reasoning that shopping centers replaced downtown districts, making them functional equivalents of public forums. 226See supra Part II.B and accompanying notes.

Looking to Facebook specifically, its primary purpose is expressly not commercial. 227 Facebook’s own words support this assertion, and a discussion of Facebook’s purported purpose and its mission statement is below. See infra note 236 and accompanying text. The Supreme Court of Pennsylvania’s reasoning in Western Pennsylvania Socialist Workers 1982 Campaign v. Connecticut General Life Insurance Co. 228 515 A.2d 1331 (Pa. 1986). helps show why. There, the court did not apply Pennsylvania’s free speech clause to a shopping center because the center invited the public onto the property only for commercial purposes and to exchange goods and services but not ideas. 229See supra notes 138–146 and accompanying text. This may be the correct interpretation of a shopping center’s invitation to the public. But, like the empirical claims about the decline of shopping centers, assuming the truth of this argument actually advances the case for applying state free speech clauses to Facebook. Unlike the shopping center’s invitation to the public as the Supreme Court of Pennsylvania found it, Facebook’s is the exact opposite: its invitation to people to use its platform is not connected to the exchange of goods and services, but to the exchange of ideas. The commercial benefits Facebook derives are ancillary to this, whether through advertisements by third parties or the data mining.

To quote the Supreme Court of New Jersey, social media is where the people “are” today 230N.J. Coalition Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 767 (N.J. 1994)—especially for purposes of political speech. That makes it important to protect political speech on these platforms. Just as the courts discussed above reasoned free speech protection would lose much of its force if not applied to the newly important shopping centers, 231See supra notes 133–134 and accompanying text. the same is applicable to political speech on social media. Indeed, it is even more important because social media platforms exist specifically to foster speech and communication, 232See, e.g., FAQs, Facebook: Inv. Rel., https://investor.fb.com/resources/default.aspx (last visited Feb. 17, 2019)., and they prosper by attracting a perpetually increasing number of users. 233 For this reason, social media platforms tend naturally toward monopolistic features. See Luigi Zingales & Guy Rolnik, A Way to Own Your Social-Media Data, N.Y. Times (June 30, 2017), https://www.nytimes.com/2017/06/30/opinion/social-data-google-facebook-europe.html. But see Sangeet Paul Choudary, Reverse Network Effects: Why Today’s Social Networks Can Fail as They Grow Larger, Wired, https://www.wired.com/insights/2014/03/reverse-network-effects-todays-social-networks-can-fail-grow-larger/ (last visited Jan. 17, 2019). In short, the point for users is to be on the platform with the most other users; scale, therefore, is crucial. Although the dominant platform might not remain so perpetually, it seems most likely that change will be through users transitioning to another, newly dominant platform, and not splintering into smaller ones (like for example the correlated decline of MySpace and rise of Facebook). The importance of this is that, even if a platform were to ban a certain viewpoint and proponents of that viewpoint created their own platform, it would be difficult for the new platform to match the scale requisite to being influential. Therefore, decisions by the owners to silence a particular viewpoint can be especially pernicious and effective.

2. Balancing Test Factors Support Applying State Free Speech Clauses to Facebook

The balancing factors from New Jersey’s and Washington’s supreme courts 234See supra Part II.C. provide an alternative route to finding social media platforms subject to state free speech clauses, especially if the given clause lacks a state action requirement. 235 As the New Jersey and Washington supreme courts found in the relevant cases. See supra notes 154 and 169. Applying those factors to Facebook makes it clear that this route is available.

First, the primary use and nature of Facebook’s property is to provide a platform for speech. 236 Facebook’s mission statement says its “mission is to give people the power to build community and bring the world closer together” through letting people “share and express what matters to them.” FAQs, Facebook: Inv. Rel., https://investor.fb.com/resources/default.aspx (last visited Feb. 17, 2019). It does not exist to sell a product to users, or even to create a marketplace for exchanging services. 237 These activities may occur on Facebook, which creates a reverse situation from shopping centers as described in cases like New Jersey Coalition Against War in the Middle East, 650 A.2d at 782 (N.J. 1994). See supra notes 136–137 and accompanying text. Instead, its primary use is to enable people to communicate and share ideas, and its nature is to be open to varying viewpoints and accepting of all users. 238 This is with the sole exception of people younger than 13 years old. See Mallory Locklear, Facebook Will Begin Actively Enforcing Its Age Policy, Engadget (July 19, 2018), https://www.engadget.com/2018/07/19/facebook—actively-enforcing-age-policy; FAQs, Facebook: Inv. Rel., https://investor.fb.com/resources/default.aspx This is similar to the private college in State v. Schmid239 423 A.2d 615 (N.J. 1980). There, the primary use of the property was to pursue truth and foster learning. 240Id. at 230. The tool for that pursuit is the communication of ideas. Facebook may not have the primary purpose of pursuing truth, but it does have the primary purpose of enabling the tool of that pursuit.

Second, the extent of the public’s invitation to use Facebook’s property is as extensive as possible, with the only limit being people under the age of thirteen. 241 An interesting thought arises when considering Facebook pre-2006, when it was only available to people with college email addresses. See Saul Hansell, Site Previously for Students Will Be Opened to Others, N.Y. Times (Sept. 12, 2006), https://www.nytimes.com/2006/09/12/technology/12online.html. In that case, its invitation to the public would be much more limited than it is currently, which could prevent a state free speech clause from reaching it under this Comment’s approach. Indeed, Facebook’s existence and success depend on a constantly increasing number of people using its property. In Schmid, the Supreme Court of New Jersey considered whether a public presence was consonant with the private college’s mission and concluded that it was because the college encouraged wide and continuous exchange of opinions. 242Schmid, 423 A.2d at 631. The same is true of Facebook.

Third, political speech is entirely compatible with Facebook’s primary use. It is in large part the very point of Facebook. In New Jersey Coalition Against War in the Middle East v. J.M.B. Realty Corp., the Supreme Court of New Jersey concluded that political speech was compatible with a shopping center’s uses because “two hundred years of compatibility between free speech and the downtown business district [was] proof enough” and it had no discernible negative financial impact. 243N.J. Coal., 650 A.2d at 775. Moreover, the shopping center could minimize any negative impact with time, place, and manner restrictions. 244Id. Although social media does not have 200 years of history, its similarity to downtown business districts makes that history informative. And Facebook can also mitigate against any possible financial detriment through restrictions like those discussed in detail below. 245See infra Part III.B and accompanying notes.

Fourth, the extent to which political speech would be undermined if Facebook were to prohibit it is an empirical question. The answer is likely that it would be significant. One concern is Facebook barring political speech totally, from all viewpoints. This would clearly significantly undermine such speech for the reasons discussed throughout this Comment; namely, that an increasing number of people turn to Facebook for political information. 246See supra notes 3–9 and accompanying text. Less obvious, but just as concerning, is if Facebook were to silence a single viewpoint. While this may not obviously undermine the amount of political speech as a whole if measured quantitatively, it would still be utterly devastating for the silenced viewpoint and qualitatively detrimental for speech overall. 247See supra note 21.

Each of these factors weighs in favor of free political speech rights over Facebook’s property rights. Moreover, a general balancing between free speech rights and private property rights—like that done in New Jersey Coalition—favors applying state free speech clauses to Facebook. What the Supreme Court of New Jersey noted there is so applicable here that it could have been written with Facebook in mind:

[T]he weight of the private property owners’ interest in controlling and limiting activities on their property has greatly diminished in view of the uses permitted and invited on that property … [They] have intentionally transformed their property into a public square … On the other side of the balance, the weight of plaintiff’s free speech interest is the most substantial in our constitutional scheme. Those interests involve speech that is central to the purpose of our right of free speech. 248N.J. Coal., 650 A.2d at 776.

B. Facebook Should Retain the Power to Moderate Content for Specific and Intentional Slurs, but not the Underlying Political Viewpoint

The above section establishes two lines of reasoning that support an application of state free speech clauses to Facebook. But were a court to do so, it would raise the question of what power Facebook would retain to moderate content on its platform. It is clear that Facebook must retain some discretion to limit content as all the relevant case law—including that of the United States Supreme Court—holds. 249See supra Part II.D. It is not as clear, though, what the contours of permissible speech restrictions may look like because existing case law discusses these restrictions in terms of “time, place, and manner,” 250See supra Part II.D and accompanying notes. which is a concept of little help in Facebook’s nonphysical context. However, the Green Party of New Jersey v. Hartz Mountain Industries, Inc. factors for balancing a property owner’s restrictions against a speaker’s speech rights are not necessarily tethered to physical space, and they guide the following analysis. 251 752 A.2d 315, 327 (N.J. 2000).

This section suggests the following balance: Broadly speaking, Facebook should not be allowed to deplatform users simply for their political speech, whether controversial or not; nor should it be able to delete posts based on the political viewpoint expressed. But Facebook should retain the ability to moderate content, not based on the viewpoint of a post, but on the specific language used if that language is particularly hateful toward a certain group and particularly unnecessary for the viewpoint’s message. Although this must still be a narrow range of discretion because it carries the potentiality of abuse, this approach has the twin benefits of preserving open, free-ranging political debate and improving the quality—and ultimately the civility—of political discourse.

1. Outline of Principles

Because the case law provides little direct guidance for how Facebook could moderate content under this Comment’s approach, it is helpful to outline a few broader principles extracted from that precedent. First, Facebook should retain full power to moderate, restrict, or otherwise ban nonpolitical speech, and this Comment has in mind only political speech. This is for the reasons discussed above—the importance and special position of political speech 252See supra Part II.A.—but also because the balancing of interests here is premised on political speech’s importance. 253 This is also evident in the cases. See supra Part II.C. For example, Facebook has a heightened interest in moderating content including nudity, vulgar language, or otherwise salacious material. The same can be said for nonpolitical language that is particularly “cruel and insensitive.” 254Community Standards, Facebook, https://www.facebook.com/communitystandards/cruel_insensitive (“[W]e have higher expectations for content that we call cruel and insensitive, which we define as content that targets victims of serious physical or emotional harm.”) (last visited Nov. 13, 2018). On the other side of the equation, users’ and society’s interest in this kind of speech is not as important as it is in political speech. This calculus, therefore, will likely always tip in favor of Facebook’s restriction. 255Cf. N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 781–82 (N.J. 1994).

Second, the traditional notion of “time, place, and manner” restrictions is out of place here for the simple reason that Facebook involves no physical space. Whereas New Jersey Coalition makes a point of limiting its holding to speech that does not involve disruptive means like soapboxes and bullhorns, 256Id. at 782. such concerns are largely irrelevant to a digital platform. Unlike when strolling through a mall’s common area where one physically cannot tune out or otherwise ignore obnoxious speakers, users on social media platforms can mute, unfollow, or even entirely block others. Moreover, videos in Facebook feeds, though automatically starting, do not include sound unless manually enabled. It is true that Facebook could create a virtual analog to time, place, and manner restrictions by limiting certain kinds of posts to specific times of day or imposing word counts. However, this seems unlikely and is not a concern of this Comment as such measures would clearly comport with all the case law unless being so limiting as to effectively stifle speech. 257See Green Party of N.J. v. Hartz Mountain Indus., Inc., 752 A.2d 315, 332 (N.J. 2000) (striking a shopping center’s restriction, in part because more than the one day a year the center allowed is needed for effective leafletting).

Although a “time, place, and manner” concept is unhelpful, the balancing factors articulated by the Supreme Court of New Jersey in Green Party of N.J. are. Rather than being tied up with physical space, they incorporate more abstract principles that can be applied to Facebook’s virtual situation. These factors look at (1) the nature of the affected right; (2) the extent to which the restriction intrudes upon it; and (3) the property owner’s need for the restriction. 258Id. at 327. These can be applied to the specifics of any given fact pattern and can offer guidance for a more general discussion of what kind of restriction might survive any challenge. This will be done below.

In addition to narrowing the focus to political speech and providing useful factors for a balancing test, the case law discussed above also contains two applicable principles that function in tandem. First, any restriction policies must be objective enough to avoid arbitrary and viewpoint-based application. 259See supra notes 207–208 and accompanying text. Second, viewpoint-based restrictions are likely invalid. 260See supra notes 205–206 and accompanying text; cf. Hill v. Colorado, 530 U.S. 703, 769–70 (2000) (Kennedy, J., dissenting) (“The Court time and again has held content-based or viewpoint-based regulations to be presumptively invalid.”).

The latter point warrants a moment of discussion when compared to the content-based restrictions criticized by the cases. While the cases address content-based restrictions, and are harsh on them, 261See supra notes 205–208 and accompanying text. the broad category is out of place when considering Facebook. This is for the same reason why “time, place, and manner” is inapplicable: There is no physical space to manage so almost any limitation will be content-based. Facebook’s speech policies are content moderation policies—they are necessarily based on content, as that is virtually all there is for Facebook to address. 262 Again, this is other than the possibility of Facebook doing things like limiting some types of posts to certain times of day.

On the other hand, a narrower subset of content-based restrictions is viewpoint-based restrictions, 263See Rosenberger v. Rector & Visitors of Univ. of Va., 515 U.S. 819, 828–29 (1995) (noting that viewpoint discrimination is “[w]hen the government targets not subject matter, but particular views taken by speakers on a subject”). See Reed v. Town of Gilbert, 135 S. Ct. 2218, 2229–30 (2015) and Pahls v. Thomas, 718 F.3d 1210, 1229 (10th Cir. 2013) for a further discussion of the distinction between content-based and viewpoint-based restrictions. and here it can serve the same role as a content-based framework does for physical locations. The case law discussed above has numerous examples of courts finding that a property owner allowing one actor to engage in political speech but not another was impermissible. 264See supra notes 205–208 and accompanying text. Although not explicitly connected in the reasoning, at least two of these cases involved antiwar protestors being prohibited from speaking despite the relevant shopping center having allowed what could be deemed more mainstream political speech. 265See Bock v. Westminster Mall Co., 819 P.2d 55, 56–57 (Colo. 1991) (noting that the shopping center previously allowed voter registration drives, salute to armed forces and displays by armed forces agencies, agency representatives, salute to presidents and display of presidential portraits, Boy Scouts and Girl Scouts, Salvation Army, controversial films); N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 765, 776 (N.J. 1994) (noting that the shopping center previously allowed voter registration drive, sixty-seven government agencies distributing literature, senator campaigning in person, Toys for Tots, military recruitment, Berlin Wall Exhibit, senate campaign voter registration drive, &c.); see also W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331, 1336 (Pa. 1986) (“Nor does [the law] yet require [the shopping center] to provide a political forum for persons or groups with views on public issues, so long as the owner does not grant unfair advantage to particular interests or groups by making his premises arbitrarily available to those he favors while excluding all others.” (emphasis added)). Another case held that a shopping center could ban all political speech, but could not “grant unfair advantage to particular interests” by making the space available to those the owner favors but not others. 266W. Pa. Socialist Workers, 515 A.2d at 1336. This supports concluding that a social media speech restriction policy premised on discriminating against a certain viewpoint should be met with skepticism, lest the restriction be used to “grant unfair advantages” and significantly sway the course of political discussion.

2. Application of Principles

With these principles in mind, this Section will consider what specific political speech restrictions by Facebook remain permissible. A clear starting point is that there is little Facebook can do to limit political speech. Whereas shopping centers can limit the times and places where political speakers can speak, 267Supra Part II.D. and moreover can limit the techniques used to speak, 268See e.g., N.J. Coal., 650 A.2d at 782 (“[O]ur ruling … does not include bullhorns, megaphones, or even a soapbox; it does not include placards, pickets, parades, and demonstrations; it does not include anything other than normal speech and then only such as is necessary to the effectiveness of the leafletting.”). Facebook’s content moderation policies and its interest in them are largely unconcerned with these kinds of restrictions. Additionally, almost any restriction will be, at some level, viewpoint-based and thus run afoul of the restricted speaker’s rights. Under this approach, Facebook has only the narrowest—if any—grounds upon which to ban accounts or limit posts based on political speech.

This outcome is supported by applying the Green Party of N.J. balancing factors. 269 This application is in the abstract and to political speech generally. But it is difficult to imagine a scenario where some specific political statement alters the equation, so the outcome would likely be the same when applied to any given example of a political post. First, the nature of the right affected—political speech—is paramount, as has been discussed throughout this Comment. Second, any restriction that deletes an account or post necessarily limits that right totally as to the specific user, 270 As to banning an account, this completely limits the user’s free speech right, at least with respect to the exercise of that right on the platform. And, when deleting a single post, it does not limit the user’s right to make other statements, but as to that statement and its viewpoint, the limitation is total. Of course, this does not mean the user’s free speech right is limited totally beyond the platform. But, given the perpetually increasing importance and centrality of social media platforms, the limitation can have far-reaching consequences. and is detrimental to political speech beyond that. 271See supra note 21. Third, Facebook’s need to delete accounts or posts because of political speech is limited. Although this may ultimately be a case-specific empirical question, it is doubtful that Facebook could show any given account or post being so detrimental to its bottom-line 272 Though Facebook’s owners may not like having certain views expressed on the platform or may see them as in tension with Facebook’s mission statement, the operative inquiry in the case law detailed above focused only on the economic impact on the private actor. Of course, it remains possible for any given platform to rebrand as specifically only accepting of one type of ideology, which would in turn change the other factors like the scope of the invitation to the public. or mission statement that the need for deleting it trumps the highly protected free speech rights at stake.

In light of the above, what is left for Facebook to do? There remains a narrow approach that would let Facebook moderate content to accomplish its goals of limiting hateful speech while preserving—or preventing Facebook from silencing—political speech and its underlying viewpoints. To give the conclusion before the support, this approach is to prohibit Facebook from restricting any content based on its political message or viewpoint while allowing Facebook to limit the use of some specific language, namely epithets and slurs.

Facebook’s current hate speech policy provides a test case for applying this approach. It prohibits hate speech, and states, in relevant part:

We define hate speech as a direct attack on people based on what we call protected characteristics— race, ethnicity, national origin, religious affiliation, sexual orientation, caste, sex, gender, gender identity, and serious disease or disability. We also provide some protections for immigration status. We define attack as violent or dehumanizing speech, statements of inferiority, or calls for exclusion or segregation. 273Community Standards, Facebook, https://www.facebook.com/communitystandards/hate_speech (last visited Feb. 17, 2019).

Facebook also considers context and the intent of the post, being less stringent when the purpose of posting hate speech is to educate, empower, or be humorous. 274Id.; see also Klonick, supra note 13, at 1630–61 (discussing in detail Facebook’s processes for content moderation); Richard Allan, Hard Questions: Who Should Decide What Is Hate Speech in an Online Global Community, Facebook: Newsroom (June 27, 2017), https://newsroom.fb.com/news/2017/06/hard-questions-hate-speech/.

Taken alone, this policy can clearly run afoul of the free speech protections this Comment argues for. But, applied carefully, it can survive and be beneficial to both Facebook’s community and political discourse generally. Two examples illustrate how and why.

First, consider anti-immigrant proponents. It is easy to imagine a well-reasoned and articulate (but disagreeable) post discussing why the United States should limit or ban immigration or increase deportations. It might say that immigrants from certain countries do not share western values, are unable to assimilate, have less work ethic, &c., and so should be excluded from the country. 275 For a list of common arguments against immigration that may fit this example (and, less relatedly, a rebuttal to each), see Alex Nowrasteh, The 14 Most Common Arguments Against Immigration and Why They’re Wrong, Cato Inst.: Cato Liberty (May 2, 2018, 11:10 AM), https://www.cato.org/blog/14-most-common-arguments-against-immigration-why-theyre-wrong. Or, if the post addressed a “caravan” of immigrants approaching the border, it might say that they are carrying diseases, would be bad for the economy, and would be a safety threat, so should be excluded. 276 These arguments were made by pundits and public intellectuals. See, e.g., Todd Bensman, Middle Eastern Migrants Part of Caravan?, Ctr. for Immigr. Stud. (Oct. 22, 2018), https://cis.org/Bensman/Middle-Eastern-Migrants-Part-Caravan; Michelle Malkin, It’s Not Unreasonable to be Worried About Disease and the Caravan, Nat’l Rev. (Oct. 31, 2018, 11:09 AM), https://www.nationalreview.com/2018/10/its-not-unreasonable-to-be-worried-about-disease-and-the-caravan/. On the other hand, it is also easy to imagine a post that does not make these arguments, but instead makes the same ultimate policy prescription (exclusion) by saying that immigrants are (or are like) “animals” 277See Julie Hirschfeld Davis, Trump Calls Some Unauthorized Immigrants ‘Animals’ in Rant, N.Y. Times (May 16, 2018), https://www.nytimes.com/2018/05/16/us/politics/trump-undocumented-immigrants-animals.html. and are from “shit hole countries.” 278See Julie Hirschfeld Davis et al., Trump Alarms Lawmakers with Disparaging Words for Haiti and Africa, N.Y. Times (Jan. 11, 2018), https://www.nytimes.com/2018/01/11/us/politics/trump-shithole-countries.html. Admittedly, using examples spoken by a president or other politician raises a complication because such speech is necessarily political and presumably entitled to great protection, even if squarely within Facebook’s hate speech policy. In such situations, the balancing test will almost necessarily favor the speech because it directly involves politicians and matters of important public information. Notably, Twitter co-founder and CEO Jack Dorsey has used this reasoning in explaining why President Trump’s tweets should almost never be taken down. See #148 – Jack Dorsey, Making Sense (Feb. 5, 2019) (downloaded using Spotify); #1236–Jack Dorsey, Joe Rogan Experience (Feb. 1, 2019) (downloaded using Spotify).

Both of the above clearly fit within Facebook’s hate speech policy, even if the latter does more obviously. Facebook’s definition of “attack” includes “statements of inferiority, or calls for exclusion or segregation.” 279Community Standards, Facebook, https://www.facebook.com/communitystandards/hate_speech (last visited Feb. 17, 2019). Arguing for less immigration (or deportation) is necessarily a call for exclusion. Further, such arguments can easily be defined as casting certain people as “inferior.” The problem is that these arguments, regardless of how they are expressed or supported, can—and under Facebook’s definition of hate speech, should—be grouped in the same category as calling immigrants “animals.” Such posts should, therefore, be banned. But this undermines free speech by limiting the speech of a significant—and not necessarily hateful—political viewpoint.

Second, consider opponents of gay marriage. On the one hand, one can imagine a post arguing against the decision in Obergefell v. Hodges 280 135 S. Ct. 2584 (2015) (holding states cannot prohibit gay marriage). because of a disagreement over interpreting the Fourteenth Amendment, 281See, e.g., id. at 2611 (Roberts, C.J., dissenting). or a post arguing that gay marriage is immoral for cited Biblical reasons and should be outlawed by states. 282See, e.g., Darrell Bock, The Bible and Same-Sex Marriage: 6 Common But Mistaken Claims, Gospel Coalition (July 27, 2015), https://www.thegospelcoalition.org/article/bible-same-sex-marriage-6-wrong-trajectories/. On the other hand, one can also imagine posts saying something utterly crass like, “fags are an abomination and going to hell and should never be allowed to marry.” Again, both arguments ultimately make the same substantive point, but with radically different presentations. And again, both fall under Facebook’s hate speech policy and should be banned, posing the same free speech problem as the immigration example.

The solution is to allow Facebook discretion to limit the latter, while making the former virtually off-limits. This effectively narrows the hate speech policy, but without sacrificing its goals—and, more importantly, without contravening free speech rights. First, by narrowing the hate speech policy to certain slurs, or otherwise more specific language, the ability for its malleable standards to be enforced arbitrarily against viewpoints the platform’s owners dislike is limited. 283Cf. supra notes 205–208 and accompanying text. Second, limiting the use of certain slurs or unnecessarily dehumanizing language is not viewpoint-based—at least not in the same sense as it is when the underlying argument itself is being prohibited. 284 It is important to distinguish between intentionally hateful use of such slurs and arguments that such slurs should be allowed to be used. If a post argues that slurs or a given slur should be allowed to be used, then prohibiting such a post becomes viewpoint-based against that viewpoint.

Finally, and crucially, this approach alters the outcome of the balancing test under the Green Party of N.J. factors. First, the nature of the affected right changes. The involved speech is less important, and Facebook’s need to moderate it is increased. Although still intertwined with political speech, the use of certain words or modes of speaking that do little to serve the underlying political viewpoint diminishes that specific speech’s importance. Second, Facebook’s need to restrict such language is greater than its need to restrict the underlying viewpoint. The extent to which any viewpoint is off-putting or alienating to users is greatly exacerbated by certain unnecessary and prejudicially charged language. 285See generally Jonathan Haidt, The Righteous Mind (2012). This could, in turn, cause an increased number of users to quit the platform, giving Facebook a greater need to moderate the content.

There is, of course, a difficulty in line-drawing. Where exactly a post crosses the line may not always be clear. But this ambiguity can be reduced by focusing on a concrete list of words intentionally used as slurs. There may also be certain types of language that are, effectively, hateful and identity-based 286 For examples, see Facebook’s listed “protected characteristics.” Community Standards, Facebook, https://www.facebook.com/communitystandards/hate_speech (last visited Feb. 17, 2019) (listing the following protected characteristics: race, ethnicity, national origin, religious affiliation, sexual orientation, caste, sex, gender, gender identity, serious disease or disability, and “some protections for immigration status”). name-calling that do not advance the underlying viewpoint’s argument. 287 This should not reach so far as to affect things like calling arguments, or even their proponents, stupid. Although this may be inflammatory, it does not fall within the “protected characteristics” promulgated by Facebook. It also creates a heightened problem of line-drawing because disagreement necessarily requires calling another viewpoint wrong, which is inherently offensive. Even though name calling may not be “decent” in a traditional sense, it does not tip the balancing test toward restricting such speech.

Any ambiguity in where this line is, however, can be beneficial for two reasons. First, it incentivizes Facebook to err on the side of caution, which builds in greater protection for free speech. Second, it forces Facebook to be especially vigilant and careful in how it moderates political speech. This encourages it to be sensitive to the needs of both the users who are posting the content and those who are bothered by it. 288 An interesting question is raised by Facebook’s algorithm and decisions to promote or prevent certain content from appearing in feeds. For example, Facebook has recently made an effort to have users’ feeds feature more posts by friends, and less by pages who pay for placement, see Adam Mosseri, Bringing People Closer Together, Facebook: Newsroom (Jan. 11, 2018), https://newsroom.fb.com/news/2018/01/news-feed-fyi-bringing-people-closer-together/, which some have alleged is inherently biased against conservative viewpoints, see Jason Schwartz, Conservative Outlets Take on Facebook, Politico (Mar. 29, 2018, 5:00 AM), https://www.politico.com/story/2018/03/29/conservatives-facebook-liberal-bias-490920. And, there are certainly general concerns that Facebook can use its algorithm or manual manipulation to suppress some viewpoints. See Laura Wagner, Is Facebook Suppressing Politically Conservative Content?, NPR (May 10, 2016, 6:06 PM), https://www.npr.org/sections/thetwo-way/2016/05/10/477525204/is-facebook-suppressing-politically-conservative-content. But see Mark Zuckerberg, Facebook (May 12, 2016), https://www.facebook.com/zuck/posts/10102830259184701 (reporting that internal investigations have found no evidence of suppression). Whether Facebook does, in fact, do this, it certainly possesses the ability to. But such actions need not necessarily conflict with free speech. For example, changes may be viewpoint-neutral decisions meant simply to make users’ feeds more attractive by including more posts from family members. But, one can also imagine a situation in which Facebook intentionally and perniciously suppresses a viewpoint, and thus accomplishes substantially the same effect as deleting it. Where exactly to draw the line at which this violates free speech protections is beyond this Comment’s scope; that determination, however, can still be resolved by applying the balancing test this Comment advances.

Finally, at first blush it may seem dubious to give Facebook the power to define what is and is not within the realm of non-hateful—or decent—political discourse. But Facebook already has that power, and more. Now, it has the power to determine the very viewpoints that get a microphone and a soapbox, with no recourse for deplatformed users, and no judicial check. This Comment’s approach changes that in two ways: First, it tells Facebook that it cannot exclude viewpoints themselves simply for being disagreeable and gives users a vehicle for redress when they are prohibited from expressing certain ideas. Second, it lets Facebook determine specific types of language that nevertheless violate Facebook’s standards, while putting state courts in the background as a check on Facebook abusing this power to arbitrarily discriminate against certain viewpoints.

C. U.S. Supreme Court Precedent Revisited

This Comment’s approach satisfies any requirements of the United States Constitution under Pruneyard Shopping Center v. Robins289 447 U.S. 74 (1980). Although First Amendment jurisprudence has evolved in the decades since Pruneyard, it does not seem that the United States Supreme Court has moved away from Pruneyard and its underlying federalist principles. At least as recently as 2008, the Court denied a petition for certiorari in a California case where the Supreme Court of California applied strict scrutiny to a shopping center’s speech restriction. See Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 751, 754 (Cal. 2007), cert. denied, 555 U.S. 819 (2008); see also Ralphs Grocery Co. v. United Food & Commercial Workers Union Local 8, 290 P.3d 1116 (Cal. 2012), cert. denied, 133 S. Ct. 2799 (2013) (denying a petition for certiorari in case using state statute to similar effect); Eugene Volokh, The Law of Compelled Speech, 97 Tex. L. Rev. 355, 358 (2018) (exploring current state of compelled speech law under the First Amendment). There is no takings problem for at least two reasons. On the one hand, like the shopping center in Pruneyard290See supra notes 91–91 and accompanying text. there is nothing to suggest that allowing political speech unreasonably impairs Facebook’s value. This is even more applicable to a social media platform, where such speech is integral and expressly invited, than it is for a shopping center, where patrons are likely not present for the purpose of receiving political information. 291See supra notes 141–146 and accompanying text. On the other hand, and again like the shopping center in Pruneyard, 292See supra note 92 and accompanying text. Facebook retains the ability to moderate content to limit any negative effect that such speech may have. There is also no due process problem for the same reasons as in Pruneyard293See supra note 94 and accompanying text. namely, the approach is not unreasonable, arbitrary, or capricious, and a state’s interest in promoting free speech protections has a substantial relation to the means.

Further, Facebook’s First Amendment rights are not violated by this Comment’s approach. Again, the main reasons given in Pruneyard 294See supra notes 95–98 and accompanying text. are applicable here. First, because Facebook is so open to the public and lets everyone post just about anything, it is unlikely that any users—or, for that matter, anyone in society—equate a Facebook user’s views with the views of Facebook itself. Indeed, anyone who uses or has heard of the platform likely knows intuitively that Facebook is just that—a platform—and not the producer or endorser of the content. Second, this approach does not mandate that Facebook display any particular message, just that it cannot discriminate against any political viewpoint. And third, Facebook can easily post its own messages expressly disavowing any speech on its platform. It can do so in a way that guarantees all users see it, whether having it show up at the top of every user’s feed as a general disclaimer or putting a disavowal directly under any given post that it disagrees with. Finally, Miami Herald Publishing Co. v. Tornillo can likewise be distinguished 295See supra note 99 and accompanying text. because rather than deterring allowing controversial political statements, this approach deters prohibiting them. 296 On the question of whether Facebook is a publisher like a newspaper, it has tried to have its cake and eat it too, arguing both that it is and is not. See, e.g., Sam Levin, Is Facebook a Publisher? In Public It Says No, But in Court It Says Yes, Guardian (July 3, 2018, 2:00 PM), https://www.theguardian.com/technology/2018/jul/02/facebook-mark-zuckerberg-platform-publisher-lawsuit. Regardless of the legal answer to that question, this Comment’s approach would survive. The balancing of Facebook’s free speech rights against a user’s free speech rights would still take on the analytical framework of Pruneyard, as it is the guiding precedent for this kind of scenario.

IV. Possible Barriers for State Courts

Assuming a state court is inclined to adopt this Comment’s framework and extend free speech protections to its state’s citizens, 297 The rights under a given state’s constitution would only be applicable to that state’s citizens. See, e.g., People v. LaValle, 817 N.E.2d 341, 366 (N.Y. 2004) (“It bears reiterating here that ‘on innumerable occasions this Court has given the State Constitution an independent construction, affording the rights and liberties of the citizens of this State even more protection than may be secured under the United States Constitution.’” (quoting Sharrock v. Dell Buick-Cadillac, 379 N.E.2d 1169, 1173 (N.Y. 1978)) (emphasis added)); see also Baldwin v. Fish & Game Comm’n, 436 U.S. 371, 383 (“It has not been suggested, however, that state citizenship or residency may never be used by a State to distinguish among persons. Suffrage, for example, always has been understood to be tied to an individual’s identification with a particular State.” (citations omitted)); Blake v. McClung, 172 U.S. 239, 256–57 (1898) (“We must not be understood as saying that a citizen of one State is entitled to enjoy in another State every privilege that may be given in the latter to its own citizens.”). there may nevertheless be procedural hurdles to its ability to do so. Because each of these presents complicated legal issues themselves, a lengthy and detailed discussion exceeds the scope of this Comment. Instead, this Part focuses on two significant potential barriers—personal jurisdiction and Section 230 of the Communications Decency Act—and makes a few brief comments about each. However, this analysis is not meant to be the final word. Instead, it is a starting point, and these issues, among others, 298 For example, other obstacles may be Facebook’s “Terms of Service” agreement, which users agree to when setting up an account. This could affect choice of forum and choice of law. See Dolin v. Facebook, Inc., 289 F. Supp. 3d 1153, 1158–61 (D. Haw. 2018) (discussing whether choice of forum clause in Facebook’s terms of service is binding); Fteja v. Facebook, Inc., 841 F. Supp. 2d 829, 834–41 (S.D.N.Y. 2012) (discussing the same). Another is that a user would need a cause of action, the specifics of which may vary by state. Compare Corum v. Univ. of N.C., 413 S.E.2d 276, 289–91 (N.C. 1992) (holding there is a direct action against the state for free speech violations under the state constitution), with Roman v. Trs. of Tufts Coll., 964 N.E.2d 331, 336–37 (Mass. 2012) (discussing statute providing cause of action for rights secured by the state or federal constitution). are ripe for future scholarship, litigation, and jurisprudence. 299 It is also crucial that this Comment not be the final word, and that the approach it proposes be explored further. Plaintiffs seeking to protect their free speech against arbitrary censorship by social media platforms stand little chance under the First Amendment, absent an unexpected and momentous reversal of entrenched precedent by the United States Supreme Court. Yet, litigants continue to bring doomed-to-fail claims based on the First Amendment. See, e.g., Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 WL 1471939, at *5 (N.D. Cal. Mar. 26, 2018) (holding YouTube is not a state actor for First Amendment purposes); Shulman v. Facebook.com, No. 17-764 (JMV), 2017 WL 5129885, at *4 (D.N.J. Nov. 6, 2017) (holding Facebook is not a state actor for First Amendment purposes and collecting cases holding the same). Instead, getting this issue before state courts may be the best—and possibly only—way for speakers and scholars to protect free speech against social media platforms’ growing power to control it. In that spirit, this Comment suggests what to say once before a state court, and offers state courts a framework to begin from, leaving the path to getting there to future trailblazers.

A. Personal Jurisdiction

A given state court may have personal jurisdiction over a social media platform through either general or specific jurisdiction. 300See generally Elma Delic, Note, Cloudy Jurisdiction: Foggy Skies in Traditional Jurisdiction Create Unclear Legal Standards for Cloud Computing and Technology, 50 Suffolk U. L. Rev. 471, 472–78 (2017) (discussing development of personal jurisdiction law generally and its application to the internet); Scott Isaacson, Note, Finding Something More in Targeted Cyberspace Activities, 68 Rutgers U. L. Rev. 905, 914–26 (2016) (discussing personal jurisdiction and applications to cyberspace). In Facebook’s case, there is general jurisdiction in California, where its principal place of business is, and Delaware, its state of incorporation. 301See Daimler AG v. Bauman, 571 U.S. 117, 137 (2014) (holding that the paradigmatic examples of general jurisdiction over a corporation are the state in which it is incorporated and in which its principal place of business is). Facebook, in litigation, has argued that there is general jurisdiction over it in California and Delaware. See Georgalis v. Facebook, Inc., 324 F. Supp. 3d 955, 959–61 (N.D. Ohio 2018) (rejecting that there is general jurisdiction over Facebook in Ohio because it has many users there); Ralls v. Facebook, 221 F. Supp. 3d 1237, 1243 (W.D. Wash. 2016). This alone is significant. California is one of the states with case law most primed for application to Facebook, 302See supra Part II. and it is the most populous state in the United States with around forty million people. 303California Population 2019, World Population Rev., http://worldpopulationreview.com/states/california-population/ (last visited Jan. 20, 2018). Successful adoption of this Comment’s approach there, if nowhere else, would mark a significant step in the protection of free speech by giving each of California’s citizens free speech protection online. 304 Additionally, Facebook’s “Terms of Service” agreement has a choice of law provision specifying that California law governs any dispute between Facebook and the user. See Terms of Service, Facebook, https://www.facebook.com/terms.php (last visited Feb. 17, 2019). Although Facebook could change this clause to apply any other state’s law, it may nevertheless be that California law continues to apply as it would be the only possible location of the dispute since both the user and the platform’s headquarters are there.

Specific jurisdiction 305 See Adam R. Kleven, Note, Minimum Virtual Contacts: A Framework for Specific Jurisdiction in Cyberspace, 116 Mich. L. Rev. 785, 789–96 (2018) for general background on specific jurisdiction case law applying it to the internet. over Facebook for its actions toward users is a much more complicated question, and the small handful 306 The case law on personal jurisdiction over social media platforms is apparently extremely limited. See Gullen v. Facebook.com, Inc., No. 15 C 7681, 2016 WL 245910, at *2 (N.D. Ill. Jan. 21, 2016) (“The Court has not found a case analyzing specific jurisdiction over a social media website.”). of district courts that have addressed it have been skeptical. 307See Georgalis v. Facebook, Inc., 324 F. Supp. 3d 955, 961 (N.D. Ohio 2018); Ralls v. Facebook, 221 F. Supp. 3d 1237, 1243–44 (W.D. Wash. 2016); Gullen, 2016 WL 245910, at *2–3 (considering a challenge to Facebook’s facial recognition technology and rejecting that personal jurisdiction exists merely because Facebook operates an interactive website accessible in the forum state). This does not mean, however, that Facebook’s actions toward a user in a certain state could never establish specific jurisdiction. For example, it could be that Facebook taking active steps to delete a user’s account or posts is found analogous to physically reaching into that user’s state and silencing him, establishing the well-known requisite “minimum contacts.” 308 This could be the kind of “targeting” of a user in a forum state that the court in Gullen found lacking (which was fatal to personal jurisdiction there). See Gullen, 2016 WL 245910, at *2. Additionally, Facebook could not hide behind an algorithm applicable to all users, because many decisions to delete posts or accounts are ultimately targeted and specific to users. See Klonick, supra note 13, at 1635–49. It could also qualify under other possible tests for personal jurisdiction. See Isaacson, supra note 299, at 926–47 (discussing ways in which personal jurisdiction may apply to various new internet related technologies, including Facebook). Or, there could be personal jurisdiction under frameworks proposed by scholars. See Kleven, supra note 304, at 800–09 (proposing two-prong framework based on virtual contacts causing harm to plaintiff and reasonableness). The difficulty is that Facebook’s platform’s presence in cyberspace, rather than a physical location, complicates traditional personal jurisdiction notions, making it unclear “where” the interaction between Facebook and the user occurs. 309See Delic, supra note 299, at 487–89 (criticizing the applicability of traditional personal jurisdiction concepts to cloud computing).

This complication raises an important point: Personal jurisdiction law has historically adapted to changing technology, 310Id. at 476 n.49 (collecting articles discussing the relationship between personal jurisdiction and growth of internet); Allison MacDonald, Comment, YouTubing Down the Stream of Commerce: Eliminating the Express Aiming Requirement for Personal Jurisdiction in User-Generated Internet Content Cases, 19 Alb. L.J. Sci. & Tech. 519, 534–43 (2009). and the recent explosion in online activity generally and social media platforms specifically should be no exception. 311 Kleven, supra note 305, at 796–800 (criticizing narrowness of current personal jurisdiction rules as applied to internet activity); see also Isaacson, supra note 300, at 911–12 n.33 (collecting articles on personal jurisdiction and the internet). Given the seismic shift in how people and corporations interact with each other, a similar transition is likely in order now, though the exact parameters of what that change should look like are not clear and are ripe for scholarship. 312 An issue related to jurisdictional questions is that fifty different states may adopt fifty different frameworks, creating a convoluted regulatory scheme for social media platforms to navigate. For example, Facebook would have to know which state a particular user is a citizen of, and then know what sort of limitation on speech that state allows. This is not, however, a defect in this Comment’s proposal, but a strength. Any variation among states would create an incentive for Facebook to simply adopt a uniform policy that complies with all. In other words, variation would pressure Facebook to err on the side of caution by providing greater protection of speech and being less willing and likely to delete political speech.

B. Section 230 of the Communications Decency Act

Section 230 of the Communications Decency Act, 313 For general background on the law and its development, see Jaime M. Freilich, Note, Section 230’s Liability Shield in the Age of Online Terrorist Recruitment, 83 Brook. L. Rev. 675, 679–89 (2018); Klonick, supra note 13, at 1603–09; Lynn C. Percival, IV, Public Policy Favoritism in the Online World: Contract Voidability Meets the Communications Decency Act, 17 Tex. Wesleyan L. Rev. 165, 167–73 (2011); Emma M. Savino, Comment, Fake News: No One Is Liable, and That Is a Problem, 65 Buff. L. Rev. 1101, 1138–51 (2017); Olivier Sylvain, Intermediary Design Duties, 50 Conn. L. Rev. 203, 231–58 (2018). For a succinct summary of Section 230 as it may apply to Facebook, see Am. Freedom Def. Initiative v. Lynch, 217 F. Supp. 3d 100, 104–05 (D.D.C. 2016). which is far from an uncriticized law, 314See, e.g., Joshua N. Azriel, Social Networking as a Communications Weapon to Harm Victims: Facebook, MySpace, and Twitter Demonstrate a Need to Amend Section 230 of the Communications Decency Act, 26 J. Marshall J. Computer & Info. L. 415 (2009); Danielle Keats Citron & Benjamin Wittes, The Internet Will Not Break: Denying Bad Samaritans § 230 Immunity, 86 Fordham L. Rev. 401, 411–23 (2017); Freilich, supra note 312, at 689–91; Mary Graw Leary, The Indecency and Injustice of Section 230 of the Communications Decency Act, 41 Harv. J.L. & Pub. Pol’y 553 (2018); Michelle Roter, Note, With Great Power Comes Great Responsibility: Imposing a “Duty to Take Down” Terrorist Incitement on Social Media, 45 Hofstra L. Rev. 1379, 1394–98 (2017); Savino, supra note 313, at 1157–60; Sylvain, supra note 313, at 269–71. shields providers and users of “interactive computer services” 315 “Interactive computer service” is defined as “any information service, system, or access software provider that provides or enables computer access by multiple users to a computer server, including specifically a service or system that provides access to the Internet….” 47 U.S.C. § 230(f)(2) (2018). from liability for content generated by third parties, providing that they will not be treated as a “publisher or speaker” of any content generated by an “information content provider.” 316See id. § 230(c)(1). “Information content provider” is defined as “any person or entity that is responsible, in whole or in part, for the creation or development of information provided through the Internet or any other interactive computer service.” Id. § 230(f)(3). Section 230 also shields civil liability for voluntary, good faith restrictions of content that is “obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected.” 317Id. § 230(c)(2)(A). And, importantly, Section 230 preempts any contrary state law. 318See id. § 230(e)(3).

In other words, interactive computer service providers cannot be held liable for material generated by third parties—or user-generated content—unless the provider takes an extensive role in creating or editing content. 319See Freilich, supra note 312, at 682–84. A notable application of Section 230 is the proliferation of so-called “fake news” on social media sites, and the liability shield has been criticized in that context. See Dallas Flick, Comment, Combatting Fake News: Alternatives to Limiting Social Media Misinformation and Rehabilitating Quality Journalism, 20 SMU Sci. & Tech. L. Rev. 375 (2017); Savino, supra note 312, at 1157–60. Section 230’s accepted purpose is to promote free speech online by shielding platforms from liability, thus removing the fear of litigation and preventing overly cautious censoring of speech by websites. 320See supra note 311. Section 230’s findings are also illuminating. See 47 U.S.C. § 230(a)(3), (5) (2018) (“The Internet and other interactive computer services offer a forum for a true diversity of political discourse, unique opportunities for cultural development, and myriad avenues for intellectual activity …. Increasingly Americans are relying on interactive media for a variety of political, educational, cultural, and entertainment services.”). Accordingly, when it comes to social media platforms, Section 230 is most often considered and applied in the context of the platform failing to take down defamatory posts, rather than a user suing the platform itself because it deleted his post. 321See Savino, supra note 313, at 1144 n.276. This means that most litigation occurs under 47 U.S.C. § 230(c)(1) rather than 47 U.S.C. § 230(c)(2)(A). See also supra note 319 (discussing application to fake news).

Because of this, there is little case law where courts have applied Section 230 when a social media platform deletes a post and that post’s creator sues. 322 For discussion of the existing case law addressing the same statute, see Nicholas Conlon, Freedom to Filter Versus User Control: Limiting the Scope of § 230(c)(2) Immunity, 2014 U. Ill. J.L. Tech. & Pol’y 105, 116–22 (2014). Instead, in at least some cases where Section 230 has been asserted as a defense, courts have resolved the case on other grounds, primarily the lack of state action (as plaintiffs consistently bring First Amendment challenges, rather than the state constitutionalism approach developed herein). 323See Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 WL 1471939, at *5 (N.D. Cal. Mar. 26, 2018) (holding YouTube is not a state actor for First Amendment purposes); Shulman v. Facebook.com, No. 17-764 (JMV), 2017 WL 5129885, at *4 (D.N.J. Nov. 6, 2017). But, useful guidance comes from Song fi Inc. v. Google, Inc.324 108 F. Supp. 3d 876 (N.D. Cal. 2015). There, YouTube removed a music video because it alleged the view count was being artificially inflated, something it deemed “otherwise objectionable.” Id. at 880. which addresses Section 230’s liability shield for removing content that is “obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable.” 325 47 U.S.C. § 230(c)(2)(A) (2018). Specifically, the case addresses the meaning of “otherwise objectionable,” holding that the phrase does not include any content the provider dislikes or that poses a “problem” for the provider. 326Song fi Inc., 108 F. Supp. 3d at 882–84. Instead, “otherwise objectionable” must be defined based on the preceding list. 327Id. at 883. Applying this construction of Section 230 here, political speech almost certainly does not fall within its terms because it will rarely bear enough resemblance to obscene, violent, or harassing content. 328 Moreover, if such speech did fall into one of those categories, this Comment’s proposed approach would probably not provide protection of it. See supra Part II.C and accompanying notes.

Finally, four additional possible reasons may prevent Section 230 from applying. First, it could be that removals of political speech are not “good faith” restrictions within the above liability shield for the removal of objectionable content. 329 See Conlon, supra note 321, at 118 for a brief discussion of court interpretations of “good faith” within Section 230(c)(2)(A). Second, it is not settled law that Facebook and similar social media platforms are “interactive computer services,” though some courts have held that they are. 330See Klayman v. Zuckerberg, 753 F.3d 1354, 1357 (D.C. Cir. 2014); Cohen v. Facebook, Inc., 252 F. Supp. 3d 140, 156 n.10 (E.D.N.Y. 2017). Third, and similarly, the same is true of the question whether users posting content on such platforms are “information content providers.” 331See Klayman, 753 F.3d at 1358; Am. Freedom Def. Initiative v. Lynch, 217 F. Supp. 3d 100, 104–05 (D.D.C. 2016). In either case, courts could find otherwise, removing the social media platform or its users from Section 230’s reach. Fourth, given that Section 230’s purpose is to incentivize platforms to not delete posts, applying it in situations where the platform has deleted a post seems perverse and contrary to the intent of the statute. 332See Song fi Inc., 108 F. Supp. 3d at 883–84.

Conclusion

As plaintiffs suing social media platforms for deleting their posts and accounts increasingly get dismissed out of federal courts because of the state action doctrine, they should turn to another route. This Comment suggests one: state constitution free speech clauses. When in federal court bringing First Amendment claims, plaintiffs run into a wall. The United States Supreme Court has required state action for the First Amendment to kick in, and has rejected the arguments plaintiffs must make. But state courts do not have to follow this precedent when interpreting their own state constitution’s free speech clause, and the United States Supreme Court has affirmed that these state clauses can apply to certain private actors. Social media platforms qualify as one of these private actors, either through having become the functional equivalent of a public forum or having property rights that do not outweigh free speech rights when balanced.

Footnotes

*This Comment received the Myron Penn Laughlin Award for Excellence in Legal Research and Writing.

1 Most Famous Social Network Sites Worldwide as of October 2018, Ranked by Number of Active Users, Statista, https://www.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/ (last visited Jan. 14, 2019). See generally Social Media Fact Sheet, Pew Res. Ctr., http://www.pewinternet.org/fact-sheet/social-media/ (last accessed Jan. 14, 2019). Facebook recently broke two billion users after a decade of rapid growth, see Felix Richter, 3 in 10 World Citizens Are on Facebook, Statista (Apr. 25, 2019), https://www.statista.com/chart/10047/facebooks-monthly-active-users/, though its growth has plateaued recently, see Kurt Wagner & Rani Molla, Facebook’s User Growth Has Hit a Wall, Vox (July 25, 2018), https://www.vox.com/2018/7/25/17614426/facebook-fb-earnings-q2-2018-user-growth-troubles. This reflects a broader trend of plateaued social media growth in the last two years. Paul Hitlin, Internet, Social Media Use and Device Ownership in U.S. Have Plateaued After Years of Growth, Pew Res. Ctr. (Sept. 28, 2018), http://www.pewresearch.org/fact-tank/2018/09/28/internet-social-media-use-and-device-ownership-in-u-s-have-plateaued-after-years-of-growth/. This is not necessarily a sign of a decline in the power of social media, as it may be attributable to the fact that the platforms have reached near-saturation levels. Id.; see also infra note 233 (discussing inherent tendency toward scale of social media platforms).

2Cf. Mikhail Bulgakov, The Master and Margarita 69 (Diana Burgin & Katherine Tiernan O’Connor trans., Ardis Publishers 1995) (1967) (“But there were worse things to be seen in the bedroom: sprawled in a relaxed pose on the pouffe that had once belonged to the jeweler’s wife was a third creature, namely, a black cat of horrific proportions with a glass of vodka in one paw and in the other a fork on which he had speared a pickled mushroom.”) (description of the character “Behemoth”).

3 Aaron Smith & Monica Anderson, Social Media Use in 2018, Pew Res. Ctr. (Mar. 1, 2018), http://www.pewinternet.org/2018/03/01/social-media-use-in-2018/.

4Id.

5Id. at Appendix A.

6See Katerina Eva Matsa & Elisa Shearer, News Use Across Social Media Platforms 2018, Pew Res. Ctr. (Sept. 10, 2018), http://www.journalism.org/2018/09/10/news-use-across-social-media-platforms-2018/.

7 Elisa Shearer, Social Media Outpaces Print Newspapers in the U.S. as a News Source, Pew Res. Ctr. (Dec. 10, 2018), http://www.pewresearch.org/fact-tank/2018/12/10/social-media-outpaces-print-newspapers-in-the-u-s-as-a-news-source/.

8See Jeffrey Gottfried & Elisa Shearer, Americans’ Online News Use Is Closing in on TV News Use, Pew Res. Ctr. (Sept. 7, 2017), http://www.pewresearch.org/fact-tank/2017/09/07/americans-online-news-use-vs-tv-news-use/.

9See Benjamin F. Jackson, Censorship and Freedom of Expression in the Age of Facebook, 44 N.M. L. Rev. 121, 125–26 (2014); Loren F. Selznick & Carolyn LaMacchia, #Mall Ruckus Tonight: Should Mall Owners be Forced to Provide a Stage for Expression in the Virtual Age?, 53 Willamette L. Rev. 239, 271–73 (2017); Monica Anderson et al., Activism in the Social Media Age, Pew Res. Ctr. (July 11, 2018), http://www.pewinternet.org/2018/07/11/activism-in-the-social-media-age/.

10See Jackson, supra note 9, at 124. For example, the Obama presidential campaigns in 2008 and 2012 used social media, and the internet generally, to great effect. See, e.g., Ed Pilkington & Amanda Michel, Obama, Facebook and the Power of Friendship: The 2012 Data Election, Guardian (Feb. 17, 2012), https://www.theguardian.com/world/2012/feb/17/obama-digital-data-machine-facebook-election.

11 Packingham v. North Carolina, 137 S. Ct. 1730, 1735 (2017). The Court went on to praise Facebook’s capacity for unlimited, low-cost communication, including political speech. Id. at 1735–36; see also Reno v. ACLU, 521 U.S. 844, 870 (1997) (“Through the use of chat rooms, any person with a phone line can become a town crier with a voice that resonates farther than it could from any soapbox.”).

12Packingham, 137 S. Ct. at 1737 (“North Carolina with one broad stroke bars access to what for many are the principal sources for knowing current events, checking ads for employment, speaking and listening in the modern public square, and otherwise exploring the vast realms of human thought and knowledge…. [T]o foreclose access to social media altogether is to prevent the user from engaging in the legitimate exercise of First Amendment rights.”). Interestingly, some courts have found that government social media pages are public forums for First Amendment purposes, at least when they have some interactive component. See Davison v. Randall, 912 F.3d 666, 682, 688 (4th Cir. 2019); Knight First Amendment Inst. v. Trump, 302 F. Supp. 3d 541, 574 (S.D.N.Y. 2018). But see Morgan v. Bevin, 298 F. Supp. 3d 1003, 1010–11 (E.D. Ky. 2018).

13See, e.g., Hudgens v. NLRB, 424 U.S. 507, 513 (1976) (“It is, of course, a commonplace that the constitutional guarantee of free speech is a guarantee only against abridgment by government, federal or state.”); Shulman v. Facebook.com, No. 17-764 (JMV), 2017 WL 5129885, at *4 (D.N.J. Nov. 6, 2017) (collecting cases of failed attempts to subject Facebook to First Amendment); Kate Klonick, The New Governors: The People, Rules, and Processes Governing Online Speech, 131 Harv. L. Rev. 1598, 1609–13 (2018) (discussing several current arguments regarding state action and the First Amendment). In recent litigation involving alleged censorship by YouTube, a federal district court dismissed based in part on YouTube not being a state actor and thus not subject to the First Amendment. Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 WL 1471939, at *8–9 (N.D. Cal. Mar. 26, 2018). In doing so, the district court provided a helpful summary of the current state of First Amendment jurisprudence as it relates to online platforms. See id. at *5–8. In a recent case involving alleged mismanagement of public access channels by a private nonprofit corporation, the Supreme Court held that simply operating public access channels did not turn the corporation into a “state actor,” and the corporation was thus not subject to the First Amendment. Manhattan Cmty. Access Corp. v. Halleck, 139 S. Ct. 1921, 1934 (2019). This suggests the Court is unlikely to expand the state action doctrine in the near future.

14See Jonathan Peters, The “Sovereigns of Cyberspace” and State Action: The First Amendment’s Application—or Lack Thereof—to Third-Party Platforms, 32 Berkeley Tech. L.J. 989, 1013–18 (2017). Senator Josh Hawley of Missouri has proposed legislation that would subject social media platforms to audits and strip certain legal protections from them if they censor viewpoints. See Senator Hawley Introduces Legislation to Amend Section 230 Immunity for Big Tech Companies, Josh Hawley (June 19, 2019), https://www.hawley.senate.gov/senator-hawley-introduces-legislation-amend-section-230-immunity-big-tech-companies. It is unclear whether this kind of legislation has any chance of becoming law, and this Comment’s focus is on case law and judiciaries, not statutes and legislatures.

15See Klonick, supra note 13, at 1616–29.

16See Jackson, supra note 9, at 129–31. It is no secret that Silicon Valley employees, for better or worse, skew heavily toward one end of the political spectrum. See, e.g., Dan Boylan, Silicon Valley an “Extremely Left-Leaning Place,” Admits Zuckerberg, Wash. Times (Apr. 10, 2018), https://www.washingtontimes.com/news/2018/apr/10/zuckerberg-admits-silicon-valley-extremely-left-le/; Kate Conger & Sheera Frenkel, Dozens at Facebook United to Challenge its “Intolerant” Liberal Culture, N.Y. Times (Aug. 28, 2018), https://www.nytimes.com/2018/08/28/technology/inside-facebook-employees-political-bias.html; Kirsten Grind & Keach Hagey, Why Did Facebook Fire a Top Executive? Hint: It Had Something To Do with Trump, Wall Street J. (Nov. 11, 2018), https://www.wsj.com/articles/why-did-facebook-fire-a-top-executive-hint-it-had-something-to-do-with-trump-1541965245; Farhad Manjoo, Silicon Valley’s Politics: Liberal, with One Big Exception, N.Y. Times (Sept. 6, 2017), https://www.nytimes.com/2017/09/06/technology/silicon-valley-politics.html.

17 A recent example of something approaching this was the collective action to delete Alex Jones’s accounts across platforms, including Facebook, Apple products, Spotify, YouTube, and, eventually, Twitter. See Kate Conger & Jack Nicas, Twitter Bars Alex Jones and Infowars, Citing Harassing Messages, N.Y. Times (Sept. 6, 2018), https://www.nytimes.com/2018/09/06/technology/twitter-alex-jones-infowars.html; Jack Nicas, Alex Jones and Infowars Content Is Removed From Apple, Facebook and YouTube, N.Y. Times (Aug. 6, 2018), https://www.nytimes.com/2018/08/06/technology/infowars-alex-jones-apple-facebook-spotify.html. Because Jones trafficked in ridiculous and demonstrably false conspiracy theories, some of which included the potential defamation of Sandy Hook parents, he falls outside any protection this Comment proposes. See also Rachel Kraus, 2018 Was the Year We (Sort of) Cleaned Up the Internet, Mashable (Dec. 26, 2018), https://mashable.com/article/deplatforming-alex-jones-2018/#GD8mPzDxkqq4 (documenting unsavory figures deplatformed in 2018).

18Id.; cf. Disinvitation Database, Fire, https://www.thefire.org/resources/disinvitation-database/ (last visited Aug. 9, 2019) (compiling speakers prevented from speaking on college campuses and showing recent increase).

19 Jason Koebler, Deplatforming Works, Vice (Aug. 10, 2018), https://vice.com/en_us/article/bjbp9d/do-social-media-bans-work.

20See Klonick, supra note 13, at 1662–70; Peters, supra note 14, at 993.

21 There is no shortage of ideologically diverse writing praising the value of free speech. See John Stuart Mill, All Minus One: John Stuart Mill’s Ideas on Free Speech Illustrated (Richard V. Reeves & Jonathan Haidt eds., 2018); Freedom of Expression, ACLU, https://www.aclu.org/other/freedom-expression (last visited Aug. 9, 2019); Harvey C. Mansfield, The Value of Free Speech, Nat’l Aff. (Fall 2018), https://www.nationalaffairs.com/publications/detail/the-value-of-free-speech. There is, of course, a range of views that spawn from those who are more absolutist to those who see more room for government regulation of things like “hate speech.” See generally Michael Kent Curtis, Critics of “Free Speech” and the Uses of the Past, 12 Const. Comment. 29, 42–65 (1995) (discussing various criticisms of free speech and responding to those criticisms). This Comment will not relitigate battles over whether free speech is a virtue and will simply assume it is. If the reader does not buy that free speech is a good, then they are beyond hope and this Comment will not save them.

22 Jackson, supra note 9, at 146–54; Trevor Puetz, Note, Facebook: The New Town Square, 44 Sw. L. Rev. 385, 396–409 (2014).

23 Peters, supra note 14, at 1022–26.

24 At least two articles have mentioned state constitutionalism as an option. See Jackson, supra note 9, at 157–59; Dawn C. Nunziato, The Death of the Public Forum in Cyberspace, 20 Berkeley Tech. L.J. 1115, 1164–67 (2005). However, both included them as briefly mentioned alternatives to their main arguments, and neither fully explored the relevant state case law or developed a fleshed-out approach based on it. Additionally, at least one plaintiff has raised a state constitutional argument against YouTube (in addition to First Amendment arguments), though the federal court declined to exercise supplemental jurisdiction over the issue. See Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 WL 1471939, at *13–14 (N.D. Cal. Mar. 26, 2018).

25See Jeffrey S. Sutton, 51 Imperfect Solutions: States and the Making of American Constitutional Law (2018); William J. Brennan, Jr., State Constitutions and the Protection of Individual Rights, 90 Harv. L. Rev. 489, 502 (1977).

26See Sutton, supra note 25, at 174–75.

27See id. at 19 (“[I]t often will be easier to vindicate a state constitutional right in a State’s high court than a federal constitutional right in the U.S. Supreme Court.”); Brennan, supra note 25, at 495.

28 William J. Brennan, Jr., The Bill of Rights and the States: The Revival of State Constitutions as Guardians of Individual Rights, 61 N.Y.U. L. Rev. 535, 548–49 (1986); see also Sutton, supra note 25, at 22–172 (detailing four instances, including compelled sterilization, in which state courts provided greater protection than federal courts).

29See, e.g., Sutton, supra note 25; Brennan, supra note 25, at 502; Brennan, supra note 28, 548–52; Paul E. McGreal, Alaska Equal Protection: Constitutional Law or Common Law?, 15 Alaska L. Rev. 209, 213–14 nn.11–12 (1998) (gathering various articles); Jeffrey S. Sutton, State Constitutions in the United States Federal System, 77 Ohio St. L.J. 195, 197-99 (2016) (symposium introduction).

30See William H. Pryor, Jr., The Importance of State Constitutions, Nat’l Rev. (June 7, 2018), https://www.nationalreview.com/2018/06/state-constitutions-important-components-of-federalism/ (Judge William H. Pryor, Jr., Circuit Judge for the Eleventh Circuit of the United States Court of Appeals); John Paul Stevens, The Other Constitutions, N.Y. Rev. Books (Dec. 2018), https://www.nybooks.com/articles/2018/12/06/other-constitutions/ (former Associate Justice of the Supreme Court John Paul Stevens). The book also features blurbs from such diverse legal thinkers as Laurence Tribe and William Baude.

31Sutton, supra note 25.

32Id. at 16.

33Id. at 16.

34Id. at 16–17. Sutton also notes that the United States Supreme Court may face a “federalism discount,” or the underenforcement of federal constitutional law, because it must be applied to all fifty states; this is a concern not present for states. Id. at 17; see also Brennan, supra note 28, at 549.

35Sutton, supra note 25, at 18. This also is the heart of the book’s title. As Judge Sutton says, “it may be more appropriate to tolerate fifty-one imperfect solutions rather than to impose one imperfect solution on the country as a whole.” Id. at 19; see also Brennan, supra note 28, at 551.

36Sutton, supra note 25, at 17–18.

37Id. at 19–20, 178.

38Id. at 211–12. Sutton maintains that state constitutionalism is compatible with all forms of judicial interpretation but criticizes Justice Brennan’s argument because it encouraged state courts to diverge from federal courts “so long as there is a progressive will.” Id. at 177. In Sutton’s view, this, combined with Brennan’s liberalism, may have alienated conservative judges and commentators at the time. Id. at 176–77. Regardless of the historical veracity of this, state constitutionalism is compatible with the textual originalism that is typically associated with conservative legal thinkers. See Jeffrey S. Sutton, 51 Imperfect Solutions: States and the Making of American Constitutional Law, 132 Harv. L. Rev. 811, 815–18 (2018) (book review); see also Brennan, supra note 28, at 550.

39 A state-based approach is particularly timely as former Attorney General Jeff Sessions recently met with a number of state attorneys general to discuss the topic. See Jessica Guynn, Sessions to Meet With State AGs on Charges Social Media Companies Stifle Free Speech, Competition, USA Today (Sept. 5, 2018, 10:51 PM), https://usatoday.com/story/tech/news/2018/09/05/sessions-meet-state-attorneys-general-charges-social-media-companies-hurt-free-speech/1204915002/.

40 Although state courts are free to interpret textually identical clauses differently from the corresponding United States Constitution clause, see Sutton, supra note 25, at 174–75, textual differences provide a strong reason to do so.

41 Bock v. Westminster Mall Co., 819 P.2d 55, 58 (Colo. 1991) (en banc).

42 Though, of course, a state court would be perfectly within its right to do this. See Sutton, supra note 25, at 174–75.

43 Of course, each state supreme court must still engage in whatever brand of constitutional analysis required by its precedent. For examples of how this might look, see Maxim Cabaret, Inc. v. City of Sandy Springs, 816 S.E.2d 31, 39–40 (Ga. 2018) (Peterson, J., concurring), which addresses Georgia’s free speech clause specifically, and Elliott v. State, 824 S.E.2d 265, 267–70, 273 (Ga. 2019) and Olevik v. State, 806 S.E.2d 505, 514–15 (Ga. 2017), which address state constitutional interpretation generally and at length.

44 U.S. Const. amend. I.

45Haw. Const. art. I, § 4; S.C. Const. art. I, § 2.

46Ind. Const. art. I, § 9; Or. Const. art. I, § 8; R.I. Const. art. I, § 21; Utah Const. art. I, § 15; W. Va. Const. art. III, § 7.

47Ala. Const. art. I, § 4; Alaska Const. art. I, § 5; Ariz. Const. art. II, § 6; Cal. Const. art. I, § 2, cl. a; Colo. Const. art. II, § 10; Conn. Const. art. I, §§ 4–5; Fla. Const. art. I, § 4; Ga. Const. art. I, § 1, para. 5; Idaho Const. art. I, § 9; Ill. Const. art. I, § 4; Iowa Const. art. I, § 7; Kan. Const. Bill of Rights § 11; La. Const. art. I, § 7; Me. Const. art. I, § 4; Mich. Const. art. I, § 5; Minn. Const. art. I, § 3; Mo. Const. art. I, § 8; Mont. Const. art. II, § 7; Neb. Const. art. I, § 5; Nev. Const. art. I, § 9; N.J. Const. art. I, para. 6; N.M. Const. art. II, § 17; N.Y. Const. art. I, § 8; N.D. Const. art. I, § 4; Ohio Const. art. I, § 11; Okla. Const. art. II, § 22; S.D. Const. art. VI, § 5; Tex. Const. art. I, § 8; Va. Const. art. I, § 12; Wash. Const. art. I, § 5; Wis. Const. art. I, § 3; Wyo. Const. art. I, § 20. Along these lines, many state constitutions include language that expressly protects defendants in libel prosecutions—both private and public—when they can prove that their statements were true. For purposes of this analysis, this aspect of these provisions does not require detailed accounting.

48 Some constitutions use words other than “impair” like “abridge” or “curtail” or “restrain” or “curtail or restrain.” For purposes of this analysis, these variations are not important.

49Ala. Const. art. I, § 4; Cal. Const. art. I, § 2, cl. a; Colo. Const. art. II, § 10; Conn. Const. art. I, §§ 4–5; Fla. Const. art. I, § 4; Ga. Const. art. I, § 1, para. 5; Iowa Const. art. I, § 7; La. Const. art. I, § 7; Mich. Const. art. I, § 5; Mo. Const. art. I, § 8; Mont. Const. art. II, § 7; Nev. Const. art. I, § 9; N.J. Const. art. I, para. 6; N.M. Const. art. II, § 17; N.Y. Const. art. I, § 8; Ohio Const. art. I, § 11; Okla. Const. art. II, § 22; Tex. Const. art. I, § 8; Va. Const. art. I, § 12; Wis. Const. art. I, § 3.

50Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Ky. Const. Bill of Rights § 8; Md. Const. Declaration of Rights art. 40; Pa. Const. art. I, § 7; Tenn. Const. art. I, § 19.

51Mass. Const. pt. I, art. 16; Miss. Const. art. III, § 13; N.H. Const. pt. I, art. 22; N.C. Const. art. I, § 14; Vt. Const. ch. I, art. XIII.

52See Maxim Cabaret, Inc. v. City of Sandy Springs, 816 S.E.2d 31, 39 (Ga. 2018) (Peterson, J., concurring).

53See supra note 47.

54See supra note 49.

55See Curious Theatre Co. v. Colo. Dep’t of Pub. Health & Env’t, 220 P.3d 544, 551 (Colo. 2009); Bock v. Westminster Mall Co., 819 P.2d 55, 57–58 (Colo. 1991) (en banc).

56See Golden Gateway Ctr. v. Golden Gateway Tenants Ass’n, 29 P.3d 797, 817–19 (Cal. 2001) (Werdegar, J., dissenting) (“[T]he latter proviso neither grammatically nor legally qualifies the simple and sweeping free speech guarantee with which [the] section … begins.”).

57See id. at 804 (plurality opinion). It is worth noting that questions of framer intent present a complicated issue of constitutional interpretation. It may be that such intent, unless unequivocally clear, should not override the plain text of these state free speech clauses. See Cologne v. Westfarms Assocs., 469 A.2d 1201, 1215–16 (Conn. 1984) (Peters, J., dissenting) (“These principles of construction would require … the intent of the draftsmen would not supersede the clear language of the constitution itself unless the intent of the framers were indisputably clear with respect to the particular provision arguably subject to implied limitation.”). It may also be that what the framers intended the words of a clause to mean is only a data point considered in a larger effort to discern the objective public meaning at the time of ratification. See Olevik v. State, 806 S.E.2d 505, 514–15 (Ga. 2017). And, such extratextual intent, if relied upon, would appropriately be discerned by reference to the relevant constitutional convention. See Golden Gateway Ctr., 29 P.3d at 804–06 (plurality opinion). This Comment takes no position on these questions of constitutional interpretation, but merely flags the issue.

58Mich. Const. art. I, § 5; Nev. Const. art. I, § 9; N.M. Const. art. II, § 17; N.Y. Const. art. I, § 8; Ohio Const. art. I, § 11; Okla. Const. art. II, § 22; Tex. Const. art. I, § 8; Wis. Const. art. I, § 3. Connecticut’s Constitution separates the two clauses entirely, putting “No law shall ever be passed to curtail or restrain the liberty of speech or of the press” in a separate section, Conn. Const. art. I, § 5, and the two have been interpreted distinctly. See Cologne, 469 A.2d at 1208–09. Therefore, this survey does not include Connecticut’s free speech clause in this category.

59Alaska Const. art. I, § 5; Ariz. Const. art. II, § 6; Idaho Const. art. I, § 9; Ill. Const. art. I, § 4; Kan. Const. Bill of Rights § 11; Me. Const. art. I, § 4; Minn. Const. art. I, § 3; Neb. Const. art. I, § 5; N.D. Const. art. I, § 4; S.D. Const. art. VI, § 5; Wash. Const. art. I, § 5; Wyo. Const. art. I, § 20. The section of Maine’s Constitution that includes the free speech clause does have negative “no laws” language, but it is specific only to the separate “freedom of the press” clause in that section. Me. Const. art. I, § 4; cf. City of Portland v. Jacobsky, 496 A.2d 646, 648 (Me. 1985) (quoting only the free speech clause without the negative language in an obscenity case).

60See supra note 56.

61See Southcenter Joint Venture v. Nat’l Democratic Policy Comm., 780 P.2d 1282, 1296 (Wash. 1989) (en banc) (Utter, J., concurring in the result) (“[F]rom a purely linguistic point of view, removing … language referring specifically to acts of state has great effect. Quite simply, it changes the facial meaning of the provision to state it in the absolute rather than in terms of state action.”); cf. Sharrock v. Dell Buick-Cadillac, Inc., 379 N.E.2d 1169, 1173–74 (N.Y. 1978) (“In contrast to the due process clause of the Fourteenth Amendment … Conspicuously absent from the [New York] State Constitution is any language requiring State action before an individual may find refuge in its protections. That is not to say, of course, that the due process clause of the State Constitution eliminates the necessity of any State involvement …. Rather, the absence … simply provides a basis to apply a more flexible State involvement requirement than … the Federal provision.”).

62Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Ky. Const. Bill of Rights § 8; Md. Const. Declaration of Rights art. 40; Pa. Const. art. I, § 7; Tenn. Const. art. I, § 19.

63Pa. Const. art. I, § 7.

64See id.

65Ky. Const. Bill of Rights § 8; Tenn. Const. art. I, § 19.

66Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Md. Const. Declaration of Rights art. 40.

67See supra note 66.

68See supra notes 63 and 65.

69N.H. Const. pt. I, art. XXII.

70Mass. Const. pt. I, art. XVI, amended by Mass. Const. art. LXXVII.

71N.C. Const. art. I, § 14.

72Miss. Const. art. III, § 13.

73Vt. Const. ch. I, art. 13.

74See Golden Gateway Ctr. v. Golden Gateway Tenants Ass’n, 29 P.3d 797, 804–06 (Cal. 2001) (plurality opinion).

75Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Ky. Const. Bill of Rights § 8; Md. Const. Declaration of Rights art. 40; Pa. Const. art. I, § 7; Tenn. Const. art. I, § 19.; N.H. Const. pt. I, art. XXII; Mass. Const. pt. I, art. XVI, amended by Mass. Const. art. LXXVII; N.C. Const. art. I, § 14; Miss. Const. art. III, § 13; Vt. Const. ch. I, art. 13; Alaska Const. art. I, § 5; Ariz. Const. art. II, § 6; Idaho Const. art. I, § 9; Ill. Const. art. I, § 4; Kan. Const. Bill of Rights § 11; Me. Const. art. I, § 4; Minn. Const. art. I, § 3; Neb. Const. art. I, § 5; N.D. Const. art. I, § 4; S.D. Const. art. VI, § 5; Wash. Const. art. I, § 5; Wyo. Const. art. I, § 20.

76Ala. Const. art. I, § 4; Alaska Const. art. I, § 5; Ariz. Const. art. II, § 6; Cal. Const. art. I, § 2, cl. a; Colo. Const. art. II, § 10; Conn. Const. art. I, §§ 4–5; Fla. Const. art. I, § 4; Ga. Const. art. I, § 1, para. 5; Idaho Const. art. I, § 9; Ill. Const. art. I, § 4; Iowa Const. art. I, § 7; Kan. Const. Bill of Rights § 11; La. Const. art. I, § 7; Me. Const. art. I, § 4; Mich. Const. art. I, § 5; Minn. Const. art. I, § 3; Mo. Const. art. I, § 8; Mont. Const. art. II, § 7; Neb. Const. art. I, § 5; Nev. Const. art. I, § 9; N.J. Const. art. I, para. 6; N.M. Const. art. II, § 17; N.Y. Const. art. I, § 8; N.D. Const. art. I, § 4; Ohio Const. art. I, § 11; Okla. Const. art. II, § 22; S.D. Const. art. VI, § 5; Tex. Const. art. I, § 8; Va. Const. art. I, § 12; Wash. Const. art. I, § 5; Wis. Const. art. I, § 3; Wyo. Const. art. I, § 20; Ark. Const. art. II, § 6; Del. Const. art. I, § 5; Ky. Const. Bill of Rights § 8; Md. Const. Declaration of Rights art. 40; Pa. Const. art. I, § 7; Tenn. Const. art. I, § 19; Mass. Const. pt. I, art. 16; Miss. Const. art. III, § 13; N.H. Const. pt. I, art. 22; N.C. Const. art. I, § 14; Vt. Const. ch. I, art. XIII.

77 Whether states with free speech clauses textually similar to the First Amendment may nevertheless choose to expand those clauses—based on evolving modern standards or otherwise—to private actors is another question, and the arguments for state constitutionalism in the introduction are applicable.

78See Jackson, supra note 9, at 142–53 for a discussion of the exceptions to the state action doctrine.

79See Hudgens v. NLRB, 424 U.S. 507, 513, 520–21 (1976) (“It is, of course, a commonplace that the constitutional guarantee of free speech is a guarantee only against abridgment by government, federal or state.”) (holding that the First Amendment does not apply to private shopping centers); Tsao v. Desert Palace, Inc., 698 F.3d 1128, 1140 (9th Cir. 2012) (noting the four narrow exceptions as the public function test, joint action test, state compulsion test, and governmental nexus test). See generally David M. Howard, Rethinking State Inaction: An In-Depth Look at the State Action Doctrine in State and Lower Federal Courts, 16 Conn. Pub. Int. L.J. 221, 221–40 (2017) (discussing state action doctrine and situations when private action becomes state action).

80 447 U.S. 74 (1980).

81Id. at 88. Also, the Court has denied petitions for certiorari in a number of subsequent cases raising the same or similar issues. See Ralphs Grocery Co. v. United Food & Commercial Workers Union Local 8, 290 P.3d 1116 (Cal. 2012), cert. denied, 133 S. Ct. 2799 (2013); Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742 (Cal. 2007), cert. denied, 555 U.S. 819 (2008); N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757 (N.J. 1994), cert. denied sub nom. Short Hills Assocs. v. N.J. Coal. Against War in the Middle E., 516 U.S. 812 (1995).

82 For consistency, when this Comment discusses cases involving shopping centers or otherwise references shopping centers, it is referring to centers with size comparable to the one involved in Pruneyard, unless expressly noted.

83Pruneyard Shopping Ctr., 447 U.S. at 77.

84Id. at 77–78.

85Id. at 77.

86Id.

87Id.

88 This decision’s reasoning is discussed below. See infra Part II and accompanying notes.

89Pruneyard Shopping Ctr., 447 U.S. at 78 (citing Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979)).

90Id. at 82, 85.

91Id. at 83.

92Id.

93Id. at 84.

94Id. at 84–85.

95Id. at 87.

96Id.

97Id.

98 418 U.S. 241 (1974).

99Pruneyard Shopping Ctr., 447 U.S. at 88.

100Id.

101 It is worth noting that Pruneyard was the culmination of a line of U.S. Supreme Court cases that first adopted reasoning similar to the reasoning of state courts discussed in Part II, see Amalgamated Food Emps. Union Local 590 v. Logan Valley Plaza, Inc., 391 U.S. 308, 319–20 (1968), then narrowed it, see Lloyd Corp. v. Tanner, 407 U.S. 551, 561–62 (1972), then overruled it, see Hudgens v. NLRB, 424 U.S. 507, 518 (1976). Because this Comment is focused on state constitutions, which need not be interpreted similarly to the First Amendments, a detailed discussion of these cases and the First Amendment would muddy the waters. For such a discussion, see Selznick & LaMacchia, supra note 9, at 252–54.

102 Bock v. Westminster Mall Co., 819 P.2d 55, 58 (Colo. 1991) (en banc).

103 Dublirer v. 2000 Linwood Ave. Owners, Inc., 103 A.3d 249 (N.J. 2014); Green Party of N.J. v. Hartz Mountain Indus., Inc., 752 A.2d 315 (N.J. 2000); N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757 (N.J. 1994), cert. denied sub nom. Short Hills Assocs. v. N.J. Coal. Against War in the Middle E., 516 U.S. 812 (1995); State v. Schmid, 423 A.2d 615 (N.J. 1980), appeal dismissed sub nom. Princeton Univ. v. Schmid, 455 U.S. 100 (1982).

104 Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742 (Cal. 2007), cert. denied, 555 U.S. 819 (2008); Robins v. Pruneyard Shopping Ctr., 592 P.2d 341 (Cal. 1979).

105 Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 533 (Wash. 1999) (en banc) (relying on state constitution’s initiative clause); Southcenter Joint Venture v. Nat’l Democratic Policy Comm., 780 P.2d 1282, 1289, 1291 (Wash. 1989) (en banc) (rejecting Alderwood plurality); Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108 (Wash. 1981) (en banc) (plurality opinion).

106Bock, 819 P.2d at 58.

107 W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331 (Pa. 1986); Commonwealth v. Tate, 432 A.2d 1382 (Pa. 1981).

108 At least two states have, at some point, found protections similar to those in Pruneyard, except under state constitution clauses other than free speech. See Glovsky v. Roche Bros. Supermarkets, Inc., 17 N.E.3d 1026 (Mass. 2014); Batchelder v. Allied Stores Int’l, Inc., 445 N.E.2d 590 (Mass. 1983); Lloyd Corp. v. Whiffen, 849 P.2d 446 (Or. 1993), abrogated by Stranahan v. Fred Meyer, Inc., 11 P.3d 228, 243 (Or. 2000). Washington has done this also and is addressed in the body.

109See Roman v. Trs. of Tufts Coll., 964 N.E.2d 331, 338 (Mass. 2012) (noting that state constitutions can provide more expansive protections of rights than the federal constitution and that the Massachusetts Supreme Court has approvingly cited cases from other jurisdictions protecting free speech rights on private colleges).

110See, e.g., Pruneyard Shopping Ctr. v. Robins, 447 U.S. 74, 77 (1980); Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 744 (Cal. 2007); Bock, 819 P.2d at 56; Green Party of N.J. v. Hartz Mountain Indus., Inc., 752 A.2d 315, 319 (N.J. 2000); N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 775 (N.J. 1994), cert. denied sub nom. Short Hills Assocs. v. N.J. Coal. Against War in the Middle E., 516 U.S. 812 (1995); State v. Schmid, 423 A.2d 615, 616 (N.J. 1980), appeal dismissed sub nom. Princeton Univ. v. Schmid, 455 U.S. 100 (1982); Tate, 432 A.2d at 1385.

111See, e.g., W. Pa. Socialist Workers, 515 A.2d at 1333; Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 525 (Wash. 1999) (en banc); Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108, 110 (Wash. 1981) (plurality opinion).

112 Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 342 (Cal. 1979).

113N.J. Coal., 650 A.2d at 762; Tate, 432 A.2d at 1384; see also Bock, 819 P.2d at 56 (“[P]ledging non-violent dissent from the federal government’s foreign policy toward Central America.”).

114Fashion Valley Mall, 172 P.3d at 743.

115Green Party of N.J., 752 A.2d at 319; Schmid, 423 A.2d at 616; see also Dublirer v. 2000 Linwood Ave. Owners, Inc., 103 A.3d 249, 250 (N.J. 2014) (campaigning for the board of a private cooperative apartment).

116N.J. Coal., 650 A.2d at 781. In the same case, the court emphasized that the speech interest involved was “the most substantial in our constitutional scheme.” Id. at 776. Also, in a previous case applying its free speech clause to a private college, the Supreme Court of New Jersey noted that political speech is particularly important and worthy of protection. See Schmid, 423 A.2d at 627. The Supreme Court of New Jersey reiterated this sentiment in a subsequent case. See Green Party of N.J., 752 A.2d at 328.

117Tate, 432 A.2d at 1390 (emphasis added).

118Bock, 819 P.2d at 57. Moreover, this free speech right holds “high rank … in the constellation of freedoms guaranteed by both the United States Constitution and [Colorado’s] state constitution.” Id.

119 Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979). The Supreme Court of California later relied on only the state’s free speech clause to hold a shopping center could not prohibit union related boycotting of its stores. Fashion Valley Mall, 172 P.3d at 754.

120See Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108, 117 (Wash. 1981) (plurality opinion).

121See Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 529–30 (Wash. 1999) (en banc); Southcenter Joint Venture v. Nat’l Democratic Policy Comm., 780 P.2d 1282, 1290 (Wash. 1989) (en banc).

122 Generally speaking, commercial speech holds a position inferior to political speech under the First Amendment. See e.g., Hoffman v. Capital Cities/ABC, Inc., 255 F.3d 1180, 1184–85 (9th Cir. 2001) (“[Commercial speech] is entitled to a measure of First Amendment protection … however, [it] do[es] not receive the same level of constitutional protection as other types of protected expression.”). Notably, although content-based restrictions are more likely to survive the First Amendment when they affect commercial speech, the Supreme Court has been skeptical of viewpoint-based restrictions and has “specifically declined to recognize a distinction between commercial and noncommercial speech that would render” the “offensiveness” of speech a government interest sufficient to justify prohibiting it. Bolger v. Youngs Drug Prods. Corp., 463 U.S. 60, 71–72 (1983). See also Iancu v. Brunetti, 139 S.Ct. 2294 (2019); Matal v. Tam, 137 S.Ct. 1744 (2017).

123See N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 781 (N.J. 1994) (“Commercial free speech at regional and community shopping centers is fundamentally so discordant with the purposes and uses of those centers as to disqualify it from constitutional protection.”). The specifics of a balancing test are discussed infra Part II.C.

124See N.J. Coal., 650 A.2d at 782 (“[C]ommercial free speech could obviously be directly in conflict with the centers’ activities, uses, and success, the most obvious example being leafletting seeking to persuade shoppers and non-shoppers to go elsewhere.”).

125See id. at 781 (“[Commercial speech] is generally discordant: the owners … carefully plan their merchandising strategy, their advertising programs, and are entitled to reap the rewards of their efforts without commercial interference, even well-intentioned commercial interference, from others.”).

126See id. at 782 (“[T]he sale of literature and the solicitation of funds on the spot (as distinguished from appeals found in the leaflets themselves) are not covered by the protection.”); see also Southcenter Joint Venture, 780 P.2d at 1306 (Utter, J., concurring in the result) (noting that “[t]he nature of this speech activity competed directly with the property interests of the mall owners” because it solicited contributions and literature sales, which distinguished it from the solicitation of signatures in Alderwood and made it incompatible with the uses of the mall and thus unprotected).

127See, e.g., Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979); Bock v. Westminster Mall Co., 819 P.2d 55, 62–63 (Colo. 1991) (en banc); N.J. Coal., 650 A.2d at 783; State v. Schmid, 423 A.2d 615, 630 (N.J. 1980); Commonwealth v. Tate, 432 A.2d 1382, 1390 (Pa. 1981); Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108, 116–17 (Wash. 1981) (plurality opinion).

128 172 P.3d 742 (Cal. 2007).

129Id. at 750.

130See id. at 743–44.

131 This “functional equivalent” argument has also been central in cases using clauses other than free speech to similar effect. See Glovsky v. Roche Bros. Supermarkets, Inc., 17 N.E.3d 1026, 1031 (Mass. 2014); Batchelder v. Allied Stores Int’l, Inc., 445 N.E.2d 590, 597 (Mass. 1983); Lloyd Corp. v. Whiffen, 849 P.2d 446, 465 (Or. 1993), abrogated by Stranahan v. Fred Meyer, Inc., 11 P.3d 228 (Or. 2000).

132See Alderwood Assocs., 635 P.2d at 114 (“Then, most public forums were located in downtown business districts, whereas now people spend more time in shopping centers than anywhere else, excepting their home and job.”). For a fun portrayal of this cultural phenomenon, see its depiction in season three of the mega-hit Netflix show Stranger Things, which take place in the 1980s. See Stranger Things (Netflix 2019).

133 N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 767 (N.J. 1994).

134See id. at 766–68 (detailing extensive empirical evidence and citing academic support); Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 344–45 (Cal. 1979); Bock v. Westminster Mall Co., 819 P.2d 55, 61 (Colo. 1991); W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331, 1336–37 (Pa. 1986); Alderwood Assocs., 635 P.2d at 114; see also James R. Spady, Free Speech, Initiative and Property Rights in Conflict—Four Alternatives to the State Action Requirement in Washington, 58 Wash. L. Rev. 587, 614–15 (1983).

135See N.J. Coal., 650 A.2d at 777 (“[W]here private ownership of property that is the functional counterpart of the downtown business district has effectively monopolized significant opportunities for free speech, the owners cannot eradicate those opportunities by prohibiting it.”); Alderwood Assocs., 635 P.2d at 114 (“Because the nature of the public forum has changed, it is necessary to reassess the underpinnings for the existing free speech principles.”).

136See N.J. Coal., 650 A.2d at 777.

137 650 A.2d 757 (N.J. 1994).

138Id. at 773.

139Id.

140 Similar to the Supreme Court of New Jersey, the Colorado Supreme Court also found the commercial purposes of the shopping centers to not override its functioning as an equivalent of a public forum. See Bock, 819 P.2d at 62 (“[T]he historical connection between the marketplace of ideas and the market for goods and services is not severed because goods and services today are bought and sold within the confines of a modern mall …. The range of activities permitted in the common areas of the Mall also indicates the extent to which the Mall effectively functions as the latter-day public forum.”); see also Lewis v. Colo. Rockies Baseball Club, Ltd., 941 P.2d 266, 271 n.8 (Colo. 1997) (en banc) (“We found the concept of ‘public forum’ to be more expansive under the Colorado Constitution and held that the private shopping mall at issue was a public forum for free speech purposes.”). When California and Washington first addressed shopping centers, they did not clearly engage in comparing the commercial to non-commercial purposes of the centers but did rely on the functional equivalent argument. See Robins, 592 P.2d at 345–47; Alderwood Assocs., 635 P.2d at 116. And, though Washington has moved away from Alderwood’s analysis of the state free speech clause, it has relied on functional equivalent reasoning to reach similar results under its initiative clause. See Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 529–33 (Wash. 1999) (en banc).

141 515 A.2d 1331 (Pa. 1986).

142See Commonwealth v. Tate, 432 A.2d 1382 (Pa. 1981).

143 W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331, 1336 (Pa. 1986).

144Id. at 1337–38.

145Id. at 1336.

146Id. at 1337.

147See id. at 1333.

148Id. at 1337–39. Washington, relying on a functional equivalent of a public forum analysis under its initiative clause, has also used commercial purpose reasoning to distinguish a large grocery store from a shopping center and to hold the grocery store was not a functional equivalent. See Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 530 (Wash. 1999) (en banc).

149 Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979).

150 Bock v. Westminster Mall Co., 819 P.2d 55, 62–63 (Colo. 1991).

151 N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 781 (N.J. 1994).

152Id.

153 The Supreme Court of Pennsylvania concisely captured the relevant balancing inquiry in holding: [I]n certain circumstances, the state may reasonably restrict the right to possess and use property in the interests of freedom of speech, assembly, and petition …. To determine whether the appropriate circumstances must exist here, we must balance the college’s right to possess and protect its property against appellants’ rights of expression in light of the compatibility of that expression with the “activity of (the) particular place at (the) particular time.” Commonwealth v. Tate, 432 A.2d 1382, 1390 (Pa. 1981) (citations omitted).

154 423 A.2d 615 (N.J. 1980).

155Id. at 626–27.

156Id. at 628 (“They are also available against unreasonably restrictive or oppressive conduct on the part of private entities that have otherwise assumed a constitutional obligation not to abridge the individual exercise of such freedoms [of speech and assembly] because of the public use of their property.”); see also N.J. Coal., 650 A.2d at 779 (“In New Jersey, we have an affirmative right of free speech, and neither government nor private entities can unreasonably restrict it.”).

157Schmid, 423 A.2d at 630.

158Id. at 617–18.

159Id. at 630.

160Id. at 631.

161Id.

162Id. at 632–33.

163 N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757 (N.J. 1994).

164Id. at 772 (“The test to determine the existence of the constitutional obligation is multi-faceted; the outcome depends on a consideration of all three [Schmid] factors … and ultimately on a balancing ….”).

165Id. at 772–74.

166Id. at 774 (“The almost limitless public use of defendant’s property … gives rise to an implied invitation of constitutional dimensions … that includes leafletting on controversial issues.”).

167Id. at 775.

168Id. at 776. The Supreme Court of New Jersey adopted a sliding scale approach, saying “as private property becomes, on a sliding scale, committed either more or less to public use and enjoyment, there is actuated, in effect, a counterbalancing between expressional and property rights.” Id. at 775 (quoting State v. Schmid, 423 A.2d 615 (N.J. 1980)). The court also quoted the famous passage from Marsh v. Alabama that “[t]he more an owner, for his advantage, opens up his property for use by the public in general, the more do his rights become circumscribed by the statutory and constitutional rights of those who use it.” Id. at 775–76 (quoting Marsh v. Alabama, 326 U.S. 501, 506 (1946)).

169Id. at 776.

170Id.

171Wash. Const. art. II, § 1(a) (detailing the initiative petition process by which the people can “reserve to themselves the power to propose bills, laws, and to enact or reject the same…”). Although this Comment does not ultimately pull directly from the case law around Washington’s initiative clause, that clause is noted here for two reasons. First, to provide a complete and accurate representation of the development of Washington’s case law, specifically its shift away from using the state’s free speech clause. Second, because it uses the same reasoning as the free speech cases, providing another data point for comparison to this Comment’s ultimate approach.

172 635 P.2d 108 (Wash. 1981) (plurality opinion).

173Id. at 114–16. The Washington free speech clause reads: “Every person may freely speak, write and publish on all subjects, being responsible for the abuse of that right.” Id. at 114 (quoting Wash. Const. art. I, § 5).

174Id. at 116 (“[B]eing sensitive to the competing speech and property rights, we conclude [the speech and petition clauses] are applicable when, after balancing all the interests, the balance favors the speech and initiative activity.”).

175Id. at 116–17.

176Id. at 117.

177Id.

178See Southcenter Joint Venture v. Nat’l Democratic Policy Comm., 780 P.2d 1282, 1289–90 (Wash. 1989) (en banc) (“A 4-member plurality of this court, i.e., less than a majority of the court, maintained that there was no ‘state action’ requirement under the free speech and initiative provisions of the state constitution …. Thus, in Alderwood, a 5-member majority of this court rejected the argument … that the free speech provision of our state constitution does not require ‘state action.’”).

179Id. at 1288–89.

180 Waremart, Inc. v. Progressive Campaigns, Inc., 989 P.2d 524, 529–30 (Wash. 1999) (en banc) (“[W]e concluded in Alderwood that [the initiative clause] specifically guarantees the right to gather signatures for initiatives at large shopping malls.”). The court also expressly refused to overturn Alderwood. Id. at 530 (“Alderwood … effectuates and complies with the express mandate of this court to preserve the people’s right to engage in the initiative process. To overturn Alderwood would substantially diminish the ability of Washington’s citizens to participate in the initiative process in a meaningful way.”).

181Id.

182Id. Ultimately, the court concluded that the property in question was not affected by the initiative clause because it was not large enough; thus, its interests won the balancing test. Id. at 533.

183 This factor is closely related with the functional equivalent of a public forum concept, as the latter helps define the former.

184 This represents a combination of a Schmid and Alderwood factor. Alderwood terms it “the nature of the speech activity” and cites Schmid, but otherwise gives the factor scant discussion. See Alderwood Assocs. v. Wash. Envtl. Council, 635 P.2d 108, 116 (Wash. 1981) (plurality opinion). Schmid terms it “the purpose of the expressional activity undertaken upon such property in relation to both the private and public use of the property.” State v. Schmid, 423 A.2d 615, 630 (N.J. 1980). A later New Jersey case further articulates this factor, stating that it “examines the compatibility of the free speech sought to be exercised with the uses of the property.” N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 775 (N.J. 1994).

185See Schmid, 423 A.2d at 630; Waremart, Inc., 989 P.2d at 530; Alderwood Assocs., 635 P.2d at 116–17.

186See Robins v. Pruneyard Shopping Ctr., 592 P.2d 341, 347 (Cal. 1979); Bock v. Westminster Mall Co., 819 P.2d 55, 63 (Colo. 1991); N.J. Coal., 650 A.2d at 783; Schmid, 423 A.2d at 630; Commonwealth v. Tate, 432 A.2d 1382, 1390 (Pa. 1981) (quoting State v. Schmid, 423 A.2d 615, 630 (N.J. 1980)); Alderwood Assocs., 635 P.2d at 116 (citing Cox v. Louisiana, 379 U.S. 536, 544-55 (1965)).

187See Pruneyard Shopping Ctr. v. Robins, 447 U.S. 74, 83–84 (1980).

188See, e.g., id. at 83 (reasoning that there was no takings violation because “[the shopping center] may restrict expressive activity by adopting time, place, and manner regulations that will minimize any interference with its commercial functions.” (emphasis added)); Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 751 (Cal. 2007) (“[A] shopping center may prohibit conduct ‘calculated to disrupt normal business operations’ or that would result in ‘obstruction of or undue interference with normal business operations.’” (quoting Diamond v. Bland, 477 P.2d 733, 741–42 (Cal. 1970))); Green Party of N.J. v. Hartz Mountain Indus., Inc., 752 A.2d 315, 327–28 (N.J. 2000) (“… designed to ensure that time, place, and manner regulation should minimize any interference with the mall’s commercial function without denying the counterbalancing interest of leafletters in expressive speech.”).

189See Fashion Valley Mall, 172 P.3d at 751, 754.

190Id.

191Id. at 751.

192Id. at 751–52.

193Id. at 754.

194Id.

195 Green Party of New Jersey v. Hartz Mountain Industries, Inc., 752 A.2d 315, 325–26 (N.J. 2000).

196Id. at 326.

197Id. at 325–26.

198Id. at 326–27.

199Id. at 326 (“Under the rule, when business judgments are made in good faith based on reasonable business knowledge, the decision makers are immune from liability from actions brought by others who have an interest in the business entity.”).

200Id. at 327.

201Id.

202Id. at 327–28.

203Id. at 317.

204Id. at 328–32.

205 A factual difference between the two cases is that the restriction the California court addressed was content-based, whereas the New Jersey court addressed a content-neutral restriction. The New Jersey court, however, made no mention of whether the restriction was content-based or content-neutral, and this consideration seems irrelevant to its reasoning.

206 For example, in Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 754 (Cal. 2007), the court says that the shopping center’s goal of maximizing profits is not compelling compared to a labor union’s right to free expression. It is difficult to imagine when or how a business property owner might have an interest other than maximizing profit, and harder still to imagine when this interest would prevail over political speech. As such, it seems too rigid a standard.

207See, e.g., Fashion Valley Mall, 172 P.3d at 751; Golden Gateway Ctr. v. Golden Gateway Tenants Ass’n, 29 P.3d 797, 821–22 (Cal. 2001) (Werdegar, J., dissenting); Bock v. Westminster Mall Co., 819 P.2d 55, 62–63 (Colo. 1991).

208 W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331, 1336 (Pa. 1986) (“Nor does it yet require [shopping centers] to provide a political forum for persons or groups with views on public issues, so long as the owner does not grant unfair advantages to particular interests or groups by making his premises arbitrarily available to those he favors while excluding others.”).

209Green Party of N.J., 752 A.2d at 329 (regulations must focus on objective criteria and avoid “general and subjective standard[s] vulnerable to arbitrary or content related determination.”); State v. Schmid, 423 A.2d 615, 632 (N.J. 1980) (“[P]rivate colleges … must be accorded a generous measure of autonomy and self-governance … In this case, however, the University regulations that were applied to Schmid … contained no standards ….”); Commonwealth v. Tate, 432 A.2d 1382, 1387 (Pa. 1981) (“The record does not reveal what standards, if any, were applicable to the granting or denial of such permission. It appears, however … the college believed itself entitled to exercise its discretion arbitrarily.”).

210Schmid, 423 A.2d at 633 (“These current amended regulations exemplify the approaches open to private educational entities seeking to protect their institutional integrity while at the same time recognizing individual rights of speech ….”). The revised restrictions were content-neutral and provided a process to get permission to distribute political literature on campus along with rules for times when and locations where this was allowed. Id. at 617 n.2.

211 The term “regulation” may seem equally appropriate here as a matter of dictionary definition. However, because that smacks of direct state action, this Comment is of the mind that “moderation” is the better word choice.

212 This does not mean, however, that the approach below is limited to Facebook, and in each instance that “Facebook” is used, it can be substituted for any given social media platform. For example, Twitter almost certainly can be analyzed identically to Facebook. On the other hand, LinkedIn may present a different analysis and outcome, given its more niche nature and limited invitation to the public.

213See supra Part II.B and accompanying notes.

214See supra Part II.C and accompanying notes.

215See supra notes 2–9 and accompanying text.

216See Nikaela Jacko Redd & Lutisha S. Vickerie, The Rise and Fall of Brick and Mortar Retail: The Impact of Emerging Technologies and Executive Choices on Business Failure, 17 J. Int’l Bus. & L. 127, 127–28 (2017); Wolf Richter, Here’s Which Brick-and-Mortar Retailers Are Getting Hit the Hardest, Bus. Insider (May 19, 2018, 11:02 AM), https://www.businessinsider.com/brick-and-mortar-retailers-getting-hit-the-hardest-2018-5.

217 Josh Sanburn, Why the Death of Malls is About More than Shopping, Time (July 20, 2017), https://time.com/4865957/death-and-life-shopping-mall/. Some malls have even become so empty and run down that they are now popular spots for filming horror movies. See J.D. Capelouto, NEW FINDINGS: North Dekalb Mall is so Empty, it’s Becoming a Hot Spot for Filming, AJC (May 3, 2019), https://www.ajc.com/news/local/north-dekalb-mall-empty-become-hot-spot-for-filming/kTWe3QStwwyAoPaALmY2QO/. See also Liz Stinson, Starcourt Mall From ‘Stranger Things’ is Actually a Real Dead Mall, Curbed (July 12, 2019, 1:13 PM), https://www.curbed.com/2019/7/12/20691222/stranger-things-starcourt-mall-real-mall-georgia.

218 Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 760 (Cal. 2007) (Chin, J., dissenting). Moreover, this dissent was written in 2007. It’s safe to say the argument has only grown stronger empirically. See also Selznick & LaMacchia, supra note 9, at 267–74.

219 So much so that extensive investigation has been dedicated to the idea that manipulation of political speech on social media effectively swayed a presidential election. See, e.g., Jane Mayer, How Russia Helped Swing the Election for Trump, New Yorker (Sept. 24, 2018), https://www.newyorker.com/magazine/2018/10/01/how-russia-helped-to-swing-the-election-for-trump.

220See supra notes 3–9 and accompanying text.

221See Jane Elizabeth, After a Decade, It’s Time to Reinvent Social Media in Newsrooms, Am. Press Inst. (Nov. 14, 2017), https://www.americanpressinstitute.org/publications/reports/strategy-studies/reinventing-social-media/single-page/. At the same time, recognizing the importance of journalists’ social media presence, some news outlets have sought to limit the kind of content their employees post. See Mathew Ingram, Social Media Crackdowns at the Times and Journal Will Backfire, Colum. Journalism Rev. (Oct. 20, 2017), https://www.cjr.org/criticism/social-media-twitter-times-journal.php.

222See, e.g., Bridget Bowman, The Future of Ads Is Digital — But Not Quite the Present, Roll Call (Dec. 11, 2018, 5:03 AM), https://www.rollcall.com/news/politics/whats-next-digital-ads-spending-increases-still-slow-change; Lindsey Erdody, Political Campaigns Boost Investment in Social Media Ads, Indianapolis Bus. J. (Sept. 21, 2018), https://www.ibj.com/articles/70545-political-campaigns-boost-investment-in-social-media-ads; Nihal Krishan, Dems Ramp Up Digital Advertising Effort for Midterms. Trump’s Doing the Same for 2020, OpenSecrets.org (Sept. 11, 2018), https://www.opensecrets.org/news/2018/09/dems-ramp-up-digital-ads-for-midterms-trumps-doing-2020/.

223See supra note 10; Election 2016: Campaigns as a Direct Source of News, Pew Res. Ctr. (July 18, 2016), http://www.journalism.org/wp-content/uploads/sites/8/2016/07/PJ_2016.07.18_election-2016_FINAL.pdf.

224See Andrew Buncombe, Donald Trump One Year on: How the Twitter President Changed Social Media and the Country’s Top Office, Indep. (Jan. 17, 2018, 9:43 PM), https://www.independent.co.uk/news/world/americas/us-politics/the-twitter-president-how-potus-changed-social-media-and-the-presidency-a8164161.html.

225See supra note 5 and accompanying text.

226See supra Part II.B and accompanying notes.

227 Facebook’s own words support this assertion, and a discussion of Facebook’s purported purpose and its mission statement is below. See infra note 236 and accompanying text.

228 515 A.2d 1331 (Pa. 1986).

229See supra notes 138–146 and accompanying text.

230N.J. Coalition Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 767 (N.J. 1994)

231See supra notes 133–134 and accompanying text.

232See, e.g., FAQs, Facebook: Inv. Rel., https://investor.fb.com/resources/default.aspx (last visited Feb. 17, 2019).

233 For this reason, social media platforms tend naturally toward monopolistic features. See Luigi Zingales & Guy Rolnik, A Way to Own Your Social-Media Data, N.Y. Times (June 30, 2017), https://www.nytimes.com/2017/06/30/opinion/social-data-google-facebook-europe.html. But see Sangeet Paul Choudary, Reverse Network Effects: Why Today’s Social Networks Can Fail as They Grow Larger, Wired, https://www.wired.com/insights/2014/03/reverse-network-effects-todays-social-networks-can-fail-grow-larger/ (last visited Jan. 17, 2019). In short, the point for users is to be on the platform with the most other users; scale, therefore, is crucial. Although the dominant platform might not remain so perpetually, it seems most likely that change will be through users transitioning to another, newly dominant platform, and not splintering into smaller ones (like for example the correlated decline of MySpace and rise of Facebook). The importance of this is that, even if a platform were to ban a certain viewpoint and proponents of that viewpoint created their own platform, it would be difficult for the new platform to match the scale requisite to being influential.

234See supra Part II.C.

235 As the New Jersey and Washington supreme courts found in the relevant cases. See supra notes 154 and 169.

236 Facebook’s mission statement says its “mission is to give people the power to build community and bring the world closer together” through letting people “share and express what matters to them.” FAQs, Facebook: Inv. Rel., https://investor.fb.com/resources/default.aspx (last visited Feb. 17, 2019).

237 These activities may occur on Facebook, which creates a reverse situation from shopping centers as described in cases like New Jersey Coalition Against War in the Middle East, 650 A.2d at 782 (N.J. 1994). See supra notes 136–137 and accompanying text.

238 This is with the sole exception of people younger than 13 years old. See Mallory Locklear, Facebook Will Begin Actively Enforcing Its Age Policy, Engadget (July 19, 2018), https://www.engadget.com/2018/07/19/facebook—actively-enforcing-age-policy; FAQs, Facebook: Inv. Rel., https://investor.fb.com/resources/default.aspx

239 423 A.2d 615 (N.J. 1980).

240Id. at 230.

241 An interesting thought arises when considering Facebook pre-2006, when it was only available to people with college email addresses. See Saul Hansell, Site Previously for Students Will Be Opened to Others, N.Y. Times (Sept. 12, 2006), https://www.nytimes.com/2006/09/12/technology/12online.html. In that case, its invitation to the public would be much more limited than it is currently, which could prevent a state free speech clause from reaching it under this Comment’s approach.

242Schmid, 423 A.2d at 631.

243N.J. Coal., 650 A.2d at 775.

244Id.

245See infra Part III.B and accompanying notes.

246See supra notes 3–9 and accompanying text.

247See supra note 21.

248N.J. Coal., 650 A.2d at 776.

249See supra Part II.D.

250See supra Part II.D and accompanying notes.

251 752 A.2d 315, 327 (N.J. 2000).

252See supra Part II.A.

253 This is also evident in the cases. See supra Part II.C.

254Community Standards, Facebook, https://www.facebook.com/communitystandards/cruel_insensitive (“[W]e have higher expectations for content that we call cruel and insensitive, which we define as content that targets victims of serious physical or emotional harm.”) (last visited Nov. 13, 2018).

255Cf. N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 781–82 (N.J. 1994).

256Id. at 782.

257See Green Party of N.J. v. Hartz Mountain Indus., Inc., 752 A.2d 315, 332 (N.J. 2000) (striking a shopping center’s restriction, in part because more than the one day a year the center allowed is needed for effective leafletting).

258Id. at 327.

259See supra notes 207–208 and accompanying text.

260See supra notes 205–206 and accompanying text; cf. Hill v. Colorado, 530 U.S. 703, 769–70 (2000) (Kennedy, J., dissenting) (“The Court time and again has held content-based or viewpoint-based regulations to be presumptively invalid.”).

261See supra notes 205–208 and accompanying text.

262 Again, this is other than the possibility of Facebook doing things like limiting some types of posts to certain times of day.

263See Rosenberger v. Rector & Visitors of Univ. of Va., 515 U.S. 819, 828–29 (1995) (noting that viewpoint discrimination is “[w]hen the government targets not subject matter, but particular views taken by speakers on a subject”). See Reed v. Town of Gilbert, 135 S. Ct. 2218, 2229–30 (2015) and Pahls v. Thomas, 718 F.3d 1210, 1229 (10th Cir. 2013) for a further discussion of the distinction between content-based and viewpoint-based restrictions.

264See supra notes 205–208 and accompanying text.

265See Bock v. Westminster Mall Co., 819 P.2d 55, 56–57 (Colo. 1991) (noting that the shopping center previously allowed voter registration drives, salute to armed forces and displays by armed forces agencies, agency representatives, salute to presidents and display of presidential portraits, Boy Scouts and Girl Scouts, Salvation Army, controversial films); N.J. Coal. Against War in the Middle E. v. J.M.B. Realty Corp., 650 A.2d 757, 765, 776 (N.J. 1994) (noting that the shopping center previously allowed voter registration drive, sixty-seven government agencies distributing literature, senator campaigning in person, Toys for Tots, military recruitment, Berlin Wall Exhibit, senate campaign voter registration drive, &c.); see also W. Pa. Socialist Workers 1982 Campaign v. Conn. Gen. Life Ins. Co., 515 A.2d 1331, 1336 (Pa. 1986) (“Nor does [the law] yet require [the shopping center] to provide a political forum for persons or groups with views on public issues, so long as the owner does not grant unfair advantage to particular interests or groups by making his premises arbitrarily available to those he favors while excluding all others.” (emphasis added)).

266W. Pa. Socialist Workers, 515 A.2d at 1336.

267Supra Part II.D.

268See e.g., N.J. Coal., 650 A.2d at 782 (“[O]ur ruling … does not include bullhorns, megaphones, or even a soapbox; it does not include placards, pickets, parades, and demonstrations; it does not include anything other than normal speech and then only such as is necessary to the effectiveness of the leafletting.”).

269 This application is in the abstract and to political speech generally. But it is difficult to imagine a scenario where some specific political statement alters the equation, so the outcome would likely be the same when applied to any given example of a political post.

270 As to banning an account, this completely limits the user’s free speech right, at least with respect to the exercise of that right on the platform. And, when deleting a single post, it does not limit the user’s right to make other statements, but as to that statement and its viewpoint, the limitation is total. Of course, this does not mean the user’s free speech right is limited totally beyond the platform. But, given the perpetually increasing importance and centrality of social media platforms, the limitation can have far-reaching consequences.

271See supra note 21.

272 Though Facebook’s owners may not like having certain views expressed on the platform or may see them as in tension with Facebook’s mission statement, the operative inquiry in the case law detailed above focused only on the economic impact on the private actor. Of course, it remains possible for any given platform to rebrand as specifically only accepting of one type of ideology, which would in turn change the other factors like the scope of the invitation to the public.

273Community Standards, Facebook, https://www.facebook.com/communitystandards/hate_speech (last visited Feb. 17, 2019).

274Id.; see also Klonick, supra note 13, at 1630–61 (discussing in detail Facebook’s processes for content moderation); Richard Allan, Hard Questions: Who Should Decide What Is Hate Speech in an Online Global Community, Facebook: Newsroom (June 27, 2017), https://newsroom.fb.com/news/2017/06/hard-questions-hate-speech/.

275 For a list of common arguments against immigration that may fit this example (and, less relatedly, a rebuttal to each), see Alex Nowrasteh, The 14 Most Common Arguments Against Immigration and Why They’re Wrong, Cato Inst.: Cato Liberty (May 2, 2018, 11:10 AM), https://www.cato.org/blog/14-most-common-arguments-against-immigration-why-theyre-wrong.

276 These arguments were made by pundits and public intellectuals. See, e.g., Todd Bensman, Middle Eastern Migrants Part of Caravan?, Ctr. for Immigr. Stud. (Oct. 22, 2018), https://cis.org/Bensman/Middle-Eastern-Migrants-Part-Caravan; Michelle Malkin, It’s Not Unreasonable to be Worried About Disease and the Caravan, Nat’l Rev. (Oct. 31, 2018, 11:09 AM), https://www.nationalreview.com/2018/10/its-not-unreasonable-to-be-worried-about-disease-and-the-caravan/.

277See Julie Hirschfeld Davis, Trump Calls Some Unauthorized Immigrants ‘Animals’ in Rant, N.Y. Times (May 16, 2018), https://www.nytimes.com/2018/05/16/us/politics/trump-undocumented-immigrants-animals.html.

278See Julie Hirschfeld Davis et al., Trump Alarms Lawmakers with Disparaging Words for Haiti and Africa, N.Y. Times (Jan. 11, 2018), https://www.nytimes.com/2018/01/11/us/politics/trump-shithole-countries.html. Admittedly, using examples spoken by a president or other politician raises a complication because such speech is necessarily political and presumably entitled to great protection, even if squarely within Facebook’s hate speech policy. In such situations, the balancing test will almost necessarily favor the speech because it directly involves politicians and matters of important public information. Notably, Twitter co-founder and CEO Jack Dorsey has used this reasoning in explaining why President Trump’s tweets should almost never be taken down. See #148 – Jack Dorsey, Making Sense (Feb. 5, 2019) (downloaded using Spotify); #1236–Jack Dorsey, Joe Rogan Experience (Feb. 1, 2019) (downloaded using Spotify).

279Community Standards, Facebook, https://www.facebook.com/communitystandards/hate_speech (last visited Feb. 17, 2019).

280 135 S. Ct. 2584 (2015) (holding states cannot prohibit gay marriage).

281See, e.g., id. at 2611 (Roberts, C.J., dissenting).

282See, e.g., Darrell Bock, The Bible and Same-Sex Marriage: 6 Common But Mistaken Claims, Gospel Coalition (July 27, 2015), https://www.thegospelcoalition.org/article/bible-same-sex-marriage-6-wrong-trajectories/.

283Cf. supra notes 205–208 and accompanying text.

284 It is important to distinguish between intentionally hateful use of such slurs and arguments that such slurs should be allowed to be used. If a post argues that slurs or a given slur should be allowed to be used, then prohibiting such a post becomes viewpoint-based against that viewpoint.

285See generally Jonathan Haidt, The Righteous Mind (2012).

286 For examples, see Facebook’s listed “protected characteristics.” Community Standards, Facebook, https://www.facebook.com/communitystandards/hate_speech (last visited Feb. 17, 2019) (listing the following protected characteristics: race, ethnicity, national origin, religious affiliation, sexual orientation, caste, sex, gender, gender identity, serious disease or disability, and “some protections for immigration status”).

287 This should not reach so far as to affect things like calling arguments, or even their proponents, stupid. Although this may be inflammatory, it does not fall within the “protected characteristics” promulgated by Facebook. It also creates a heightened problem of line-drawing because disagreement necessarily requires calling another viewpoint wrong, which is inherently offensive. Even though name calling may not be “decent” in a traditional sense, it does not tip the balancing test toward restricting such speech.

288 An interesting question is raised by Facebook’s algorithm and decisions to promote or prevent certain content from appearing in feeds. For example, Facebook has recently made an effort to have users’ feeds feature more posts by friends, and less by pages who pay for placement, see Adam Mosseri, Bringing People Closer Together, Facebook: Newsroom (Jan. 11, 2018), https://newsroom.fb.com/news/2018/01/news-feed-fyi-bringing-people-closer-together/, which some have alleged is inherently biased against conservative viewpoints, see Jason Schwartz, Conservative Outlets Take on Facebook, Politico (Mar. 29, 2018, 5:00 AM), https://www.politico.com/story/2018/03/29/conservatives-facebook-liberal-bias-490920. And, there are certainly general concerns that Facebook can use its algorithm or manual manipulation to suppress some viewpoints. See Laura Wagner, Is Facebook Suppressing Politically Conservative Content?, NPR (May 10, 2016, 6:06 PM), https://www.npr.org/sections/thetwo-way/2016/05/10/477525204/is-facebook-suppressing-politically-conservative-content. But see Mark Zuckerberg, Facebook (May 12, 2016), https://www.facebook.com/zuck/posts/10102830259184701 (reporting that internal investigations have found no evidence of suppression). Whether Facebook does, in fact, do this, it certainly possesses the ability to. But such actions need not necessarily conflict with free speech. For example, changes may be viewpoint-neutral decisions meant simply to make users’ feeds more attractive by including more posts from family members. But, one can also imagine a situation in which Facebook intentionally and perniciously suppresses a viewpoint, and thus accomplishes substantially the same effect as deleting it. Where exactly to draw the line at which this violates free speech protections is beyond this Comment’s scope; that determination, however, can still be resolved by applying the balancing test this Comment advances.

289 447 U.S. 74 (1980). Although First Amendment jurisprudence has evolved in the decades since Pruneyard, it does not seem that the United States Supreme Court has moved away from Pruneyard and its underlying federalist principles. At least as recently as 2008, the Court denied a petition for certiorari in a California case where the Supreme Court of California applied strict scrutiny to a shopping center’s speech restriction. See Fashion Valley Mall, LLC v. NLRB, 172 P.3d 742, 751, 754 (Cal. 2007), cert. denied, 555 U.S. 819 (2008); see also Ralphs Grocery Co. v. United Food & Commercial Workers Union Local 8, 290 P.3d 1116 (Cal. 2012), cert. denied, 133 S. Ct. 2799 (2013) (denying a petition for certiorari in case using state statute to similar effect); Eugene Volokh, The Law of Compelled Speech, 97 Tex. L. Rev. 355, 358 (2018) (exploring current state of compelled speech law under the First Amendment).

290See supra notes 91–91 and accompanying text.

291See supra notes 141–146 and accompanying text.

292See supra note 92 and accompanying text.

293See supra note 94 and accompanying text.

294See supra notes 95–98 and accompanying text.

295See supra note 99 and accompanying text.

296 On the question of whether Facebook is a publisher like a newspaper, it has tried to have its cake and eat it too, arguing both that it is and is not. See, e.g., Sam Levin, Is Facebook a Publisher? In Public It Says No, But in Court It Says Yes, Guardian (July 3, 2018, 2:00 PM), https://www.theguardian.com/technology/2018/jul/02/facebook-mark-zuckerberg-platform-publisher-lawsuit. Regardless of the legal answer to that question, this Comment’s approach would survive. The balancing of Facebook’s free speech rights against a user’s free speech rights would still take on the analytical framework of Pruneyard, as it is the guiding precedent for this kind of scenario.

297 The rights under a given state’s constitution would only be applicable to that state’s citizens. See, e.g., People v. LaValle, 817 N.E.2d 341, 366 (N.Y. 2004) (“It bears reiterating here that ‘on innumerable occasions this Court has given the State Constitution an independent construction, affording the rights and liberties of the citizens of this State even more protection than may be secured under the United States Constitution.’” (quoting Sharrock v. Dell Buick-Cadillac, 379 N.E.2d 1169, 1173 (N.Y. 1978)) (emphasis added)); see also Baldwin v. Fish & Game Comm’n, 436 U.S. 371, 383 (“It has not been suggested, however, that state citizenship or residency may never be used by a State to distinguish among persons. Suffrage, for example, always has been understood to be tied to an individual’s identification with a particular State.” (citations omitted)); Blake v. McClung, 172 U.S. 239, 256–57 (1898) (“We must not be understood as saying that a citizen of one State is entitled to enjoy in another State every privilege that may be given in the latter to its own citizens.”).

298 For example, other obstacles may be Facebook’s “Terms of Service” agreement, which users agree to when setting up an account. This could affect choice of forum and choice of law. See Dolin v. Facebook, Inc., 289 F. Supp. 3d 1153, 1158–61 (D. Haw. 2018) (discussing whether choice of forum clause in Facebook’s terms of service is binding); Fteja v. Facebook, Inc., 841 F. Supp. 2d 829, 834–41 (S.D.N.Y. 2012) (discussing the same). Another is that a user would need a cause of action, the specifics of which may vary by state. Compare Corum v. Univ. of N.C., 413 S.E.2d 276, 289–91 (N.C. 1992) (holding there is a direct action against the state for free speech violations under the state constitution), with Roman v. Trs. of Tufts Coll., 964 N.E.2d 331, 336–37 (Mass. 2012) (discussing statute providing cause of action for rights secured by the state or federal constitution).

299 It is also crucial that this Comment not be the final word, and that the approach it proposes be explored further. Plaintiffs seeking to protect their free speech against arbitrary censorship by social media platforms stand little chance under the First Amendment, absent an unexpected and momentous reversal of entrenched precedent by the United States Supreme Court. Yet, litigants continue to bring doomed-to-fail claims based on the First Amendment. See, e.g., Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 WL 1471939, at *5 (N.D. Cal. Mar. 26, 2018) (holding YouTube is not a state actor for First Amendment purposes); Shulman v. Facebook.com, No. 17-764 (JMV), 2017 WL 5129885, at *4 (D.N.J. Nov. 6, 2017) (holding Facebook is not a state actor for First Amendment purposes and collecting cases holding the same). Instead, getting this issue before state courts may be the best—and possibly only—way for speakers and scholars to protect free speech against social media platforms’ growing power to control it. In that spirit, this Comment suggests what to say once before a state court, and offers state courts a framework to begin from, leaving the path to getting there to future trailblazers.

300See generally Elma Delic, Note, Cloudy Jurisdiction: Foggy Skies in Traditional Jurisdiction Create Unclear Legal Standards for Cloud Computing and Technology, 50 Suffolk U. L. Rev. 471, 472–78 (2017) (discussing development of personal jurisdiction law generally and its application to the internet); Scott Isaacson, Note, Finding Something More in Targeted Cyberspace Activities, 68 Rutgers U. L. Rev. 905, 914–26 (2016) (discussing personal jurisdiction and applications to cyberspace).

301See Daimler AG v. Bauman, 571 U.S. 117, 137 (2014) (holding that the paradigmatic examples of general jurisdiction over a corporation are the state in which it is incorporated and in which its principal place of business is). Facebook, in litigation, has argued that there is general jurisdiction over it in California and Delaware. See Georgalis v. Facebook, Inc., 324 F. Supp. 3d 955, 959–61 (N.D. Ohio 2018) (rejecting that there is general jurisdiction over Facebook in Ohio because it has many users there); Ralls v. Facebook, 221 F. Supp. 3d 1237, 1243 (W.D. Wash. 2016).

302See supra Part II.

303California Population 2019, World Population Rev., http://worldpopulationreview.com/states/california-population/ (last visited Jan. 20, 2018).

304 Additionally, Facebook’s “Terms of Service” agreement has a choice of law provision specifying that California law governs any dispute between Facebook and the user. See Terms of Service, Facebook, https://www.facebook.com/terms.php (last visited Feb. 17, 2019). Although Facebook could change this clause to apply any other state’s law, it may nevertheless be that California law continues to apply as it would be the only possible location of the dispute since both the user and the platform’s headquarters are there.

305 See Adam R. Kleven, Note, Minimum Virtual Contacts: A Framework for Specific Jurisdiction in Cyberspace, 116 Mich. L. Rev. 785, 789–96 (2018) for general background on specific jurisdiction case law applying it to the internet.

306 The case law on personal jurisdiction over social media platforms is apparently extremely limited. See Gullen v. Facebook.com, Inc., No. 15 C 7681, 2016 WL 245910, at *2 (N.D. Ill. Jan. 21, 2016) (“The Court has not found a case analyzing specific jurisdiction over a social media website.”).

307See Georgalis v. Facebook, Inc., 324 F. Supp. 3d 955, 961 (N.D. Ohio 2018); Ralls v. Facebook, 221 F. Supp. 3d 1237, 1243–44 (W.D. Wash. 2016); Gullen, 2016 WL 245910, at *2–3 (considering a challenge to Facebook’s facial recognition technology and rejecting that personal jurisdiction exists merely because Facebook operates an interactive website accessible in the forum state).

308 This could be the kind of “targeting” of a user in a forum state that the court in Gullen found lacking (which was fatal to personal jurisdiction there). See Gullen, 2016 WL 245910, at *2. Additionally, Facebook could not hide behind an algorithm applicable to all users, because many decisions to delete posts or accounts are ultimately targeted and specific to users. See Klonick, supra note 13, at 1635–49. It could also qualify under other possible tests for personal jurisdiction. See Isaacson, supra note 299, at 926–47 (discussing ways in which personal jurisdiction may apply to various new internet related technologies, including Facebook). Or, there could be personal jurisdiction under frameworks proposed by scholars. See Kleven, supra note 304, at 800–09 (proposing two-prong framework based on virtual contacts causing harm to plaintiff and reasonableness).

309See Delic, supra note 299, at 487–89 (criticizing the applicability of traditional personal jurisdiction concepts to cloud computing).

310Id. at 476 n.49 (collecting articles discussing the relationship between personal jurisdiction and growth of internet); Allison MacDonald, Comment, YouTubing Down the Stream of Commerce: Eliminating the Express Aiming Requirement for Personal Jurisdiction in User-Generated Internet Content Cases, 19 Alb. L.J. Sci. & Tech. 519, 534–43 (2009).

311 Kleven, supra note 305, at 796–800 (criticizing narrowness of current personal jurisdiction rules as applied to internet activity); see also Isaacson, supra note 300, at 911–12 n.33 (collecting articles on personal jurisdiction and the internet).

312 An issue related to jurisdictional questions is that fifty different states may adopt fifty different frameworks, creating a convoluted regulatory scheme for social media platforms to navigate. For example, Facebook would have to know which state a particular user is a citizen of, and then know what sort of limitation on speech that state allows. This is not, however, a defect in this Comment’s proposal, but a strength. Any variation among states would create an incentive for Facebook to simply adopt a uniform policy that complies with all. In other words, variation would pressure Facebook to err on the side of caution by providing greater protection of speech and being less willing and likely to delete political speech.

313 For general background on the law and its development, see Jaime M. Freilich, Note, Section 230’s Liability Shield in the Age of Online Terrorist Recruitment, 83 Brook. L. Rev. 675, 679–89 (2018); Klonick, supra note 13, at 1603–09; Lynn C. Percival, IV, Public Policy Favoritism in the Online World: Contract Voidability Meets the Communications Decency Act, 17 Tex. Wesleyan L. Rev. 165, 167–73 (2011); Emma M. Savino, Comment, Fake News: No One Is Liable, and That Is a Problem, 65 Buff. L. Rev. 1101, 1138–51 (2017); Olivier Sylvain, Intermediary Design Duties, 50 Conn. L. Rev. 203, 231–58 (2018). For a succinct summary of Section 230 as it may apply to Facebook, see Am. Freedom Def. Initiative v. Lynch, 217 F. Supp. 3d 100, 104–05 (D.D.C. 2016).

314See, e.g., Joshua N. Azriel, Social Networking as a Communications Weapon to Harm Victims: Facebook, MySpace, and Twitter Demonstrate a Need to Amend Section 230 of the Communications Decency Act, 26 J. Marshall J. Computer & Info. L. 415 (2009); Danielle Keats Citron & Benjamin Wittes, The Internet Will Not Break: Denying Bad Samaritans § 230 Immunity, 86 Fordham L. Rev. 401, 411–23 (2017); Freilich, supra note 312, at 689–91; Mary Graw Leary, The Indecency and Injustice of Section 230 of the Communications Decency Act, 41 Harv. J.L. & Pub. Pol’y 553 (2018); Michelle Roter, Note, With Great Power Comes Great Responsibility: Imposing a “Duty to Take Down” Terrorist Incitement on Social Media, 45 Hofstra L. Rev. 1379, 1394–98 (2017); Savino, supra note 313, at 1157–60; Sylvain, supra note 313, at 269–71.

315 “Interactive computer service” is defined as “any information service, system, or access software provider that provides or enables computer access by multiple users to a computer server, including specifically a service or system that provides access to the Internet….” 47 U.S.C. § 230(f)(2) (2018).

316See id. § 230(c)(1). “Information content provider” is defined as “any person or entity that is responsible, in whole or in part, for the creation or development of information provided through the Internet or any other interactive computer service.” Id. § 230(f)(3).

317Id. § 230(c)(2)(A).

318See id. § 230(e)(3).

319See Freilich, supra note 312, at 682–84. A notable application of Section 230 is the proliferation of so-called “fake news” on social media sites, and the liability shield has been criticized in that context. See Dallas Flick, Comment, Combatting Fake News: Alternatives to Limiting Social Media Misinformation and Rehabilitating Quality Journalism, 20 SMU Sci. & Tech. L. Rev. 375 (2017); Savino, supra note 312, at 1157–60.

320See supra note 311. Section 230’s findings are also illuminating. See 47 U.S.C. § 230(a)(3), (5) (2018) (“The Internet and other interactive computer services offer a forum for a true diversity of political discourse, unique opportunities for cultural development, and myriad avenues for intellectual activity …. Increasingly Americans are relying on interactive media for a variety of political, educational, cultural, and entertainment services.”).

321See Savino, supra note 313, at 1144 n.276. This means that most litigation occurs under 47 U.S.C. § 230(c)(1) rather than 47 U.S.C. § 230(c)(2)(A). See also supra note 319 (discussing application to fake news).

322 For discussion of the existing case law addressing the same statute, see Nicholas Conlon, Freedom to Filter Versus User Control: Limiting the Scope of § 230(c)(2) Immunity, 2014 U. Ill. J.L. Tech. & Pol’y 105, 116–22 (2014).

323See Prager Univ. v. Google LLC, No. 17-CV-06064-LHK, 2018 WL 1471939, at *5 (N.D. Cal. Mar. 26, 2018) (holding YouTube is not a state actor for First Amendment purposes); Shulman v. Facebook.com, No. 17-764 (JMV), 2017 WL 5129885, at *4 (D.N.J. Nov. 6, 2017).

324 108 F. Supp. 3d 876 (N.D. Cal. 2015). There, YouTube removed a music video because it alleged the view count was being artificially inflated, something it deemed “otherwise objectionable.” Id. at 880.

325 47 U.S.C. § 230(c)(2)(A) (2018).

326Song fi Inc., 108 F. Supp. 3d at 882–84.

327Id. at 883.

328 Moreover, if such speech did fall into one of those categories, this Comment’s proposed approach would probably not provide protection of it. See supra Part II.C and accompanying notes.

329 See Conlon, supra note 321, at 118 for a brief discussion of court interpretations of “good faith” within Section 230(c)(2)(A).

330See Klayman v. Zuckerberg, 753 F.3d 1354, 1357 (D.C. Cir. 2014); Cohen v. Facebook, Inc., 252 F. Supp. 3d 140, 156 n.10 (E.D.N.Y. 2017).

331See Klayman, 753 F.3d at 1358; Am. Freedom Def. Initiative v. Lynch, 217 F. Supp. 3d 100, 104–05 (D.D.C. 2016).

332See Song fi Inc., 108 F. Supp. 3d at 883–84.

*J.D. Candidate, Class of 2020.