Social media are aflame with accusations and counteraccusations: Liberals have a stranglehold on American news! Conservatives are driving a fascist insurrection! Silicon Valley elites silence the voice of people! Red-staters are book-burning simpletons! For good and for ill, ideas on social media spread like wildfire—especially emotionally charged, half-true ones. The power of online speech puts speech itself at the center of public debate, with rhetoric at a fever pitch.

Were the stakes not so very high, we might all stop for a moment and smile: Flag-burning, free-speech-loving Democrats of yesteryear condemn today’s charged online rhetoric as a danger to American democracy, while law-and-order Republicans charge into the breach to defend free-speech ramparts built by the sweat and blood of 1970s radicals. But this is no time to revel in the irony. The great online speech debate is upon us—flush with substance, compelling arguments all around, and real consequences for American democratic government.

The Supreme Question

Soon the question will land in the Supreme Court, which has before it petitions to review Florida and Texas laws that would require social-media platforms to host user speech without regard to its viewpoint. The laws are being challenged by NetChoice, a tech-industry trade association whose members include Twitter, Meta, Pinterest, and Google. The challengers to the Texas and Florida laws argue that even were their content-moderation decisions politically motivated, decisions about what content to promote and what content to censor are protected from interference by the platforms’ own First Amendment interests in free expression.

That argument is the crux of the debate. Americans demand freedom to express their views online, even misinformed or disfavored ones. Yet platforms have their own contrary interests at stake. They are trying to build online environments that people actually want to visit. Lunatics shouting, spewing nonsense, and running madly through the virtual hallways tend to drive away business. The problem is that although online platforms are private entities, with an obvious interest in curating user content, they are so large and so dominant that they have become chokepoints of online discourse.

Under the Court’s compelled-association cases, the viability of the Florida and Texas laws depends on whether one thinks Twitter is more like a marching band or like a public shopping mall—the circumstances of two leading First Amendment decisions of the Supreme Court. First, In Hurley v. Irish-American Gay, Lesbian & Bisexual Group of Boston, Justice Souter, writing for a unanimous Court, forbade Massachusetts from forcing private parade organizers to include a gay-, lesbian-, and bisexual- pride group in their parade. The Court acknowledged the validity of Massachusetts’ public accommodations law, but concluded that, when applied to expressive conduct like a private parade, “where every participating unit affects the message conveyed by the private organizers,” the law would require the parade organizers to “alter the expressive content of their parade” and therefore violate “the fundamental rule of protection under the First Amendment, that a speaker has the autonomy to choose the content of his own message.”

Another guidepost is PruneYard Shopping Center v. Robins, in which the Supreme Court considered a provision in California’s state constitution that creates a right for members of the public to engage in “free speech and petition rights” in privately owned shopping centers. The appellant in that case was a shopping mall owner who enforced a general policy against noncommercial expressive activity to exclude from his property a group of students who were distributing pamphlets. The owner argued that the California provision interfered with his right under the First Amendment “to refrain from speaking” and “not to be forced by the state to use his property as a forum for the speech of others.” Writing for the Court, Justice Rehnquist rejected the owner’s arguments, noting that the property was “not limited to the owner’s personal use,” but was instead “a business establishment that is open to the public” and that “the views expressed by members of the public in passing out pamphlets . . . would not likely be identified with those of the owner.”

So which is Twitter, a parade or a shopping center? The Eleventh Circuit thought a parade and the Fifth Circuit a shopping mall. But nobody really knows. Online platforms are different from both and more complicated than either. Like parade organizers, online platforms tailor their content feeds to attract eyeballs and create atmospheres that are consistent with their brands. They have reputations to uphold and users to attract. But almost all of the decisions are made algorithmically, to maximize views and clicks, not manually, to express a message. And, like shopping malls, online platforms are businesses open to the public; readers will attribute posts to their third-party speakers, not the platforms’ owners; and, unlike a moving parade where “disclaimers would be quite curious,” platform disclaimers of third-party views are not only possible, but already in place.

Pitfalls of the Easy Way

One way the Supreme Court could rule in the NetChoice cases—perhaps the easiest way—is narrowly, by pure application of precedent: The Florida and Texas laws violate the First Amendment because they require social-media companies to alter the expressive content on their platforms; or, alternatively, they are consistent with the First Amendment because social-media companies have opened their platforms to the public and the views of platform users will not likely be attributed to platform owners. This approach exemplifies classic judicial minimalism. Stay within the contours of the present case and avoid speculating about the future because you might very well be wrong.

But taking the easy path might be problematic for a few reasons. First, a basic application of precedent either to approve or disapprove of the Texas and Florida social-media laws would leave in place the deep tension in the Court’s cases, which simultaneously guarantee First Amendment protection against government-compelled speech, but also recognize the government’s power under Pruneyard and related cases to force private property owners to host third-party speech—even speech they disagree with, and even where speech is a central part of their own mission (as in Rumsfeld v. Forum for Academic and Institutional Rights, which upheld a law forcing law schools to allow military recruiters on their campuses).

Second, a terse decision under contested precedents might appear to be driven by the Court’s political priors rather than established law—an ironic result given that a goal of Robertsian judicial minimalism is to produce apolitical, balls-and-strikes judging as often as is possible. But judicial umpiring can only work if we know what game we are playing. The last four decades have seen the Court call compelled-association claims as balls and strikes without ever explaining the rules of the game or whether it is even baseball we’re playing. Resolving another case without a theory of how to distinguish permissible from impermissible speech compulsion will continue that trend into the politically salient context of online speech. The result could be a public backlash à la Kelo or Roe that risks undermining the Court’s credibility.

Third, platforms are not monolithic. Part of what makes platforms so attractive is their wide-ranging functionality. Much like blogs, platforms allow users to author content and post it publicly to their individual pages. These individual user pages also serve another function. Members of the public who visit users’ pages are able to post their own comments or pictures in response, thereby expanding and livening (or, just as likely, disrupting and frustrating) the user’s original message. To help users and commenters craft eyeball-attracting content, platforms pitch in by providing GIF databases and photo-editing functionality. And, because making money is important too, platforms curate and repackage the body of material posted by their users into real-time content feeds that are customized for each user and prominently displayed (mixed with targeted ads) on the main page of the app or website. Social media’s cornucopia of functionality makes it poorly suited to minimalist resolution as parade- or shopping-mall–like.

The Common-Carrier Proposal

In principle, the Court could just split up the platform’s functionality into pieces and classify each separately. But that approach would only make more glaring the lack of a unifying theory to explain why parades and shopping malls are treated differently in the first place. Why does the First Amendment permit the government to compel a private shopping mall owner or private law school to serve as a platform for unwanted speech, but not a private parade organizer?

One of the most intriguing paths the Court could take is common-carrier doctrine—the idea that social-media platforms, like telecommunications companies, public utilities, transportation companies, and the like, might be so critical to the public well-being that they must provide service equally to all comers, despite companies’ typical freedom to accept or reject customers as they see fit.

In a 2021 decision, Justice Clarence Thomas raised the possibility that a similar principle should govern social media. In Biden v. Knight First Amendment Institute at Columbia University he observed that social-media platforms “provide avenues for historically unprecedented amounts of speech,” but that never before has “control of so much speech [been] in the hands of a few private parties.” He went on to suggest that perhaps social-media companies, like railroads and telephone companies, should be required to accept all users equally, without regard to their viewpoints. Most people, after all, would not want telephone companies to deny service to individuals based on their political or religious views. Perhaps online platforms should be treated similarly.

But common-carrier doctrine is a difficult path. The Court’s compelled-association cases may lack a unifying theoretical principle, but the line of precedent is well-developed and comparatively robust. By contrast, the common-carrier doctrine is rarely discussed in the Court’s decisions and is dramatically undertheorized. Developing a full-fledged theory of common-carrier doctrine as related to First Amendment freedom of association and compelled speech cases would be an enormous undertaking.

Moreover, common-carrier treatment would be a costly solution to the social-media problem. Treating entities as common-carriers interferes with their freedom to respond to market demands and reduces consumer welfare; undermines corporate profitability and spending on innovation by increasing compliance and litigation costs; and imposes licensing requirements and other barriers that solidify market leaders and benefit incumbents. Such requirements are accepted as necessary for the public good but are burdensome and applied sparingly, typically where industry incumbents hold significant market power and some economic or social barrier makes a market-driven solution unlikely.

Yet common-carrier doctrine may be a better mode of analysis. Unlike the shopping mall–parade distinction, the common-carrier doctrine goes to the very heart of the problem: Have large social-media platforms become so dominant and so critical to modern public discourse that they can be forced to convey messages expressing viewpoints they disagree with, or have they not? That is the real issue, and common-carrier doctrine is the right analytical tool to address it. No one cares whether online platforms are parades or shopping malls or purple petunias. What people want to know is whether platforms’ role in societal discourse is so pivotal and their market power so strong as to warrant them being forced to carry the public’s messages.

When the government may compel private parties to host speech is a question that has bedeviled commentators for decades. But the stars are aligning for the Court to take up the question again. The Court is all but certain to review the challenged Florida and Texas laws. And it has already agreed next term to review a pair of decisions from the Sixth and Ninth Circuits that consider whether government officials may block users from commenting on their Twitter feeds. The rush of social-media questions shows no sign of slowing. Whichever way it decides the merits, the Court should resist the temptation to fall back on precedential distinctions between parades and shopping malls. Online platforms are a generational technology that defies analogy and requires fresh consideration via appropriate doctrinal tools.