Over one hundred years ago, in a case called Schenck v. United States, Justice Oliver Wendell Holmes created a meme.
Holmes wanted to illustrate why freedom of speech was not—and could never be—absolute. “The most stringent protection of free speech,” Holmes wrote, “would not protect a man in falsely shouting fire in a theatre and causing a panic.” This was because, Holmes explained, “[t]he question in every case is whether the words used are used in such circumstances and are of such a nature as to create a clear and present danger that they will bring about the substantive evils that Congress has a right to prevent.” A false cry of fire in a theater, he implied, surely posed this kind of clear and present danger.
It might have surprised Holmes to know that more than a century later, his claim about the constitutionality of false cries of fire in theaters has become one of the most famous hypotheticals in American constitutional law. And it has acquired a remarkable significance in debates about speech regulation. On a near-daily basis, the fire meme is invoked to support restricting a wide variety of speech, from health misinformation, to former presidents’ social media posts, to Tucker Carlson’s television show.
In response to these near-constant invocations of the hypothetical, many have suggested that the meme should be struck from our lexicon in conversations about free speech. These arguments typically come in three forms. The first is that the meme is so tainted by the shameful nature of the outcome in Schenck, which essentially rubber-stamped the Wilson administration’s persecution of political dissidents during World War I, that reference to it is “historically ignorant” and embarrassing. The second is that regardless of its rotten roots, the meme is not, and perhaps never was, a good description of the law: that those who invoke Holmes’ famous hypothetical in public debates are relying upon “an empty rhetorical device … in a long-overturned case about jailing draft protestors.” The third is that it doesn’t matter whether or not it’s good law—the meme is so often used to attempt to justify blatantly unconstitutional laws, it is dangerous for this reason alone.
As we show in this post, none of these arguments for banishing the analogy from free speech debates withstands scrutiny. Without a doubt, the case that birthed the fire meme, Schenck, is as big an affront to free speech as they come. But despite these origins, the fire meme is actually still good law—to an extent that surprised us both when researching this piece. Holmes’ comment may have been dicta when he said it, but it has since been (and continues to be) repeatedly relied upon by courts in First Amendment cases and provides the justification for important forms of speech regulation—including laws banning bomb threats, harassment, and other types of coercive speech. In other words, for courts, the fire meme plays an important, and almost entirely uncontroversial, role in delimiting the boundaries of First Amendment protection. For this reason, the analogy also plays—or at least can play—an instructive role in popular debates about speech regulation. Analogies provide a powerful means of encapsulating complex ideas in a concrete form. The fire meme is one such analogy: It is a vivid illustration of a limitation on First Amendment protection (and one that, contrary to what is commonly assumed, actually has nothing to do with the limits that apply to speech that incites unlawful action).It is not merely a stand-in for the truism that “not all speech is protected,” but says something deeper about how to think about the value of speech.
This doesn’t mean, of course, that the fire meme doesn’t get invoked problematically to justify what would clearly be unconstitutional laws. But the problem in these cases is not the invokers’ understanding of the law, but their application of the law to the facts.
In what follows, we attempt to add light (but not heat!) to the conversation about false cries of fire in crowded theaters by taking a close look at how the fire meme was employed in Schenck and how it has been subsequently interpreted. As we explain, the problem is not that those who invoke the fire meme are relying on a “legal irrelevance” that “hasn’t been the law in the U.S. for almost 50 years.” The problem is that those who invoke it tend to rely on an overly reductive view of how public discourse works in the digital age (one which conceives of its participants as lemmings, more or less).
Another way of putting this is to say that looking again at the fire meme pushes us to think about the limits of the First Amendment in more nuanced ways than the inside baseball conversation about false cries of fire in crowded theaters often permits.
How the Fire Meme Survives Its Rotten Roots
We begin at the beginning—with Schenck. Any discussion of the fire analogy has to acknowledge the fact that the analogy has rotten roots.
Schenck involved the criminal prosecution of two socialists, Charles Schenck and Elizabeth Baer, who distributed roughly 15,000 leaflets to army conscripts during World War I that argued that conscription violated the 13th Amendment and urged them to assert their rights. The two were charged with, and ultimately convicted of, impeding military recruitment, in violation of the Espionage Act that Congress passed shortly after the United States entered World War I. Schenck and Baer challenged their convictions on First Amendment grounds, but the Supreme Court—in an opinion written by Holmes—held that, because the pamphlets could have plausibly impeded the war effort, the two defendants could be punished for their speech.
This conclusion generated strong criticism at the time because, as its critics recognized and seems obvious now, it could be used to justify the repression of any anti-war speech that might plausibly have an impact on the success of the war effort. These criticisms had an effect. In later Espionage Act cases, Holmes reinterpreted Schenck’s clear and present danger test to require the government to show that the speech in question posed not just a plausible but a very grave and imminent threat to the war effort before it could be punished.
Holmes continued to cite Schenck as good law, however, and never rejected the context-specific view of freedom of speech it relied upon.And although the clear and present danger test that Holmes introduced in Schenck was famously overruled 50 years later in Brandenburg v. Ohio, the fire meme survived.
Indeed, the continuing validity of the claim that there is a category of unprotected speech analogous to false cries of fire in theaters was explicitly affirmed by the famously speech-absolutist Justice Douglas in the concurring opinion he wrote in Brandenburg—an opinion, it is worth noting, that was joined by Justice Black, another notably speech-protective justice. In that opinion, Douglas sharply distinguished cases like Schenck (and Brandenburg itself), in which the government sought to punish speech because of the dangerous ideas it communicated, from cases in which speech was prohibited because it caused harm in different ways. In the first set of cases, Douglas argued, the government had no power to tell people what to think. However, in the latter set of cases, Douglas argued, the speech could be “subject to regulation.” Douglas pointed specifically to the “example … of one who falsely shouts fire in a crowded theatre.” This was, he wrote, “a classic case where speech is brigaded with action” and could therefore incur liability for the harms it caused.
Later Supreme Court opinions affirmed Douglas’ claim that false cries of fire remained unprotected speech even after Brandenburg. For example, the Court’s opinion in Claiborne Hardware v. NAACP cited Schenck and Holmes’ hypo for the proposition that, while “mere advocacy of the use of force … does not remove speech from the protection of the First Amendment, words that create an immediate panic are not entitled to constitutional protection.” A few years later, Justice Scalia (no First Amendment softie!) noted, in an opinion for the Court, that “a ban on shouting fire can be a core exercise of the State’s police power to protect the public safety.” Lower courts repeatedly reached the same conclusion about the continuing salience of Holmes’ hypothetical.
As these cases make clear, the fire meme survived Schenck’s (effective) overruling because, although Holmes used the example of false cries of fire to lend support to his conclusion that anti-war advocacy could be prosecuted, the two kinds of speech can and should be distinguished, as they were in Brandenburg. Speech that encourages unlawful behavior—like the speech the government prosecuted Schenck and Baer for—is typically believed to cause harm by persuading others of the truth of its ideas about matters of broad public concern (such as the constitutionality of conscription, or, as in Brandenburg, the “suppress[ion] of the white Caucasian race”). But it is generally constitutionally protected, because it leaves its audience with time to reflect on the ideas it disseminates, to get other opinions, to decide for themselves what to believe.
Contrast this with false cries of fire in crowded theaters, or other speech that causes public panic. This kind of speech communicates only the idea that the theater is on fire, or the plane is being hijacked, or the letter is full of anthrax—not matters that we typically think of as touching on important questions of public policy—and it leaves its audience with no time at all to evaluate the claim’s veracity. Who is going to stop and verify anything when the theater may already be on fire, or the plane about to be hijacked, or the letter you are holding spreading, as you read it, a deadly disease? This is not idle speculation about human behavior. When Holmes was writing, as one commentator notes, “[f]alse shouts of ‘fire’ were a pervasive problem that plagued theaters throughout the United States and the United Kingdom, resulting in hundreds of deaths and injuries.” To Holmes, the hypothetical was not actually a hypothetical. This kind of speech was especially harmful, and foreseeably so. And it continues to cause harm.
This difference in how the two kinds of speech operate explains why courts continue to assume that most encouragement to unlawful action enjoys strong constitutional protection but speech that causes panic not so much. While the former contributes something to the “public … debate about matters of public concern” that the First Amendment protects, and leaves it to the audience to choose to act or not act in response to it, the latter does not. It is in this respect that this kind of speech has, as Holmes put it, “all the effect of force.” It is more akin to a gun to the head than a proposition.
Understanding what distinguishes false cries of fire in crowded theaters from speech that encourages unlawful action also helps explain why courts have invoked the fire meme to justify the denial of constitutional protection to other kinds of speech that act to coerce and intimidate, rather than to persuade. For example, courts have found that harassment constitutes unprotected speechbecause, like a false cry of fire in a crowded theater, it impacts its audience not by reason of the ideas it communicates but by the inconvenience and fear it causes. Courts have similarly invoked the fire meme to explain why threats are unprotected speech. These are also examples of speech that is, as Douglas would say, “brigaded with action.” All of which is to say that the fire meme is regularly invoked to support an important, and not so narrow, set of exceptions to First Amendment protection for speech.
This means that when politicians or online commentators claim that false cries of fire in crowded theaters aren’t constitutionally protected, they are making an entirely accurate statement about First Amendment law. And an important one: If you are in a crowded theater, you cannot, in fact, yell “fire!” knowing that there isn’t one (... unless you’re conducting an authorized fire drill, which is a wise exception). Nor can you display or threaten to detonate a false bomb, unless you are a police officer engaged in a training exercise (... another reasonable loophole). You also cannot call the authorities and claim, falsely, that there is a bomb or a biological weapon on an airplane. Nor, for that matter, can you falsely claim that a fire or bomb will “kill, injure, or intimidate any individual or unlawfully … damage or destroy any building, vehicle, or other real or personal property.” All of these laws, and many more like them, are (hopefully uncontroversially) very good things.
This doesn’t mean, of course, that there is no problem with how the fire meme gets used in contemporary debates. The meme has, well, become a meme and taken on a life of its own, popping up in all sorts of unsuitable contexts. But the problem does not stem from the mistaken belief that First Amendment protection doesn’t extend to speech that has all the effects of force, or that such a category is so marginal and negligible as to be all but irrelevant. The problem is a failure to understand why this category of speech is unprotected, and therefore what it can be said to include. The reference to shouts of fire is not just another way of saying “the First Amendment is not absolute”—the analogy says something important about why such speech is unprotected, and when speech is no longer simply speech but an act.
What Else is Like a False Cry of Fire in a Crowded Theater?
The boundaries of the category of “speech brigaded with action”—or what some courts and commentators describe simply as “verbal acts”—have always been contested, and continue to be contested today. This is because lots of speech acts on its audience by scaring or coercing it, and yet much of that speech clearly also contributes to public debate about public matters. (“They’re going to take our guns,” for example.) The result is that courts have to figure out when speech is so coercive and so without value to public discussion of public affairs that it can be denied constitutional protection without threatening First Amendment values. There can be, and has been, significant disagreement about where to draw this line.
But there is nothing about the fire meme itself that makes these disagreements any harder to resolve. These disagreements are hard to solve because defining the boundaries of permissible discourse in a democratic society is a very difficult thing to do. And it’s not getting any easier.
The rise of social media platforms has generated all kinds of anxieties about how democratic discourse works in the contemporary public sphere. There is a commonly rolled-out set of observations about the effects of social media on public debate: it has democratized platforms for speech, for good and for ill, and dramatically sped up its dissemination. It has “unwittingly dissolved the mortar of trust, belief in institutions, and shared stories that had held a large and diverse secular democracy together.” And it has, via the use of algorithmic feeds that are optimized for engagement, driven people into echo chambers and toward extremism. When a user reads these posts without context or counterspeech, they can be manipulated into believing whatever information they see in their feed. Foreign trolls are taking advantage of this to trick people into mindlessly taking political action they otherwise would not. These tropes, and many more you can probably fill in yourself, have themselves become memes.
It is the ascendance of this picture of online discourse that has made public figures and others increasingly prone to invoke Holmes’ fire meme to justify the suppression of speech. When Sen. Amy Klobuchar compares the algorithmic amplification of misinformation to crying fire in a crowded theater, or the governor of New York says that platforms hosting white supremacist propaganda is the same, or a Democratic representative says allowing Donald Trump back on Twitter would yet again be equivalent, they are relying on these ideas that users of social media are left little meaningful choice but to believe what they read and are manipulated into action in a way that defies conscious choice.
If that’s your conception of online speech, the intuitive appeal of the fire meme is understandable. And it is definitely true that social media platforms are hardly the model of rational public discourse. These spaces are unquestionably rife with manipulative content and designed for commercial, not democratic, ends.
Nevertheless, the effort to extend the exception for speech that has all the effects of force to Donald Trump’s tweets, or to vaccine disinformation, or to foreign trolling, rests on a ridiculously reductive view of the contemporary speech ecosystem. Do the readers of Donald Trump’s tweets really lack the same freedom of choice that someone faced with an (apparently imminent) threat of fire lacks? Of course not—we can readily attest to that as people who have read far too many of Trump’s tweets. Similarly, in almost every case of coronavirus misinformation, people have plenty of opportunity to seek out other sources to check whether the information is correct. The speech, in this respect, is clearly less coercive than a false cry in a movie theater and much more relevant to debate on matters of public concern. To treat it as the same is to risk denying constitutional protection to huge swathes of the digital public sphere.
This is why politicians and others should hesitate before invoking the Schenck analogy willy-nilly to justify restricting speech: not because it states something false about the world, but because it states something true. Speech can coerce, rather than persuade, and for that reason lose First Amendment protection. But the category of speech that is so coercive that it loses its constitutional protection should be carefully drawn so that it does not impinge too much on valuable public debate. This explains why courts have tended to require proof that the speaker knew or was reckless about the panic he or she would cause, and have refused to apply the exception to panic-inducing but nevertheless valuable speech like Orson Welles’ “War of the Worlds” broadcast. (Although note that, contrary to what is often assumed, actually causing a panic is not always a necessary requirement for conviction.)
This does not mean, however, that the fire analogy can never be relevant to online or broadcast speech. Take the example of a congressional campaign manager who falsely tweeted during the peak hours of hurricane Sandy that already-panicked citizens were about to lose power. Or the Federal Communications Commission’s hoax rule that prohibits knowingly broadcasting false information concerning a crime or a catastrophe, passed after a radio announcer broadcast a mock emergency bulletin warning of a nuclear attack during the first Gulf War. Or false rumors, including some circulating on social media, that there was an active shooter at a boxing match in Brooklyn’s Barclays Center—at a time when the nation was already on edge from two recent mass shootings—which caused a stampede. If done with the requisite intent, it is not obviousthat all these false cries should be considered protected speech or that those who suggest otherwise are simply ignorant about the First Amendment.
When drawing the line between speech and verbal acts, commentators should remember the reason why false cries of fire are excluded from First Amendment protection is not merely that they are harmful. After all, the First Amendment protects lots of harmful speech. It is that false cries of this kind bypass reasoning such as to have the effect of force and contribute very little to public debate about public affairs. It is this that fundamentally distinguishes them as “speech brigaded with action” rather than just speech. Few that invoke the meme are likely aware of the importance of this distinction, but it remains a crucial part of modern First Amendment doctrine that we should not forget.
Analogies get misused all the time. Take, for example, the other famous metaphor Holmes introduced into American free speech law: namely, the metaphor of the “free trade,” or marketplace of ideas. This metaphor is frequently invoked on the other side of free speech debates, to suggest that speech regulation is somehow illegitimate because it interferes with the operation of an ideal speech market. Like the analogy to false cries of fire in a theater, the evocative power of this metaphor makes it a powerful weapon in free speech debates, and one that gets used to justify a deregulatory conception of free speech that ignores plenty of evidence of market failures in many arenas.
That analogies can be misused does not mean, however, that we should—or could—embrace an analogy-free legal debate. Analogies are also useful because they translate difficult conceptual ideas into more accessible language. They concretize complicated concepts and therefore broaden the base of people who can engage in the relevant debates. They can even, as is the case with the fire meme, be enlightening by highlighting the false factual premises that the analogizer is relying on.
This is why, even though we had some hesitation about adding fuel to the (ahem) fire of the debate about Holmes’ analogy, we thought it worthwhile to explore what not-so-flickering light the fire meme sheds on First Amendment law. To dismiss the invocation of the meme as a sign of ignorance or stupidity is to try to transform an important legal, public, and cultural debate about the proper regulation of speech in modern society into a technocratic debate about “holdings” and “dicta” (which, to reiterate, is a poor description of the meme’s doctrinal status today!) and to distract from the substantive issue.
The focus should remain on what is really troubling about most uses of the meme: not the invocation of a hypothetical that states a valid principle of First Amendment law, but the mistaken supposition about how people are affected by information they consume that motivates many of those who invoke the analogy in contemporary debates. False and inflammatory speech online is equated with speech acts that have coercive force. Participants in public discourse are too often characterized as mere automatons who can be made to join a stampede (or storming of the Capitol) at any moment. It is better to confront this incorrect view head-on even if that is harder to do than attempting to extinguish cries of false cries of fire in theaters.
This assumption is understandable given the cases that the fire meme is most commonly associated with, Schenck and Brandenburg, are cases in which the speech was believed to cause harm by inciting its audience to take unlawful action. But in fact, the meme is an illustration of a different exception to First Amendment protection than the low-value category of incitement.
In his famous dissent in Abrams v. United States, Holmes insisted that he had “never … seen any reason to doubt that the questions of law … before this Court in the Case of Schenck … were rightly decided.” Abrams v. United States, 250 U.S. 616, 627 (1919) (Holmes J., dissenting). Holmes also relied heavily upon the famous clear and present language from Schenck in his dissent in Gitlow v. New York, 268 U.S. 652, 672-73 (1925) and signed on to a concurring opinion written by Justice Brandeis that described the rule articulated in Schenck as “the rule of reason.” Schaefer v. United States, 251 U.S. 466, 482 (1920) (Brandeis J., concurring).
For example, in 1975, the Fifth Circuit affirmed the constitutionality of a federal law that made it a crime to convey false information “concerning an attempt or alleged attempt” to hijack a plane because it found that the speech the statute prohibited “approximates the false cry of fire” and therefore could be criminalized without posing a First Amendment problem. United States v. Irving, 509 F.2d 1325, 1331 (5th Cir. 1975). A few years later, an Ohio appeals court similarly relied upon the fire meme to affirm the constitutionality of a state law that made it a crime to “cause serious public inconvenience or alarm by... initiating or circulating a report … of an alleged or impending fire, explosion, crime, or other catastrophe, knowing that such report … is false.” State v. Loless, 31 Ohio App. 3d 5, 6 (Ohio Ct. App. 1986). And just a few years ago, the Ninth Circuit relied upon the fire meme to justify its conclusion that a federal law that made it a crime to convey false or misleading information about the presence of a biological weapon in interstate commerce was not inconsistent with the First Amendment. United States v. Keyser, 704 F.3d 631, 637 (9th Cir. 2012).
On this point see Judge Wood’s concurring opinion in the famous Skokie Nazi rally case, in which he argued that the Nazi rally was “not equivalent to the sudden and unfounded cry of “fire” in a crowded and unsuspecting theatre” because “there [was] ample warning of the proposed event.” Collin v. Smith, 578 F.2d 1197, 1210 (7th Cir. 1978) (Wood J., concurring) (order of quotes altered).
As the Supreme Court of Appeals of West Virginia put it so eloquently: “A citizen must have the right to petition his government and to petition it forcefully and repeatedly without any danger of being found guilty of a crime. … Nevertheless, there is a point where legitimate inquiry ends and harassment begins. There comes a point where one cannot repeatedly call a public servant and threaten to fry him in oil.” State v Thorne, 175 W. Va. 452, 455 (1985) (noting that, [t]o achieve order, the legislature may properly limit certain activities such as falsely shouting fire in a crowded theater, announcing the sailing date of military transports or the number and location of troops, and … harassing others over a telephone.”). See also Lyle v. Warner Bros. Television Prods., 132 P.3d 211, 234 (Ca. 2006) (Chin J., concurring) (“Just as criminal threats are not protected, just as no one has the right to falsely shout ‘fire’ in a crowded theater, … [harassing] speech that is directed, or ‘aimed at a particular employee because of her race, sex, religion, or national origin,’ is not protected.”).
Haughwout v. Tordenti, 332 Conn. 559, 561, 211 A.3d 1, 3 (2019) (noting that “recent history … has led federal and state courts to describe threats of gun violence and mass shootings as the twenty-first century equivalent to the shout of fire in a crowded theater … envisioned by Holmes”); Garcia v. State, 583 S.W.3d 170, 175 (Tex. App. 2018) (citing the fire meme as support for the claim that “even if the threat of violence is a subjectively false statement of fact, it is not worthy of constitutional protection”); Milo v. New York, 59 F.Supp. 3d 513, 527 (EDNY 2014) (threat to bring a machine gun to school, given the context in which it occurred, “was tantamount to ‘yelling fire in a theatre” and therefore did not enjoy constitutional protection); State ex rel. RT, 781 So. 2d 1239, 1243 (La. 2001) (“Words which by their very utterance may cause alarm, public disruption, or constitute a signal to prompt unlawful action fall within the principle of the false cry of “fire” in a crowded theater and are characterized as verbal acts unprotected by constitutional prohibitions against restraint of free speech. We have no trouble concluding that the state has a legitimate interest in criminalizing apparently serious, albeit false, bomb threats, notwithstanding that the crime is committed through the medium of speech. The First Amendment does not protect criminal activity, even when carried out with words.”)
evelyn douek is a senior research fellow at the Knight Institute.
Genevieve Lakier is the Knight Institute's senior visiting research scholar, 2021-2022.