Making True Threats Is a Crime

But what counts as true?

Illustration showing a gun, but the thing coming out of it is not a bullet but a flag that says "only kidding"
Illustration by The Atlantic. Source: Getty.

The internet is full of threats—violent, personal, severe threats, the recipients of which are left to assess the threats’ seriousness, to figure out how to protect themselves. Perhaps the most well-documented and well-known threats were those made to election workers during the 2020 presidential campaign. In voice messages, emails, and social-media posts, election deniers told state and local officials they would “get popped,” “hang from a tree,” and be “executed for treason.” One Facebook message said, “You raided an office. You broke the law. STOP USING YOUR TACTICS. STOP NOW. Watch your back. I KNOW WHERE YOU SLEEP, I SEE YOU SLEEPING. BE AFRAID, BE VERRY AFFRAID. I hope you die.”

I was part of a group of legal experts asked by Reuters reporters to help determine whether these threats were unlawful. When contacted by reporters, many who sent these messages were unapologetic; some were even proud of their messages, viewing them as patriotic. They downplayed the intimidating nature of their communications and characterized them as harmless “trolling.” Law enforcement showed little interest in tracking down the speakers, reasoning that the communications were lawful—a conclusion that, in many instances, I and other First Amendment scholars disagreed with.

It’s admittedly a tough determination to make. The First Amendment does not protect threats—that much is clear. But the Supreme Court has provided little guidance on what qualifies as a threat. That may soon change, when the Court decides Counterman v. Colorado, which will be argued next Wednesday.

In a 2003 case involving cross burnings, the Supreme Court defined a threat as a statement in which “the speaker means to communicate a serious expression of an intent to commit an act of unlawful violence to a particular individual or group of individuals.” But which expressions are “serious”? And whose perspective counts when making that determination—the speaker’s or the recipient’s? A small minority of courts have concluded that the First Amendment requires the speaker to intend to make the recipient fear violence or know that would be the result (sometimes referred to as a “subjective” or “specific intent” standard). Most courts, including Colorado’s, have focused instead on whether a reasonable recipient would view the message, in context, as a threat (sometimes referred to as an “objective” or “reasonable recipient” standard). Counterman will decide which standard satisfies the First Amendment.

That may sound like a narrow legal technicality, but the stakes are very high. The standard the Court adopts will affect not just speakers’ liability but also the safety and peace of mind of the recipients of threatening speech and the nature of online and offline speech cultures. The Supreme Court must adopt a standard that protects political and other valuable expression while at the same time shielding targets from the substantial harm that threats inflict.

Adopting a standard that focuses on what was in the speaker’s mind would allow individuals more latitude to vent and lash out, in part because it is hard to prove speaker intent. Like the election denier, many speakers could claim that their intent was to troll, rant, or even entertain rather than threaten acts of violence. But focusing solely on what the speaker intended would leave vulnerable targets at the mercy of threatening speech.

The Court should follow the majority of lower courts and adopt a “reasonable recipient” standard like Colorado’s. That standard protects vulnerable targets from the harms the Court has identified, including “the fear of violence and the disruption that fear engenders.” These harms relate not to the speaker’s state of mind but to the recipient’s well-being. Threats can turn targets’ lives upside down, forcing them to quit their job, hide their identity, and stop going out in public.

Consider the facts in Counterman. In April 2017, Billy Raymond Counterman was convicted of making threatening statements to a local musician on Facebook over the course of two years. Among other things, he criticized the woman’s “arrogance” when she did not respond to his messages, asked her “What do you fear?,” indicated he had seen her in public, and accused her of “not being good for human relations.” When she did not respond, Counterman told the musician to “Die” and “Fuck off permanently.” The musician contacted law enforcement, obtained a protective order against Counterman, and canceled public performances. The First Amendment generally protects nasty and abusive language, in part to allow speakers to engage in political rhetoric. But regulating unwanted communications like these, repeatedly directed to an individual over social media, does not threaten to suppress valuable expression on matters of public concern. Indeed, threatening communications chill public discussion and instill terror in their recipients. An objective standard recognizes this reality. At the same time, it does not criminalize all violent and abusive language or suppress political speech. Courts and juries must consider the full context of speech, including the words used, the recipient’s reaction, the course and manner of communication, and the relationship between speaker and recipient.

The Supreme Court’s own precedents have focused on contextual factors, such as the conditional nature of a threat, the audience’s reaction, and the general context of the communication—for example, whether the speech was part of a public political rally or a civil-rights boycott. This approach protects political hyperbole and rhetoric. It also makes it less likely that speakers will be subject to punishment or liability based solely on their use of crude language or a good-faith misunderstanding of words taken out of context. Such a standard is in line with the First Amendment’s goal of protecting robust public debate on important matters while excluding speech that threatens physical and other harms.

As Counterman shows, threats are not limited to the political context. Anyone can be subjected to threats online, though women seem to receive a disproportionate share. And what happens online does not necessarily stay online; over and over, threats proffered on social media have been linked with real-world violence.

Speakers should not be jailed or subject to civil liability merely for their use of abusive or violent language. But neither should they escape liability when their words reasonably cause someone to fear serious bodily harm or death. Consider a voice message left for employees at a voting-machine company: “You’re all fucking dead. We’re going to fucking kill you all.” Or the graphic threats of rape and murder frequently sent to women on social media. These communications can’t be brushed off as harmless venting or trolling. Nor are they part of any robust and uninhibited political discourse protected by the First Amendment. They are unprotected speech, and the law should treat them as such.