Notice: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.
SUPREME COURT OF THE UNITED STATES
No. 22–138
BILLY RAYMOND COUNTERMAN, PETITIONER v. COLORADO
ON WRIT OF CERTIORARI TO THE COURT OF APPEALS OF COLORADO
[June 27, 2023]
Justice Kagan delivered the opinion of the Court.
True threats of violence are outside the bounds of First Amendment protection and punishable as crimes. Today we consider a criminal conviction for communications falling within that historically unprotected category. The question presented is whether the First Amendment still requires proof that the defendant had some subjective understanding of the threatening nature of his statements. We hold that it does, but that a mental state of recklessness is sufficient. The State must show that the defendant consciously disregarded a substantial risk that his communications would be viewed as threatening violence. The State need not prove any more demanding form of subjective intent to threaten another.
I
From 2014 to 2016, petitioner Billy Counterman sent hundreds of Facebook messages to C. W., a local singer and musician. The two had never met, and C. W. never responded. In fact, she repeatedly blocked Counterman. But each time, he created a new Facebook account and resumed his contacts. Some of his messages were utterly prosaic (“Good morning sweetheart”; “I am going to the store would you like anything?”)—except that they were coming from a total stranger. 3 App. 465. Others suggested that Counterman might be surveilling C. W. He asked “[w]as that you in the white Jeep?”; referenced “[a] fine display with your partner”; and noted “a couple [of] physical sightings.” 497 P. 3d 1039, 1044 (Colo. App. 2021). And most critically, a number expressed anger at C. W. and envisaged harm befalling her: “Fuck off permanently.” Ibid. “Staying in cyber life is going to kill you.” Ibid. “You’re not being good for human relations. Die.” Ibid.
The messages put C. W. in fear and upended her daily existence. She believed that Counterman was “threat[ening her] life”; “was very fearful that he was following” her; and was “afraid [she] would get hurt.” 2 App. 177, 181, 193. As a result, she had “a lot of trouble sleeping” and suffered from severe anxiety. Id., at 200; see id., at 194–198. She stopped walking alone, declined social engagements, and canceled some of her performances, though doing so caused her financial strain. See id., at 182–183, 199, 201–206, 238–239. Eventually, C. W. decided that she had to contact the authorities. Id., at 184.
Colorado charged Counterman under a statute making it unlawful to “[r]epeatedly … make[] any form of communication with another person” in “a manner that would cause a reasonable person to suffer serious emotional distress and does cause that person … to suffer serious emotional distress.” Colo. Rev. Stat. §18–3–602(1)(c) (2022). The only evidence the State proposed to introduce at trial were his Facebook messages.[1]
Counterman moved to dismiss the charge on First Amendment grounds, arguing that his messages were not “true threats” and therefore could not form the basis of a criminal prosecution. In line with Colorado law, the trial court assessed the true-threat issue using an “objective ‘reasonable person’ standard.” People v. Cross, 127 P. 3d 71, 76 (Colo. 2006). Under that standard, the State had to show that a reasonable person would have viewed the Facebook messages as threatening. By contrast, the State had no need to prove that Counterman had any kind of “subjective intent to threaten” C. W. In re R. D., 464 P. 3d 717, 731, n. 21 (Colo. 2020). The court decided, after “consider[ing] the totality of the circumstances,” that Counterman’s statements “r[o]se to the level of a true threat.” 497 P. 3d, at 1045. Because that was so, the court ruled, the First Amendment posed no bar to prosecution. The court accordingly sent the case to the jury, which found Counterman guilty as charged.
The Colorado Court of Appeals affirmed. Counterman had urged the court to hold that the First Amendment required the State to show that he was aware of the threatening nature of his statements. Relying on its precedent, the court turned the request down: It “decline[d] today to say that a speaker’s subjective intent to threaten is necessary” under the First Amendment to procure a conviction for threatening communications. Id., at 1046 (quoting R. D., 464 P. 3d, at 731, n. 21). Using the established objective standard, the court then approved the trial court’s ruling that Counterman’s messages were “true threats” and so were not protected by the First Amendment. 497 P. 3d, at 1050. The Colorado Supreme Court denied review.
Courts are divided about (1) whether the First Amendment requires proof of a defendant’s subjective mindset in true-threats cases, and (2) if so, what mens rea standard is sufficient. We therefore granted certiorari. 598 U. S. ___ (2023).
II
True threats of violence, everyone agrees, lie outside the bounds of the First Amendment’s protection. And a statement can count as such a threat based solely on its objective content. The first dispute here is about whether the First Amendment nonetheless demands that the State in a true-threats case prove that the defendant was aware in some way of the threatening nature of his communications.[2] Colorado argues that there is no such requirement. Counterman contends that there is one, based mainly on the likelihood that the absence of such a mens rea requirement will chill protected, non-threatening speech. Counterman’s view, we decide today, is the more consistent with our precedent. To combat the kind of chill he references, our decisions have often insisted on protecting even some historically unprotected speech through the adoption of a subjective mental-state element. We follow the same path today, holding that the State must prove in true-threats cases that the defendant had some understanding of his statements’ threatening character. The second issue here concerns what precise mens rea standard suffices for the First Amendment purpose at issue. Again guided by our precedent, we hold that a recklessness standard is enough. Given that a subjective standard here shields speech not independently entitled to protection—and indeed posing real dangers—we do not require that the State prove the defendant had any more specific intent to threaten the victim.
A
“From 1791 to the present,” the First Amendment has “permitted restrictions upon the content of speech in a few limited areas.” United States v. Stevens, 559 U. S. 460, 468 (2010). These “historic and traditional categories” are “long familiar to the bar” and perhaps, too, the general public. Ibid. One is incitement—statements “directed [at] producing imminent lawless action,” and likely to do so. Brandenburg v. Ohio, 395 U. S. 444, 447 (1969) (per curiam). Another is defamation—false statements of fact harming another’s reputation. See Gertz v. Robert Welch, Inc., 418 U. S. 323, 340, 342 (1974). Still a third is obscenity—valueless material “appeal[ing] to the prurient interest” and describing “sexual conduct” in “a patently offensive way.” Miller v. California, 413 U. S. 15, 24 (1973). This Court has “often described [those] historically unprotected categories of speech as being of such slight social value as a step to truth that any benefit that may be derived from them is clearly outweighed by the social interest” in their proscription. Stevens, 559 U. S., at 470 (internal quotation marks omitted; emphasis deleted).
“True threats” of violence is another historically unprotected category of communications. Virginia v. Black, 538 U. S. 343, 359 (2003); see United States v. Alvarez, 567 U. S. 709, 717–718 (2012) (plurality opinion). The “true” in that term distinguishes what is at issue from jests, “hyperbole,” or other statements that when taken in context do not convey a real possibility that violence will follow (say, “I am going to kill you for showing up late”). Watts v. United States, 394 U. S. 705, 708 (1969) (per curiam). True threats are “serious expression[s]” conveying that a speaker means to “commit an act of unlawful violence.” Black, 538 U. S., at 359. Whether the speaker is aware of, and intends to convey, the threatening aspect of the message is not part of what makes a statement a threat, as this Court recently explained. See Elonis v. United States, 575 U. S. 723, 733 (2015). The existence of a threat depends not on “the mental state of the author,” but on “what the statement conveys” to the person on the other end. Ibid. When the statement is understood as a true threat, all the harms that have long made threats unprotected naturally follow. True threats subject individuals to “fear of violence” and to the many kinds of “disruption that fear engenders.” Black, 538 U. S., at 360 (internal quotation marks omitted). The facts of this case well illustrate how.[3]
Yet the First Amendment may still demand a subjective mental-state requirement shielding some true threats from liability. The reason relates to what is often called a chilling effect. Prohibitions on speech have the potential to chill, or deter, speech outside their boundaries. A speaker may be unsure about the side of a line on which his speech falls. Or he may worry that the legal system will err, and count speech that is permissible as instead not. See Philadelphia Newspapers, Inc. v. Hepps, 475 U. S. 767, 777 (1986). Or he may simply be concerned about the expense of becoming entangled in the legal system. The result is “self-censorship” of speech that could not be proscribed—a “cautious and restrictive exercise” of First Amendment freedoms. Gertz, 418 U. S., at 340. And an important tool to prevent that outcome—to stop people from steering “wide[] of the unlawful zone”—is to condition liability on the State’s showing of a culpable mental state. Speiser v. Randall, 357 U. S. 513, 526 (1958). Such a requirement comes at a cost: It will shield some otherwise proscribable (here, threatening) speech because the State cannot prove what the defendant thought. But the added element reduces the prospect of chilling fully protected expression. As this Court has noted, the requirement lessens “the hazard of self-censorship” by “compensat[ing]” for the law’s uncertainties. Mishkin v. New York, 383 U. S. 502, 511 (1966). Or said a bit differently: “[B]y reducing an honest speaker’s fear that he may accidentally [or erroneously] incur liability,” a mens rea requirement “provide[s] ‘breathing room’ for more valuable speech.” Alvarez, 567 U. S., at 733 (Breyer, J., concurring in judgment).
That kind of “strategic protection” features in our precedent concerning the most prominent categories of historically unprotected speech. Gertz, 418 U. S., at 342. Defamation is the best known and best theorized example. False and defamatory statements of fact, we have held, have “no constitutional value.” Id., at 340; see Alvarez, 567 U. S., at 718–719 (plurality opinion). Yet a public figure cannot recover for the injury such a statement causes unless the speaker acted with “knowledge that it was false or with reckless disregard of whether it was false or not.” New York Times Co. v. Sullivan, 376 U. S. 254, 280 (1964); see Garrison v. Louisiana, 379 U. S. 64, 74 (1964) (using the same standard for criminal libel). That rule is based on fear of “self-censorship”—the worry that without such a subjective mental-state requirement, the uncertainties and expense of litigation will deter speakers from making even truthful statements. Sullivan, 376 U. S., at 279. The First Amendment, we have concluded, “requires that we protect some falsehood in order to protect speech that matters.” Gertz, 418 U. S., at 341.
The same idea arises in the law respecting obscenity and incitement to unlawful conduct. Like threats, incitement inheres in particular words used in particular contexts: Its harm can arise even when a clueless speaker fails to grasp his expression’s nature and consequence. But still, the First Amendment precludes punishment, whether civil or criminal, unless the speaker’s words were “intended” (not just likely) to produce imminent disorder. Hess v. Indiana, 414 U. S. 105, 109 (1973) (per curiam); see Brandenburg, 395 U. S., at 447; NAACP v. Claiborne Hardware Co., 458 U. S. 886, 927–929 (1982). That rule helps prevent a law from deterring “mere advocacy” of illegal acts—a kind of speech falling within the First Amendment’s core. Brandenburg, 395 U. S., at 449. And for a similar reason, the First Amendment demands proof of a defendant’s mindset to make out an obscenity case. Obscenity is obscenity, whatever the purveyor’s mental state. But we have repeatedly recognized that punishment depends on a “vital element of scienter”—often described as the defendant’s awareness of “the character and nature” of the materials he distributed. Hamling v. United States, 418 U. S. 87, 122–123 (1974); see Elonis, 575 U. S., at 739 (reiterating Hamling). The rationale should by now be familiar. Yes, “obscene speech and writings are not protected.” Smith v. California, 361 U. S. 147, 152 (1959). But punishing their distribution without regard to scienter would “have the collateral effect of inhibiting” protected expression. Id., at 151. Given “the ambiguities inherent in the definition of obscenity,” the First Amendment “requires proof of scienter to avoid the hazard of self-censorship.” Mishkin, 383 U. S., at 511.[4]
The same reasoning counsels in favor of requiring a subjective element in a true-threats case. This Court again must consider the prospect of chilling non-threatening expression, given the ordinary citizen’s predictable tendency to steer “wide[] of the unlawful zone.” Speiser, 357 U. S., at 526. The speaker’s fear of mistaking whether a statement is a threat; his fear of the legal system getting that judgment wrong; his fear, in any event, of incurring legal costs—all those may lead him to swallow words that are in fact not true threats. Some 50 years ago, Justice Marshall made the point when reviewing a true-threats prosecution arguably involving only political hyperbole. See Rogers v. United States, 422 U. S. 35 (1975). The Court in Rogers reversed the conviction on other grounds, but Justice Marshall focused on the danger of deterring non-threatening speech. An objective standard, turning only on how reasonable observers would construe a statement in context, would make people give threats “a wide berth.” Id., at 47 (concurring opinion). And so use of that standard would discourage the “uninhibited, robust, and wide-open debate that the First Amendment is intended to protect.” Id., at 48 (quoting Sullivan, 376 U. S., at 270).
The reasoning—and indeed some of the words—came straight from this Court’s decisions insisting on a subjective element in other unprotected-speech cases, whether involving defamation, incitement, or obscenity. No doubt, the approach in all of those cases has a cost: Even as it lessens chill of protected speech, it makes prosecution of otherwise proscribable, and often dangerous, communications harder. And the balance between those two effects may play out differently in different contexts, as the next part of this opinion discusses. But the ban on an objective standard remains the same, lest true-threats prosecutions chill too much protected, non-threatening expression.
B
The next question concerns the type of subjective standard the First Amendment requires. The law of mens rea offers three basic choices. Purpose is the most culpable level in the standard mental-state hierarchy, and the hardest to prove. A person acts purposefully when he “consciously desires” a result—so here, when he wants his words to be received as threats. United States v. Bailey, 444 U. S. 394, 404 (1980). Next down, though not often distinguished from purpose, is knowledge. Ibid. A person acts knowingly when “he is aware that [a] result is practically certain to follow”—so here, when he knows to a practical certainty that others will take his words as threats. Ibid. (internal quotation marks omitted). A greater gap separates those two from recklessness. A person acts recklessly, in the most common formulation, when he “consciously disregard[s] a substantial [and unjustifiable] risk that the conduct will cause harm to another.” Voisine v. United States, 579 U. S. 686, 691 (2016) (internal quotation marks omitted). That standard involves insufficient concern with risk, rather than awareness of impending harm. See Borden v. United States, 593 U. S. ___, ___ (2021) (plurality opinion) (slip op., at 5). But still, recklessness is morally culpable conduct, involving a “deliberate decision to endanger another.” Voisine, 579 U. S., at 694. In the threats context, it means that a speaker is aware “that others could regard his statements as” threatening violence and “delivers them anyway.” Elonis, 575 U. S., at 746 (Alito, J., concurring in part and dissenting in part).[5]
Among those standards, recklessness offers the right path forward. We have so far mostly focused on the constitutional interest in free expression, and on the correlative need to take into account threat prosecutions’ chilling effects. But the precedent we have relied on has always recognized—and insisted on “accommodat[ing]”—the “competing value[]” in regulating historically unprotected expression. Gertz, 418 U. S., at 348. Here, as we have noted, that value lies in protecting against the profound harms, to both individuals and society, that attend true threats of violence—as evidenced in this case. See supra, at 2, 6. The injury associated with those statements caused history long ago to place them outside the First Amendment’s bounds. When despite that judgment we require use of a subjective mental-state standard, we necessarily impede some true-threat prosecutions. And as we go up the subjective mens rea ladder, that imposition on States’ capacity to counter true threats becomes still greater—and, presumably, with diminishing returns for protected expression. In advancing past recklessness, we make it harder for a State to substantiate the needed inferences about mens rea (absent, as is usual, direct evidence). And of particular importance, we prevent States from convicting morally culpable defendants. See Elonis, 575 U. S., at 745 (opinion of Alito, J.). For reckless defendants have done more than make a bad mistake. They have consciously accepted a substantial risk of inflicting serious harm.
Using a recklessness standard also fits with the analysis in our defamation decisions. As noted earlier, the Court there adopted a recklessness rule, applicable in both civil and criminal contexts, as a way of accommodating competing interests. See supra, at 7–8. In the more than half-century in which that standard has governed, few have suggested that it needs to be higher—in other words, that still more First Amendment “breathing space” is required. Gertz, 418 U. S., at 342. And we see no reason to offer greater insulation to threats than to defamation. See Elonis, 575 U. S., at 748 (opinion of Alito, J.). The societal interests in countering the former are at least as high. And the protected speech near the borderline of true threats (even though sometimes political, as in Rogers) is, if anything, further from the First Amendment’s central concerns than the chilled speech in Sullivan-type cases (i.e., truthful reputation-damaging statements about public officials and figures).
It is true that our incitement decisions demand more—but the reason for that demand is not present here. When incitement is at issue, we have spoken in terms of specific intent, presumably equivalent to purpose or knowledge. See Hess, 414 U. S., at 109; supra, at 8. In doing so, we recognized that incitement to disorder is commonly a hair’s-breadth away from political “advocacy”—and particularly from strong protests against the government and prevailing social order. Brandenburg, 395 U. S., at 447. Such protests gave rise to all the cases in which the Court demanded a showing of intent. See ibid.; Hess, 414 U. S., at 106; Claiborne Hardware Co., 458 U. S., at 888, 928. And the Court decided those cases against a resonant historical backdrop: the Court’s failure, in an earlier era, to protect mere advocacy of force or lawbreaking from legal sanction. See, e.g., Whitney v. California, 274 U. S. 357 (1927); Gitlow v. New York, 268 U. S. 652 (1925); Abrams v. United States, 250 U. S. 616 (1919). A strong intent requirement was, and remains, one way to guarantee history was not repeated. It was a way to ensure that efforts to prosecute incitement would not bleed over, either directly or through a chilling effect, to dissenting political speech at the First Amendment’s core. But the potency of that protection is not needed here. For the most part, the speech on the other side of the true-threats boundary line—as compared with the advocacy addressed in our incitement decisions—is neither so central to the theory of the First Amendment nor so vulnerable to government prosecutions. It is not just that our incitement decisions are distinguishable; it is more that they compel the use of a distinct standard here.[6]
That standard, again, is recklessness. It offers “enough ‘breathing space’ for protected speech,” without sacrificing too many of the benefits of enforcing laws against true threats. Elonis, 575 U. S., at 748 (opinion of Alito, J.). As with any balance, something is lost on both sides: The rule we adopt today is neither the most speech-protective nor the most sensitive to the dangers of true threats. But in declining one of those two alternative paths, something more important is gained: Not “having it all”—because that is impossible—but having much of what is important on both sides of the scale.[7]
III
It is time to return to Counterman’s case, though only a few remarks are necessary. Counterman, as described above, was prosecuted in accordance with an objective standard. See supra, at 3. The State had to show only that a reasonable person would understand his statements as threats. It did not have to show any awareness on his part that the statements could be understood that way. For the reasons stated, that is a violation of the First Amendment.
We accordingly vacate the judgment of the Colorado Court of Appeals and remand the case for further proceedings not inconsistent with this opinion.
- ↑ The statute Counterman was charged with violating is titled a “stalking” statute and also prohibits “[r]epeatedly follow[ing], approach[ing], contact[ing], [or] plac[ing] under surveillance” another person. §18–3–602(1)(c). But the State had no evidence, beyond what Counterman claimed, that he actually had followed or surveilled C. W. For example, C. W. had never noticed anything of that kind. So the prosecution based its case solely on Counterman’s “[r]epeated[] … communication[s]” with C. W. Ibid.
- ↑ A preliminary clarification may be useful, concerning the difference between awareness of a communication’s contents and awareness of its threatening nature. Everyone agrees, again, that the State must prove the former—and Colorado law appears to hold as much. See Colo. Rev. Stat. §18–3–602(1)(c); Brief for Respondent 18. So, for example, if a defendant delivers a sealed envelope without knowing that a threatening letter is inside, he cannot be liable for the communication. So too (though this common example seems fairly preposterous) if a “foreigner, ignorant of the English language, who would not know the meaning of the words,” somehow manages to convey an English-language threat. Elonis v. United States, 575 U. S. 723, 738 (2015) (internal quotation marks omitted). The question in this case arises when the defendant (unlike in those hypotheticals) understands the content of the words, but may not grasp that others would find them threatening. Must he do so, under the First Amendment, for a true-threats prosecution to succeed?
- ↑ The concurrence relies on Virginia v. Black, 538 U. S. 343 (2003), to argue that the category of true threats itself incorporates a mens rea element. See post, at 9–11, 14 (Sotomayor, J., concurring in part and concurring in judgment). But that claim is based on a misreading. The statements the concurrence quotes merely reflect that the statute involved in the case required a showing of intent. Black did not address whether the First Amendment demands such a showing, or why it might do so. See United States v. Jeffries, 692 F. 3d 473, 479–480 (CA6 2012) (Sutton, J.); see also post, at 9–10, and n. 4 (Barrett, J., dissenting) (explaining that Black concerned a different part of the statute, preventing consideration of contextual factors in assessing whether a statement was a threat).
- ↑ The dissent, in urging an objective standard here, reads the obscenity decisions as requiring merely that the defendant know “what the material depicts” (as a speaker must know a communication’s contents). Post, at 5–6 (opinion of Barrett, J.) (relying on Hamling, 418 U. S., at 120–123). But see the statements quoted above: That is not what they say. And indeed, this Court recently rejected the dissent’s revisionist reading, explaining in detail—and in response to a near-identical argument—that the obscenity decisions demand awareness of “the character of [the materials,] not simply [their] contents.” Elonis, 575 U. S., at 739–740 (discussing Hamling, 418 U. S., at 120–123, and Mishkin, 383 U. S., at 510).
The dissent’s use of two other First Amendment categories—fighting words and false commercial speech—to support an objective test also falls flat. See post, at 3–4 (opinion of Barrett, J.). This Court has not upheld a conviction under the fighting-words doctrine in 80 years. At the least, that doctrine is today a poor candidate for spinning off other First Amendment rules. False commercial speech is also a poor analog, though for different reasons. Put aside that the line of cases the dissent invokes has never been listed among the historically unprotected categories of speech. See, e.g., United States v. Stevens, 559 U. S. 460, 468 (2010); see supra, at 5. Yet more relevant, the Court has often noted that commercial speech is less vulnerable to chill than most other speech is. See, e.g., Board of Trustees of State Univ. of N. Y. v. Fox, 492 U. S. 469, 481 (1989). And it is the fear of chill that has led to state-of-mind requirements in the context of unprotected speech.
- ↑ Just to complete the mens rea hierarchy, the last level is negligence—but that is an objective standard, of the kind we have just rejected. A person acts negligently if he is not but should be aware of a substantial risk—here, that others will understand his words as threats. See Borden, 593 U. S., at ___ (plurality opinion) (slip op., at 5). That makes liability depend not on what the speaker thinks, but instead on what a reasonable person would think about whether his statements are threatening in nature. See Elonis, 575 U. S., at 738 (“Having liability turn on whether a reasonable person regards the communication as a threat—regardless of what the defendant thinks—reduces culpability … to negligence” (internal quotation marks omitted)).
- ↑ Our obscenity decisions are of no help in this inquiry, because the Court has never determined the precise mens rea needed to impose punishment. In arguing to the contrary, the concurrence relies mainly on Hamling. Post, at 18–19 (opinion of Sotomayor, J.). But if the dissent is wrong in saying that Hamling (and other obscenity decisions) allowed an objective inquiry, see supra, at 9, n. 4, the concurrence is wrong in suggesting that it required use of a purpose or knowledge standard. As to the concurrence’s claim, Hamling held only that a statute with that standard was “constitutionally sufficient.” 418 U. S., at 123. The decision said nothing about whether it was constitutionally necessary, or instead whether a recklessness standard would suffice as well.
- ↑ The dissent accuses the Court of making a “Goldilocks judgment” in favoring a recklessness standard. Post, at 13 (opinion of Barrett, J.). But in law, as in life, there are worse things than being “just right.”