ELONIS v. UNITED STATES

SUPREME COURT OF THE UNITED STATES

_________________

No. 13–983

_________________

ANTHONY DOUGLAS ELONIS, PETITIONER v. UNITED STATES

on writ of certiorari to the united states court of appeals for the third circuit

 Justice Thomas, dissenting.

 We granted certiorari to resolve a conflict in the lower courts over the appropriate mental state for threat prosecutions under 18 U. S. C. §875(c). Save two, every Circuit to have considered the issue—11 in total—has held that this provision demands proof only of general intent, which here requires no more than that a defendant knew he transmitted a communication, knew the words used in that communication, and understood the ordinary meaning of those words in the relevant context. The outliers are the Ninth and Tenth Circuits, which have concluded that proof of an intent to threaten was necessary for conviction. Adopting the minority position, Elonis urges us to hold that §875(c) and the First Amendment require proof of an intent to threaten. The Government in turn advocates a general-intent approach.

 Rather than resolve the conflict, the Court casts aside the approach used in nine Circuits and leaves nothing in its place. Lower courts are thus left to guess at the appropriate mental state for §875(c). All they know after today’s decision is that a requirement of general intent will not do. But they can safely infer that a majority of this Court would not adopt an intent-to-threaten requirement, as the opinion carefully leaves open the possibility that recklessness may be enough. See ante, at 16–17.

  This failure to decide throws everyone from appellate judges to everyday Facebook users into a state of uncertainty. This uncertainty could have been avoided had we simply adhered to the background rule of the common law favoring general intent. Although I am sympathetic to my colleagues’ policy concerns about the risks associated with threat prosecutions, the answer to such fears is not to discard our traditional approach to state-of-mind requirements in criminal law. Because the Court of Appeals properly applied the general-intent standard, and because the communications transmitted by Elonis were “true threats” unprotected by the First Amendment, I would affirm the judgment below.

I

A

 Enacted in 1939, §875(c) provides, “Whoever transmits in interstate or foreign commerce any communication containing any threat to kidnap any person or any threat to injure the person of another, shall be fined under this title or imprisoned not more than five years, or both.” Because §875(c) criminalizes speech, the First Amendment requires that the term “threat” be limited to a narrow class of historically unprotected communications called “true threats.” To qualify as a true threat, a communication must be a serious expression of an intention to commit unlawful physical violence, not merely “political hyperbole”; “vehement, caustic, and sometimes unpleasantly sharp attacks”; or “vituperative, abusive, and inexact” statements. Watts v. United States, 394 U. S. 705, 708 (1969) (per curiam) (internal quotation marks omitted). It also cannot be determined solely by the reaction of the recipient, but must instead be “determined by the interpretation of a reasonable recipient familiar with the context of the communication,” United States v. Darby, 37 F. 3d 1059, 1066 (CA4 1994) (emphasis added), lest histor ically protected speech be suppressed at the will of an eggshell observer, cf. Cox v. Louisiana, 379 U. S. 536, 551 (1965) (“[C]onstitutional rights may not be denied simply because of hostility to their assertion or exercise” (internal quotation marks omitted)). There is thus no dispute that, at a minimum, §875(c) requires an objective showing: The communication must be one that “a reasonable observer would construe as a true threat to another.” United States v. Jeffries, 692 F. 3d 473, 478 (CA6 2012). And there is no dispute that the posts at issue here meet that objective standard.

 The only dispute in this case is about the state of mind necessary to convict Elonis for making those posts. On its face, §875(c) does not demand any particular mental state. As the Court correctly explains, the word “threat” does not itself contain a mens rea requirement. See ante, at 8–9. But because we read criminal statutes “in light of the background rules of the common law, in which the requirement of some mens rea for a crime is firmly embedded,” we require “some indication of congressional intent, express or implied, . . . to dispense with mens rea as an element of a crime.” Staples v. United States, 511 U. S. 600, 605–606 (1994) (citation omitted). Absent such indicia, we ordinarily apply the “presumption in favor of scienter” to require only “proof of general intent—that is, that the defendant [must] posses[s] knowledge with respect to the actus reus of the crime.” Carter v. United States, 530 U. S. 255, 268 (2000).

 Under this “conventional mens rea element,” “the defendant [must] know the facts that make his conduct illegal,” Staples, supra, at 605, but he need not know that those facts make his conduct illegal. It has long been settled that “the knowledge requisite to knowing violation of a statute is factual knowledge as distinguished from knowledge of the law.” Bryan v. United States, 524 U. S. 184, 192 (1998) (internal quotation marks omitted). For  instance, in Posters ‘N’ Things, Ltd. v. United States, 511 U. S. 513 (1994), the Court addressed a conviction for selling drug paraphernalia under a statute forbidding anyone to “ ‘make use of the services of the Postal Service or other interstate conveyance as part of a scheme to sell drug paraphernalia,’ ” id., at 516 (quoting 21 U. S. C. §857(a)(1) (1988 ed.)). In applying the presumption in favor of scienter, the Court concluded that “although the Government must establish that the defendant knew that the items at issue are likely to be used with illegal drugs, it need not prove specific knowledge that the items are ‘drug paraphernalia’ within the meaning of the statute.” 511 U. S., at 524.

 Our default rule in favor of general intent applies with full force to criminal statutes addressing speech. Well over 100 years ago, this Court considered a conviction under a federal obscenity statute that punished anyone “ ‘who shall knowingly deposit, or cause to be deposited, for mailing or delivery,’ ” any “ ‘obscene, lewd, or lascivious book, pamphlet, picture, paper, writing, print, or other publication of an indecent character.’ ” Rosen v. United States, 161 U. S. 29, 30 (1896) (quoting Rev. Stat. §3893). In that case, as here, the defendant argued that, even if “he may have had . . . actual knowledge or notice of [the paper’s] contents” when he put it in the mail, he could not “be convicted of the offence . . . unless he knew or believed that such paper could be properly or justly characterized as obscene, lewd, and lascivious.” 161 U. S., at 41. The Court rejected that theory, concluding that if the material was actually obscene and “deposited in the mail by one who knew or had notice at the time of its contents, the offence is complete, although the defendant himself did not regard the paper as one that the statute forbade to be carried in the mails.” Ibid. As the Court explained, “Congress did not intend that the question as to the character of the paper should depend upon the opinion or belief of  the person who, with knowledge or notice of [the paper’s] contents, assumed the responsibility of putting it in the mails of the United States,” because “[e]very one who uses the mails of the United States for carrying papers or publications must take notice of . . . what must be deemed obscene, lewd, and lascivious.” Id., at 41–42.

 This Court reaffirmed Rosen’s holding in Hamling v. United States, 418 U. S. 87 (1974), when it considered a challenge to convictions under the successor federal statute, see id., at 98, n. 8 (citing 18 U. S. C. §1461 (1970 ed.)). Relying on Rosen, the Court rejected the argument that the statute required “proof both of knowledge of the contents of the material and awareness of the obscene character of the material.” 418 U. S., at 120 (internal quotation marks omitted). In approving the jury instruction that the defendants’ “belief as to the obscenity or non-obscenity of the material is irrelevant,” the Court declined to hold “that the prosecution must prove a defendant’s knowledge of the legal status of the materials he distributes.” Id., at 120–121 (internal quotation marks omitted). To rule otherwise, the Court observed, “would permit the defendant to avoid prosecution by simply claiming that he had not brushed up on the law.” Id., at 123.

 Decades before §875(c)’s enactment, courts took the same approach to the first federal threat statute, which prohibited threats against the President. In 1917, Congress enacted a law punishing anyone

“who knowingly and willfully deposits or causes to be deposited for conveyance in the mail . . . any letter, paper, writing, print, missive, or document containing any threat to take the life of or to inflict bodily harm upon the President of the United States, or who knowingly and willfully otherwise makes any such threat against the President.” Act of Feb. 14, 1917, ch. 64, 39 Stat. 919.

 Courts applying this statute shortly after its enactment appeared to require proof of only general intent. In Ragansky v. United States, 253 F. 643 (CA7 1918), for instance, a Court of Appeals held that “[a] threat is knowingly made, if the maker of it comprehends the meaning of the words uttered by him,” and “is willfully made, if in addition to comprehending the meaning of his words, the maker voluntarily and intentionally utters them as the declaration of an apparent determination to carry them into execution,” id., at 645. The court consequently rejected the defendant’s argument that he could not be convicted  when his language “[c]oncededly . . . constituted such a threat” but was meant only “as a joke.” Id., at 644. Likewise, in United States v. Stobo, 251 F. 689 (Del. 1918), a District Court rejected the defendant’s objection that there was no allegation “of any facts . . . indicating any intention . . . on the part of the defendant . . . to menace the President of the United States,” id., at 693 (internal quotation marks omitted). As it explained, the defendant “is punishable under the act whether he uses the words lightly or with a set purpose to kill,” as “[t]he effect upon the minds of the hearers, who cannot read his inward thoughts, is precisely the same.” Ibid. At a minimum, there is no historical practice requiring more than general intent when a statute regulates speech.

B

 Applying ordinary rules of statutory construction, I would read §875(c) to require proof of general intent. To “know the facts that make his conduct illegal” under §875(c), see Staples, 511 U. S., at 605, a defendant must know that he transmitted a communication in interstate or foreign commerce that contained a threat. Knowing that the communication contains a “threat”—a serious expression of an intention to engage in unlawful physical violence—does not, however, require knowing that a jury  will conclude that the communication contains a threat as a matter of law. Instead, like one who mails an “obscene” publication and is prosecuted under the federal obscenity statute, a defendant prosecuted under §875(c) must know only the words used in that communication, along with their ordinary meaning in context.

 General intent divides those who know the facts constituting the actus reus of this crime from those who do not. For example, someone who transmits a threat who does not know English—or who knows English, but perhaps does not know a threatening idiom—lacks the general intent required under §875(c). See Ragansky, supra, at 645 (“[A] foreigner, ignorant of the English language, repeating [threatening] words without knowledge of their meaning, may not knowingly have made a threat”). Likewise, the hapless mailman who delivers a threatening letter, ignorant of its contents, should not fear prosecution. A defendant like Elonis, however, who admits that he “knew that what [he] was saying was violent” but supposedly “just wanted to express [him]self,” App. 205, acted with the general intent required under §875(c), even if he did not know that a jury would conclude that his communication constituted a “threat” as a matter of law.

 Demanding evidence only of general intent also corresponds to §875(c)’s statutory backdrop. As previously discussed, before the enactment of §875(c), courts had read the Presidential threats statute to require proof only of general intent. Given Congress’ presumptive awareness of this application of the Presidential threats statute—not to mention this Court’s similar approach in the obscenity context, see Rosen, 161 U. S., at 41–42—it is difficult to conclude that the Congress that enacted §875(c) in 1939 understood it to contain an implicit mental-state requirement apart from general intent. There is certainly no textual evidence to support this conclusion. If anything, the text supports the opposite inference, as §875(c), unlike  the Presidential threats statute, contains no reference to knowledge or willfulness. Nothing in the statute suggests that Congress departed from the “conventional mens rea element” of general intent, Staples, supra, at 605; I would not impose a higher mental-state requirement here.

C

 The majority refuses to apply these ordinary background principles. Instead, it casts my application of general intent as a negligence standard disfavored in the criminal law. Ante, at 13–16. But that characterization misses the mark. Requiring general intent in this context is not the same as requiring mere negligence. Like the mental-state requirements adopted in many of the cases cited by the Court, general intent under §875(c) prevents a defendant from being convicted on the basis of any fact beyond his awareness. See, e.g., United States v. X-Citement Video, Inc., 513 U. S. 64, 73 (1994) (knowledge of age of persons depicted in explicit materials); Staples, supra, at 614–615 (knowledge of firing capability of  weapon); Morissette v. United States, 342 U. S. 246, 270–271 (1952) (knowledge that property belonged to another). In other words, the defendant must know—not merely be reckless or negligent with respect to the fact—that he is committing the acts that constitute the actus reus of the offense.

 But general intent requires no mental state (not even a negligent one) concerning the “fact” that certain words meet the legal definition of a threat. That approach is particularly appropriate where, as here, that legal status is determined by a jury’s application of the legal standard of a “threat” to the contents of a communication. And convicting a defendant despite his ignorance of the legal—or objective—status of his conduct does not mean that he is being punished for negligent conduct. By way of example, a defendant who is convicted of murder despite claim ing that he acted in self-defense has not been penalized under a negligence standard merely because he does not know that the jury will reject his argument that his “belief in the necessity of using force to prevent harm to himself [was] a reasonable one.” See 2 W. LaFave, Substantive Criminal Law §10.4(c), p. 147 (2d ed. 2003).

 The Court apparently does not believe that our traditional approach to the federal obscenity statute involved a negligence standard. It asserts that Hamling “approved a state court’s conclusion that requiring a defendant to know the character of the material incorporated a ‘vital element of scienter’ so that ‘not innocent but calculated purveyance of filth . . . is exorcised.’ ” Ante, at 15 (quoting Hamling, 418 U. S., at 122 (in turn quoting Mishkin v. New York, 383 U. S. 502, 510 (1966)). According to the Court, the mental state approved in Hamling thus “turns on whether a defendant knew the character of what was sent, not simply its contents and context.” Ante, at 15. It is unclear what the Court means by its distinction between “character” and “contents and context.” “Character” cannot mean legal obscenity, as Hamling rejected the argument that a defendant must have “awareness of the obscene character of the material.” 418 U. S., at 120 (internal quotation marks omitted). Moreover, this discussion was not part of Hamling’s holding, which was primarily a reaffirmation of Rosen. See 418 U. S., at 120–121; see also Posters ’N’ Things, 511 U. S., at 524–525 (characterizing Hamling as holding that a “statute prohibiting mailing of obscene materials does not require proof that [the] defendant knew the materials at issue met the legal definition of ‘obscenity’ ”).

 The majority’s treatment of Rosen is even less persuasive. To shore up its position, it asserts that the critical portion of Rosen rejected an “ ‘ignorance of the law’ defense,” and claims that “no such contention is at issue here.” Ante, at 15. But the thrust of Elonis’ challenge is  that a §875(c) conviction cannot stand if the defendant’s subjective belief of what constitutes a “threat” differs from that of a reasonable jury. That is akin to the argument the defendant made—and lost—in Rosen. That defendant insisted that he could not be convicted for mailing the paper “unless he knew or believed that such paper could be properly or justly characterized as obscene.” 161 U. S., at 41. The Court, however, held that the Government did not need to show that the defendant “regard[ed] the paper as one that the statute forbade to be carried in the mails,” because the obscene character of the material did not “depend upon the opinion or belief of the person who . . . assumed the responsibility of putting it in the mails.” Ibid. The majority’s muddying of the waters cannot obscure the fact that today’s decision is irreconcilable with Rosen and Hamling.

D

 The majority today at least refrains from requiring an intent to threaten for §875(c) convictions, as Elonis asks us to do. Elonis contends that proof of a defendant’s intent to put the recipient of a threat in fear is necessary for conviction, but that element cannot be found within the statutory text. “[W]e ordinarily resist reading words or elements into a statute that do not appear on its face,” including elements similar to the one Elonis proposes. E.g., Bates v. United States, 522 U. S. 23, 29 (1997) (declining to read an “intent to defraud” element into a criminal statute). As the majority correctly explains, nothing in the text of §875(c) itself requires proof of an intent to threaten. See ante, at 8–9. The absence of such a requirement is significant, as Congress knows how to require a heightened mens rea in the context of threat offenses. See §875(b) (providing for the punishment of “[w]hoever, with intent to extort . . . , transmits in interstate or foreign commerce any communication containing  any threat to kidnap any person or any threat to injure the person of another”); see also §119 (providing for the punishment of “[w]hoever knowingly makes restricted personal information about [certain officials] . . . publicly available . . . with the intent to threaten”).

 Elonis nonetheless suggests that an intent-to-threaten element is necessary in order to avoid the risk of punishing innocent conduct. But there is nothing absurd about punishing an individual who, with knowledge of the words he uses and their ordinary meaning in context, makes a threat. For instance, a high-school student who sends a letter to his principal stating that he will massacre his classmates with a machine gun, even if he intended the letter as a joke, cannot fairly be described as engaging in innocent conduct. But see ante, at 4–5, 16 (concluding that Elonis’ conviction under §875(c) for discussing a plan to “ ‘initiate the most heinous school shooting ever imagined’ ” against “ ‘a Kindergarten class’ ” cannot stand without proof of some unspecified heightened mental state).

 Elonis also insists that we read an intent-to-threaten element into §875(c) in light of the First Amendment. But our practice of construing statutes “to avoid constitutional questions . . . is not a license for the judiciary to rewrite language enacted by the legislature,” Salinas v. United States, 522 U. S. 52, 59–60 (1997) (internal quotation marks omitted), and ordinary background principles of criminal law do not support rewriting §875(c) to include an intent-to-threaten requirement. We have not altered our traditional approach to mens rea for other constitutional provisions. See, e.g., Dean v. United States, 556 U. S. 568, 572–574 (2009) (refusing to read an intent-to-discharge-the-firearm element into a mandatory minimum provision concerning the discharge of a firearm during a particular crime). The First Amendment should be treated no  differently.

 II

 In light of my conclusion that Elonis was properly convicted under the requirements of §875(c), I must address his argument that his threatening posts were nevertheless protected by the First Amendment.

A

 Elonis does not contend that threats are constitutionally protected speech, nor could he: “From 1791 to the present, . . . our society . . . has permitted restrictions upon the content of speech in a few limited areas,” true threats being one of them. R. A. V. v. St. Paul, 505 U. S. 377, 382–383 (1992); see id., at 388. Instead, Elonis claims that only intentional threats fall within this particular historical exception.

 If it were clear that intentional threats alone have been punished in our Nation since 1791, I would be inclined to agree. But that is the not the case. Although the Federal Government apparently did not get into the business of regulating threats until 1917, the States have been doing so since the late 18th and early 19th centuries. See, e.g., 1795 N. J. Laws p. 108; Ill. Rev. Code of Laws, Crim. Code §108 (1827) (1827 Ill. Crim. Code); 1832 Fla. Laws pp. 68–69. And that practice continued even after the States amended their constitutions to include speech protections similar to those in the First Amendment. See, e.g., Fla. Const., Art. I, §5 (1838); Ill. Const., Art. VIII, §22 (1818), Mich. Const., Art. I, §7 (1835); N. J. Const., Art. I, §5 (1844); J. Hood, Index of Colonial and State Laws of New Jersey 1203, 1235, 1257, 1265 (1905); 1 Ill. Stat., ch. 30, div. 9, §31 (3d ed. 1873). State practice thus provides at least some evidence of the original meaning of the phrase “freedom of speech” in the First Amendment. See Roth v. United States, 354 U. S. 476, 481–483 (1957) (engaging in a similar inquiry with respect to obscenity).

 Shortly after the founding, several States and Territo ries enacted laws making it a crime to “knowingly send or deliver any letter or writing, with or without a name subscribed thereto, or signed with a fictitious name, . . . threatening to maim, wound, kill or murder any person, or to burn his or her [property], though no money, goods or chattels, or other valuable thing shall be demanded,” e.g., 1795 N. J. Laws §57, at 108; see also, e.g., 1816 Ga. Laws p. 178; 1816 Mich. Territory Laws p. 128; 1827 Ill. Crim. Code §108; 1832 Fla. Laws, at 68–69. These laws appear to be the closest early analogue to §875(c), as they penalize transmitting a communication containing a threat without proof of a demand to extort something from the victim. Threat provisions explicitly requiring proof of a specific “intent to extort” appeared alongside these laws, see,  e.g., 1795 N. J. Laws §57, at 108, but those provisions  are simply the predecessors to §875(b) and §875(d),  which likewise expressly contain an intent-to-extort  requirement.

 The laws without that extortion requirement were copies of a 1754 English threat statute subject to only a general-intent requirement. The statute made it a capital offense to “knowingly send any Letter without any Name subscribed thereto, or signed with a fictitious Name . . . threatening to kill or murder any of his Majesty’s Subject or Subjects, or to burn their [property], though no Money or Venison or other valuable Thing shall be demanded.” 27 Geo. II, ch. 15, in 7 Eng. Stat. at Large 61 (1754); see also 4 W. Blackstone, Commentaries on the Laws of England 144 (1768) (describing this statute). Early English decisions applying this threat statute indicated that the appropriate mental state was general intent. In King v. Girdwood, 1 Leach 142, 168 Eng. Rep. 173 (K. B. 1776), for example, the trial court instructed the jurors that, “if they were of opinion that” the “terms of the letter conveyed an actual threat to kill or murder,” “and that the prisoner knew the contents of it, they ought to find him  guilty; but that if they thought he did not know the contents, or that the words might import any thing less than to kill or murder, they ought to acquit,” id., at 143, 168 Eng. Rep., at 173. On appeal following conviction, the judges “thought that the case had been properly left to the Jury.” Ibid., 168 Eng. Rep., at 174. Other cases likewise appeared to consider only the import of the letter’s language, not the intent of its sender. See, e.g., Rex v. Boucher, 4 Car. & P. 562, 563, 172 Eng. Rep. 826, 827 (K. B. 1831) (concluding that an indictment was sufficient because “th[e] letter very plainly conveys a threat to kill and murder” and “[n]o one who received it could have any doubt as to what the writer meant to threaten”); see also 2 E. East, A Treatise of the Pleas of the Crown 1116 (1806) (discussing Jepson and Springett’s Case, in which the judges disagreed over whether “the letter must be understood as . . . importing a threat” and whether that was “a necessary construction”).

 Unsurprisingly, these early English cases were well known in the legal world of the 19th century United States. For instance, Nathan Dane’s A General Abridgement of American Law—“a necessary adjunct to the library of every American lawyer of distinction,” 1 C. Warren, History of the Harvard Law School and of Early Legal Conditions in America 414 (1908)—discussed the English threat statute and summarized decisions such as Girdwood. 7 N. Dane, A General Abridgement of American Law 31–32 (1824). And as this Court long ago recognized, “It is doubtless true . . . that where English statutes . . . have been adopted into our own legislation; the known and settled construction of those statutes by courts of law, has been considered as silently incorporated into the acts, or has been received with all the weight of authority.” Pennock v. Dialogue, 2 Pet. 1, 18 (1829); see also, e.g., Commonwealth v. Burdick, 2 Pa. 163, 164 (1846) (considering English cases persuasive authority in interpreting similar  state statute creating the offense of obtaining property through false pretenses). In short, there is good reason to believe that States bound by their own Constitutions to protect freedom of speech long ago enacted general-intent threat statutes.

 Elonis disputes this historical analysis on two grounds, but neither is persuasive. He first points to a treatise stating that the 1754 English statute was “levelled against such whose intention it was, (by writing such letters, either without names or in fictitious names,) to conceal themselves from the knowledge of the party threatened, that they might obtain their object by creating terror in [the victim’s] mind.” 2 W. Russell & D. Davis, A Treatise on Crimes & Misdemeanors 1845 (1st Am. ed. 1824). But the fact that the ordinary prosecution under this provision involved a defendant who intended to cause fear does not mean that such a mental state was required as a matter of law. After all, §875(c) is frequently deployed against people who wanted to cause their victims fear, but that fact does not answer the legal question presented in this case. See, e.g., United States v. Sutcliffe, 505 F. 3d 944, 952 (CA9 2007); see also Tr. of Oral Arg. 53 (counsel for the Government noting that “I think Congress would well have understood that the majority of these cases probably [involved] people who intended to threaten”).

 Elonis also cobbles together an assortment of older American authorities to prove his point, but they fail to stand up to close scrutiny. Two of his cases address the offense of breaching the peace, Ware v. Loveridge, 75 Mich. 488, 490–493, 42 N. W. 997, 998 (1889); State v. Benedict, 11 Vt. 236, 239 (1839), which is insufficiently similar to the offense criminalized in §875(c) to be of much use. Another involves a prosecution under a blackmailing statute similar to §875(b) and §875(c) in that it expressly required an “intent to extort.” Norris v. State, 95 Ind. 73, 74 (1884). And his treatises do not clearly distinguish  between the offense of making threats with the intent to extort and the offense of sending threatening letters without such a requirement in their discussions of threat statutes, making it difficult to draw strong inferences about the latter category. See 2 J. Bishop, Commentaries on the Criminal Law §1201, p. 664, and nn. 5–6 (1877); 2 J. Bishop, Commentaries on the Law of Criminal Procedure §975, p. 546 (1866); 25 The American and English Encyclopædia of Law 1073 (C. Williams ed. 1894).

 Two of Elonis’ cases appear to discuss an offense of sending a threatening letter without an intent to extort, but even these fail to make his point. One notes in passing that character evidence is admissible “to prove guilty knowledge of the defendant, when that is an essential element of the crime; that is, the quo animo, the intent or design,” and offers as an example that in the context of “sending a threatening letter, . . . prior and subsequent letters to the same person are competent in order to show the intent and meaning of the particular letter in question.” State v. Graham, 121 N. C. 623, 627, 28 S. E. 409, 409 (1897). But it is unclear from that statement whether that court thought an intent to threaten was required, especially as the case it cited for this proposition—Rex v. Boucher, 4 Car. & P. 562, 563, 172 Eng. Rep. 826, 827 (K. B. 1831)—supports a general-intent approach. The other case Elonis cites involves a statutory provision that had been judicially limited to “ ‘pertain to one or the other acts which are denounced by the statute,’ ” namely, terroristic activities carried out by the Ku Klux Klan. Commonwealth v. Morton, 140 Ky. 628, 630, 131 S. W. 506, 507 (1910) (quoting Commonwealth v. Patrick, 127 Ky. 473, 478, 105 S. W. 981, 982 (1907)). That case thus provides scant historical support for Elonis’ position.

B

 Elonis also insists that our precedents require a mental  state of intent when it comes to threat prosecutions  under §875(c), primarily relying on Watts, 394 U. S. 705, and Virginia v. Black, 538 U. S. 343 (2003). Neither of those decisions, however, addresses whether the First Amendment requires a particular mental state for threat prosecutions.

 As Elonis admits, Watts expressly declined to address the mental state required under the First Amendment for a “true threat.” See 394 U. S., at 707–708. True, the Court in Watts noted “grave doubts” about Raganksy’s construction of “willfully” in the presidential threats statute. 394 U. S., at 707–708. But “grave doubts” do not make a holding, and that stray statement in Watts is entitled to no precedential force. If anything, Watts continued the long tradition of focusing on objective criteria in evaluating the mental requirement. See ibid.

 The Court’s fractured opinion in Black likewise says little about whether an intent-to-threaten requirement is constitutionally mandated here. Black concerned a Virginia cross-burning law that expressly required “ ‘an intent to intimidate a person or group of persons,’ ” 538 U. S., at 347 (quoting Va. Code Ann. §18.2–423 (1996)), and the Court thus had no occasion to decide whether such an element was necessary in threat provisions silent on the matter. Moreover, the focus of the Black decision was on the statutory presumption that “any cross burning [w]as prima facie evidence of intent to intimidate.” 538 U. S., at 347–348. A majority of the Court concluded that this presumption failed to distinguish unprotected threats from protected speech because it might allow convictions “based solely on the fact of cross burning itself,” including cross burnings in a play or at a political rally. Id., at 365–366 (plurality opinion); id., at 386 (Souter, J., concurring in judgment in part and dissenting in part) (“The provision will thus tend to draw nonthreatening ideological expression within the ambit of the prohibition of intimidating  expression”). The objective standard for threats under §875(c), however, helps to avoid this problem by “forc[ing] jurors to examine the circumstances in which a statement is made.” Jeffries, 692 F. 3d, at 480.

 In addition to requiring a departure from our precedents, adopting Elonis’ view would make threats one of the most protected categories of unprotected speech, thereby sowing tension throughout our First Amendment doctrine. We generally have not required a heightened mental state under the First Amendment for historically unprotected categories of speech. For instance, the Court has indicated that a legislature may constitutionally prohibit “ ‘fighting words,’ those personally abusive epithets which, when addressed to the ordinary citizen, are, as a matter of common knowledge, inherently likely to provoke violent reaction,” Cohen v. California, 403 U. S. 15, 20 (1971)—without proof of an intent to provoke a violent reaction. Because the definition of “fighting words” turns on how the “ordinary citizen” would react to the language, ibid., this Court has observed that a defendant may be guilty of a breach of the peace if he “makes statements likely to provoke violence and disturbance of good order, even though no such eventuality be intended,” and that the punishment of such statements “as a criminal act would raise no question under [the Constitution],” Cantwell v. Connecticut, 310 U. S. 296, 309–310 (1940); see  also Chaplinsky v. New Hampshire, 315 U. S. 568, 572–573 (1942) (rejecting a First Amendment challenge to a general-intent construction of a state statute punishing “ ‘fighting’ words”); State v. Chaplinsky, 91 N. H. 310, 318, 18 A. 2d 754, 758 (1941) (“[T]he only intent required for conviction . . . was an intent to speak the words”). The Court has similarly held that a defendant may be convicted  of mailing obscenity under the First Amendment with- out proof that he knew the materials were legally obscene. Hamling, 418 U. S., at 120–124. And our precedents allow  liability in tort for false statements about private persons on matters of private concern even if the speaker acted negligently with respect to the falsity of those statements. See Philadelphia Newspapers, Inc. v. Hepps, 475 U. S. 767, 770, 773–775 (1986). I see no reason why we should give threats pride of place among unprotected speech.

*  *  *

 There is always a risk that a criminal threat statute may be deployed by the Government to suppress legitimate speech. But the proper response to that risk is to adhere to our traditional rule that only a narrow class of true threats, historically unprotected, may be constitutionally proscribed.

 The solution is not to abandon a mental-state requirement compelled by text, history, and precedent. Not only does such a decision warp our traditional approach to mens rea, it results in an arbitrary distinction between threats and other forms of unprotected speech. Had Elonis mailed obscene materials to his wife and a kindergarten class, he could have been prosecuted irrespective of whether he intended to offend those recipients or reck- lessly disregarded that possibility. Yet when he threatened to kill his wife and a kindergarten class, his intent to terrify those recipients (or reckless disregard of that risk) suddenly becomes highly relevant. That need not—and should not—be the case.

 Nor should it be the case that we cast aside the mental-state requirement compelled by our precedents yet offer nothing in its place. Our job is to decide questions, not create them. Given the majority’s ostensible concern for protecting innocent actors, one would have expected it to announce a clear rule—any clear rule. Its failure to do so reveals the fractured foundation upon which today’s decision rests.

 I respectfully dissent.