Searching for Truth in the First Amendment’s True Threat Doctrine

Threats of violence, even when not actually carried out, can inflict real damage. As such, state and federal laws criminalize threats in a wide range of circumstances. But threats are also speech, and free speech is broadly protected by the First Amendment. The criminalization of threats is nonetheless possible because of Supreme Court precedents denying First Amendment protection to “true threats.” Yet a crucial question remains unanswered: What counts as a true threat?

This Note examines courts’ attempts to answer this question and identifies the many ambiguities that have resulted from those attempts. In particular, this piece highlights three frontiers of judicial confusion that are likely to arise in a true threat case: (1) what type of intent the First Amendment requires, (2) the proper standard of review on appeals of true threat convictions, and (3) the contextual analyses in which courts engage to assess whether a threat is “true” (and, by extension, whether a threat conviction was constitutional). This third frontier is discussed most extensively, as it has the greatest impact on a case’s ultimate outcome. This Note also proposes a new framework for inquiries into the context of true threats, adapted from defamation law, in order to increase consistency and ensure adequate protection of speech rights within the chaotic true threat doctrine.


A threat is a grotesquely powerful weapon. Apart from any physical violence that may follow, the threat itself induces fear and can inflict genuine harm on its target. A terrorist group’s bomb threat can destabilize a community and set citizens on edge, regardless of whether an explosive detonates;1See generally Alex Schmid, Terrorism as Psychological Warfare, 1 Democracy & Sec. 137 (2005). a domestic abuser’s threat to harm his wife traumatizes her, regardless of whether he strikes her in the moment;2See TK Logan, “If I Can’t Have You Nobody Will”: Explicit Threats in the Context of Coercive Control, 32 Violence & Victims 126, 132–34 (2017); Jessica Miles, Straight Outta SCOTUS: Domestic Violence, True Threats, and Free Speech, 74 U. Mia. L. Rev. 711, 733–40 (2020). an online troll’s death threat to a reporter can prevent her from doing her job and wreak havoc on her mental health, regardless of whether the violence ever materializes.3See Philip Eil, The Worst Effects of Online Death Threats Are Things No One Can See, Vice (July 22, 2019, 7:00 AM), []; Jason Wilson, Doxxing, Assault, Death Threats: The New Dangers Facing US Journalists Covering Extremism, Guardian (June 14, 2018, 10:41 AM),‌/doxxing-assault-death-threats-the-new-dangers-facing-us-journalists-covering-extremism [].

The rapid expansion of communication technology means that threats, and all their attendant harms, can now be inflicted with a tap on a screen. In-person threats still occur, of course, but alternative methods—ranging from a threatening phone call to an especially vicious tweet—are innumerable, costless, and potentially anonymous.4See Micah Lee, How to Run a Rogue Government Twitter Account with an Anonymous Email Address and a Burner Phone, Intercept (Feb. 20, 2017, 9:53 AM), []. With this in mind, the impulse to punish individuals who threaten others is understandable, perhaps even imperative.

In the United States, the rational desire to sanction threatening language runs up against a formidable barrier: the First Amendment. The amendment protects a vast range of speech from government restriction, including speech that some may find upsetting, inaccurate, or offensive.5See Watts v. United States, 394 U.S. 705, 708 (1969); N.Y. Times Co. v. Sullivan, 376 U.S. 254, 270 (1964). “Speech” and “statements” are used loosely throughout this Note to describe many different types of expression, including written material and even expressive conduct. It is not difficult to imagine how laws restricting threats could infringe upon this “prized American privilege to speak one’s mind.”6Bridges v. California, 314 U.S. 252, 270 (1941). Speech that one person perceives as a threat might reasonably be interpreted by others as a mere expression of feeling or as a controversial idea. The Supreme Court, however, has recognized that certain types of speech remain “constitutionally proscribable.”7R.A.V. v. City of St. Paul, 505 U.S. 377, 383 (1992); see also Chaplinsky v. New Hampshire, 315 U.S. 568, 571–72 (1942). “True threats” are one such category,8Virginia v. Black, 538 U.S. 343, 359 (2003). paving the way for federal and state statutes that criminalize threatening speech.9See infra note 33. Making true threats an exception to First Amendment protection nonetheless raises the question: How do we know a true threat when we see it?

So far, the Court has failed to provide a practicable answer, leaving lower courts in the lurch. Labeling a threat as “true” is meaningless at best and misleading at worst. “True” implies some degree of actualization—that the statement must reflect a real intention to go through with the threatened act.10See True, Merriam-Webster, []. But the Supreme Court has clarified that the defendant in a true threat case does not need to have an intent to act in order to be convicted.11Black, 538 U.S. at 359–60. Instead, the government need only prove a specific “intent of placing the victim in fear of bodily harm or death.”12Id. at 360. Lower courts disagree about what exactly that intent looks like, though.13See infra Section II.A. Beyond this threshold question about intent, courts have also struggled to reach consensus on downstream issues such as the standard of review for assessing the constitutionality of a threat conviction and the test that should be used to evaluate the relevant context in true threat cases.14See infra Sections II.B, II.C. This contextual analysis is particularly difficult for courts as the inquiry is highly fact intensive, and no single approach to this crucial analytical step has emerged. Despite the uncertainty pervading every phase of a true threat case, American courts regularly convict defendants for “threatening” speech.15See, e.g., United States v. Khan, 937 F.3d 1042 (7th Cir. 2019); Commonwealth v. Knox, 190 A.3d 1146 (Pa. 2018); United States v. Elonis, 841 F.3d 589 (3d Cir. 2016); Looney v. State, 785 S.E.2d 432 (Ga. Ct. App. 2016).

The recent case of Commonwealth v. Knox demonstrates the hazards of current true threat doctrine.16See Knox, 190 A.3d at 1146. There, teenage defendant Jamal Knox wrote and recorded a rap song entitled “F–k the Police,” with violent lyrics directed at the police in general and at two Pittsburgh police officers in particular.17Id. at 1149. After the Pittsburgh police came across the song on Facebook, Knox was charged with witness intimidation and making “terroristic threats” under Pennsylvania state law.18Id. at 1150. Knox was convicted and sentenced to multiple years in prison.192 Sentenced in Threatening Rap Video Case, CBS Pittsburgh (Feb. 6, 2014, 1:06 PM), [‌/R5YZ-TQ9Q]. He appealed all the way to the Pennsylvania Supreme Court, arguing that his speech was protected under the First Amendment.20Knox, 190 A.3d at 1148. The court affirmed his conviction, finding the evidence competent to show his subjective intent to threaten the officers based on the court’s interpretation of the lyrics themselves and four “contextual factors.”21Id. at 1153, 1159–61. But breaking down the opinion reveals that the court’s analysis of context was surface level at best.22See infra Section II.D.3. Still, the U.S. Supreme Court denied Knox’s petition for writ of certiorari.23Knox v. Pennsylvania, 139 S. Ct. 1547 (2019).

This Note argues that courts’ haphazard approach to analyzing context in true threat cases risks the unconstitutional criminalization of protected speech. Part I discusses the Supreme Court’s sparse case law creating and defining the true threat doctrine. Using Knox as a representative example, Part II examines three areas of disagreement that have emerged between courts struggling to apply the Court’s inscrutable true threat holdings. Specifically, this Part emphasizes the importance of contextual factors in determining whether particular speech counts as a “true threat” and concludes that courts’ ad hoc approach to this contextual analysis does not work. To help courts more consistently and thoroughly identify true threats in a manner that aligns with First Amendment principles, Part III proposes a four-part framework borrowed from defamation law to analyze the complex context surrounding a threat.

I.        Supreme Court Development of the True Threat Doctrine

The protections enshrined in the First Amendment are broad, prohibiting Congress and state legislatures from making any law “abridging the freedom of speech.”24U.S. Const. amend. I; see also Police Dep’t of Chi. v. Mosley, 408 U.S. 92, 95–96 (1972). The Supreme Court has read the Amendment to express “a profound national commitment to the principle that debate on public issues should be uninhibited, robust, and wide-open,” even where such debate includes “vehement, caustic, and sometimes unpleasantly sharp attacks on government and public officials.”25N.Y. Times Co. v. Sullivan, 376 U.S. 254, 270 (1964). In practice, this constitutional mandate prevents state actors from punishing or restricting the expression of an idea “simply because society finds the idea itself offensive or disagreeable.”26Texas v. Johnson, 491 U.S. 397, 414 (1989).

But the First Amendment’s reach is not unlimited.27R.A.V. v. City of St. Paul, 505 U.S. 377, 382–83 (1992) (“From 1791 to the present, however, our society, like other free but civilized societies, has permitted restrictions upon the content of speech in a few limited areas . . . .”). The Supreme Court has held that certain categories of speech should not receive the typical protection, generally because these categories have “no essential part of any exposition of ideas, and are of such slight social value as a step to truth” that any possible benefits of the speech are “clearly outweighed by the social interest in order and morality.”28Chaplinsky v. New Hampshire, 315 U.S. 568, 572 (1942). These narrow exceptions to the general rule of First Amendment protection include obscenity, defamation, fighting words, and—most relevant here—true threats.29See Virginia v. Black, 538 U.S. 343, 359 (2003); R.A.V., 505 U.S. at 383.

Threats are, at the most basic level, “pure speech.”30Watts v. United States, 394 U.S. 705, 707 (1969).They are statements made strictly for the purpose of conveying a particular idea, like “I am going to hurt you.”31See Speech, Black’s Law Dictionary (11th ed. 2019) (defining pure speech as “[w]ords or conduct limited in form to what is necessary to convey the idea”). Traditionally, pure speech receives the highest degree of First Amendment protection.32Id. Yet states and the federal government have enacted myriad laws criminalizing threats in one form or another,33See, e.g., 18 U.S.C. § 875; 18 Pa. Stat and Cons. Stat. Ann. § 2706 (2015); Cal. Penal Code § 71 (West 2014); Model Penal Code § 211.3 (Am. L. Inst. 1962). prosecuting speakers based solely on their words as well as the meaning that those words conveyed. This is possible because the Supreme Court has carved out “true threats” as categorically unprotected by the Constitution due to their lack of value and countervailing high costs to society.34See Chaplinsky v. New Hampshire, 315 U.S. 568, 572 (1942). As a result, a defendant’s liberty may hinge on which threats count as “true”—a naturally complicated question. This thorniness is exacerbated by the fact that the Court has only ever heard three true threat cases: Watts v. United States,35394 U.S. 705 (1969). Virginia v. Black,36538 U.S. 343 (2003). and Elonis v. United States.37575 U.S. 723 (2015). To lay out what the Court has (and, more importantly, has not) established about true threats, this Part explains the relevant holdings of each case in turn.

A.     Watts v. United States

When the Supreme Court confronted a criminal threat conviction for the first time in 1969, it created as much ambiguity as it resolved. The case, Watts v. United States, involved a federal statute that prohibited “knowingly and willfully” making “any threat to take the life of or to inflict bodily harm upon the President of the United States.”38Watts, 394 U.S. at 705. Defendant Robert Watts, who had just been drafted into the military during the Vietnam War, was charged with stating that “[i]f they ever make me carry a rifle the first man I want to get in my sights is L. B. J.,” referring to then-President Lyndon B. Johnson.39Id. at 706. Watts argued that his conviction violated his First Amendment rights.

The Court accepted the government’s stated interest in protecting the president as compelling.40Id. at 707. It also agreed that, in the most literal sense, Watts had clearly made a threat against President Johnson.41See id. at 706. But the Court then proceeded to scrutinize the conviction with the “commands of the First Amendment clearly in mind.”42Id. at 707. As a result, the Court overturned Watts’s conviction, finding that his speech was “the kind of political hyperbole” that could not constitute a punishable true threat.43Id. at 708.

Yet the Court hardly explained why Watts’s speech qualified as mere political hyperbole. Its reasoning spanned a few short sentences and highlighted just three key facts of the case: the location of Watts’s statement in the “political arena,” the “expressly conditional” nature of his statement, and the fact that listeners laughed in reaction to his statement.44Id. Courts and scholars have since interpreted Watts as an instruction to consider broader contextual circumstances, and those three factors in particular, when evaluating whether a statement rises to the level of a criminally punishable true threat.45See Jing Xun Quek, Elonis v. United States: The Next Twelve Years, 31 Berkeley Tech. L.J. 1109, 1113 (2016). Section II.C explores the development of these factors by lower courts in more detail. Still, the holding is difficult to generalize, and lower courts unsurprisingly have struggled to apply the doctrine in the years since Watts.46See Adrienne Scheffey, Note, Defining Intent in 165 Characters or Less: A Call for Clarity in the Intent Standard of True Threats After Virginia v. Black, 69 U. Mia. L. Rev. 861, 872 (2015).

B.     Virginia v. Black

When the Supreme Court returned to the issue of true threats in 2003, its decision again failed to resolve much of the doctrine’s uncertainty. The Court in Virginia v. Black considered the constitutionality of a law that banned cross burning done with “an intent to intimidate a person or group of persons.”47538 U.S. 343, 347 (2003). The law also stated that any cross burning served as “prima facie evidence” of the requisite intent to intimidate.48Black, 538 U.S. at 348. Accordingly, the trial court instructed the jury that “the burning of a cross by itself is sufficient evidence from which you may infer the required intent.”49Id. at 349. The Court ultimately held that this jury instruction violated the First Amendment on its face.50Id. at 367 (plurality opinion). The problem, according to the Court, was that a fact-specific inquiry into intent was the only way to distinguish an unprotected true threat from “lawful political speech at the core of what the First Amendment is designed to protect.”51Id. at 365. The jury instruction relieved the jury of its constitutional duty to perform this inquiry, rendering the defendants’ convictions invalid.52See id. A key holding of Black, then, was that a contextual, fact-intensive analysis of a defendant’s intent is required to identify and punish a true threat within the bounds of the Constitution.

While this holding may seem like a step toward clarity, the Black Court did not explicitly name the contextual factors that might be determinative in finding an intent to intimidate. Instead, the Court listed hypothetical contexts in which cross burnings might occur, implying that some may count as constitutionally proscribable true threats while others may not.53Id. at 366–67. The Court concluded that failing to distinguish between the intent reflected by those differing contexts would constitute a “shortcut” not permitted by the First Amendment. Id. Read broadly, the list suggests that relevant factors include the defendant’s stated purpose, the size of the defendant’s audience, and the location where the speech occurred.54See id. Despite emphasizing the importance of contextual factors in true threat cases, the Court in Black said very little as to what those factors might look like or how they should be weighed.

Black was also significant because it marked the first time the Court affirmatively defined the phrase “true threat.” According to the Court, a true threat is a “statement[] where 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.”55Id. at 359 (majority opinion). The Court further explained that true threats are not protected by the First Amendment mainly because the fear of violence they foster is itself a harm, one the state can seek to prevent.56Id. at 360 (“[A] prohibition on true threats ‘protect[s] individuals from the fear of violence’ and ‘from the disruption that fear engenders,’ in addition to protecting people ‘from the possibility that the threatened violence will occur.’” (second alteration in original) (quoting R.A.V. v. City of St. Paul, 505 U.S. 377, 388 (1992))). This description made clear that a threat is not “true” based on whether or not the speaker “actually intend[ed] to carry out the threat.”57Id.; see also Paul T. Crane, Note, “True Threats” and the Issue of Intent, 92 Va. L. Rev. 1225, 1261 (2006); Scheffey, supra note 46, at 873–74. Instead, what matters is whether the speaker “mean[t] to communicate” a serious threat of violence.58Black, 538 U.S. at 359. In this sense, “true” is a misleading label for the sort of threat that the First Amendment does not protect. More accurately (albeit less academically), a threat is punishable only when it is very scary and was intended to be so.59You may suspect that scariness is not an easy thing for a court to ascertain. This intuition would be correct. See infra Part II.

C.     Elonis v. United States

The Supreme Court managed to shed even less light on the doctrine in its most recent case involving an alleged true threat, Elonis v. United States.60575 U.S. 723 (2015), rev’g 730 F.3d 321 (3d Cir. 2013). There, defendant Anthony Elonis was charged under the federal threat statute for posting rap lyrics to Facebook that included apparent threats of violence toward his ex-wife.61Elonis, 575 U.S. at 726–31. The main issue was that the statute did not explicitly include an intent requirement.62Id. at 732. The Court cited criminal law and statutory-interpretation principles, not free-speech doctrine, in its holding that a defendant could only be convicted under the federal threat statute if the government proved that the defendant communicated with the purpose or knowledge that the statement would be viewed as a threat.63Id. at 734–35, 737–40. But the majority did not rely on the First Amendment at all, cabining its holding to the single statute at issue.64Id. at 740; see also Marley N. Brison, Note, Elonis v. United States: The Need for a Recklessness Standard in True Threats Jurisprudence, 78 U. Pitt. L. Rev. 493, 505 (2017). Consequently, Black and Watts remain the only binding Supreme Court cases regarding true threats from a First Amendment perspective.

Elonis does, however, provide some insight into how the justices on the Court in 2015 were thinking about true threats. Justice Alito authored a separate opinion that reached the broader First Amendment implications of the case.65Elonis, 575 U.S. at 742 (Alito, J., concurring in part and dissenting in part). He would have held that free-speech principles required a prosecutor to show that the defendant was at least reckless in making a statement that could reasonably be perceived as a threat.66Id. at 746–48. He also rejected Elonis’s argument that any intent requirement short of a subjective “intent to harm” would unconstitutionally punish protected speech.67Id. at 746–47. Justice Alito pointed to other areas of First Amendment law, such as defamation, in which the Court had approved a recklessness standard as sufficiently protective of speech while still allowing governments to punish unprotected statements.68Id. at 748 (citing N.Y. Times Co. v. Sullivan, 376 U.S. 254, 279–80 (1964), and Garrison v. Louisiana, 379 U.S. 64, 74–75 (1964)).

Lastly, Justice Alito shot down Elonis’s argument that his rap lyrics were not a threat but rather a “work[] of art” entitled to full First Amendment protection. In his reasoning, Justice Alito cited Watts and its emphasis on context.69Id. at 747. He noted that certain facts in the record reflected Elonis’s intent to threaten rather than an intent to create “art.”70Id. at 748. These facts included Elonis’s history of abusing his wife and his apparent efforts to make sure that his wife saw his posts.71See id. at 747–48. According to Justice Alito, threatening statements in song lyrics, on social media, and elsewhere must be “[t]aken in context” to avoid “grant[ing] a license to anyone who is clever enough to dress up a real threat in the guise of rap lyrics, a parody, or something similar.”72Id. at 747. Justice Alito’s assertion that “context matters” is indisputable, but it does not explain how courts can use context to unmask disguised threats in practice.

II.      Frontiers of Judicial Confusion in the True Threat Doctrine

Supreme Court precedent ultimately clarifies very little about how to apply the true threat doctrine. As a result, lower-court opinions in true threat cases are, collectively, a mess. This Part sorts the array of unresolved doctrinal questions into three distinct issues and explores how lower courts have attempted to handle them.

The first issue, discussed in Section II.A, involves the intent requirement for true threats. Specifically, this Section collects case law addressing whether the First Amendment requires the government to show that the defendant subjectively intended to threaten or if the speech must be objectively threatening.

After a court takes a side in the formalistic intent debate, there remains the harder task of applying the facts to the law. Section II.B explains the standard of review by which appellate courts should analyze a true threat conviction to evaluate whether it offends the First Amendment. Section II.C then examines the contextual factors appellate courts deploy when interpreting a true threat defendant’s intent. Finally, Section II.D focuses on one recent case, Commonwealth v. Knox,73190 A.3d 1146 (Pa. 2018). to demonstrate how the judicial confusion in true threat jurisprudence risks inadequate protection of First Amendment rights.

A.     Intent Requirement

Most judicial and scholarly debate regarding the true threat doctrine surrounds the legal question of requisite intent.74Crane, supra note 57, at 1261. After Black, a circuit split developed over whether the First Amendment required the government to prove that the defendant had an objective intent to threaten (based on a “reasonable person’s” perception of the speech) or a subjective intent to threaten (based on the defendant’s own state of mind).75Id. at 1261–65; Scheffey, supra note 46. The majority of federal appellate courts initially adopted some version of the objective test.76United States v. Clemens, 738 F.3d 1, 12 (1st Cir. 2013) (“[T]his court has applied an objective defendant vantage point standard post-Black.”); United States v. Davila, 461 F.3d 298, 305 (2d Cir. 2006) (“The test is . . . whether an ordinary, reasonable recipient . . . familiar with the context . . . would interpret it as a threat . . . .” (quoting United States v. Malik, 16 F.3d 45, 49 (2d Cir. 1994))); United States v. Elonis, 730 F.3d 321, 331 n.7 (3d Cir. 2013) (“[O]ur test asks whether a reasonable speaker would foresee the statement would be understood as a threat.”), rev’d, 575 U.S. 723 (2015); United States v. White, 670 F.3d 498, 507 (4th Cir. 2012) (describing test as whether “an ordinary reasonable recipient . . . familiar with the context . . . would interpret [the statement] as a threat” (alteration in original) (quoting United States v. Armel, 585 F.3d 182, 185 (4th Cir. 2009))); Porter v. Ascension Par. Sch. Bd., 393 F.3d 608, 616 (5th Cir. 2004) (“Speech is a ‘true threat’ and therefore unprotected if an objectively reasonable person would interpret the speech as a ‘serious expression of an intent to cause . . . harm.’” (quoting Doe v. Pulaski Cnty. Special Sch. Dist., 306 F.3d 616, 622 (8th Cir. 2002))); United States v. Jeffries, 692 F.3d 473, 479 (6th Cir. 2012) (describing test based on perceptions of “reasonable person”); United States v. Parr, 545 F.3d 491, 499 (7th Cir. 2008) (describing “objective ‘reasonable person’ test”); United States v. Nicklas, 713 F.3d 435, 440 (8th Cir. 2013) (describing test as whether “a reasonable recipient would have interpreted the defendant’s communication as a serious threat”); United States v. Martinez, 736 F.3d 981, 988 (11th Cir. 2013) (describing test based on “position of an objective, reasonable person”). But the Ninth and Tenth Circuits instead applied a test that focused on the speaker’s subjective intent to threaten a person or group with his or her statement.77See United States v. Cassel, 408 F.3d 622, 633 (9th Cir. 2005) (describing test as whether “speaker subjectively intended the speech as a threat”); United States v. Magleby, 420 F.3d 1136, 1139 (10th Cir. 2005) (describing test based on “speaker’s intent . . . of placing the victim in fear of bodily harm or death’” (quoting Virginia v. Black, 538 U.S. 343, 360 (2003))); cf. United States v. Heineman, 767 F.3d 970, 978 (10th Cir. 2014) (quoting reasonable-recipient test but adding requirement that the defendant intend that the recipient feel threatened). But see United States v. Wheeler, 776 F.3d 736, 743 n.3 (10th Cir. 2015) (denying that subjective-intent requirement was part of Heineman’s First Amendment analysis).

By mandating a subjective-intent standard in only a narrow set of true threat cases brought under the federal threat statute, Elonis failed to resolve this circuit split.78See supra Section I.C. The divergence in standards of intent that courts apply in true threat cases is an alarming issue, and it is one that the Supreme Court must promptly address.79See Elonis v. United States, 575 U.S. 723, 750 (2015) (Thomas, J., dissenting) (“This failure to decide throws everyone from appellate judges to everyday Facebook users into a state of uncertainty.”). It is also worth noting that, in 2017, Justice Sotomayor suggested that she believes the First Amendment requires a subjective-intent standard. Perez v. Florida, 137 S. Ct. 853, 855 (2017) (Sotomayor, J., concurring in denial of certiorari) (“Together, Watts and Black make clear that . . . some level of intent is required. And these two cases strongly suggest that it is not enough that a reasonable person might have understood the words as a threat—a jury must find that the speaker actually intended to convey a threat.”). Consistency among lower courts is essential for ensuring that the speech rights of every true threat defendant are adequately and equally protected. Scholarship debating the merits of the different standards and proposing new alternatives abounds.80See, e.g., Scheffey, supra note 46; Crane, supra note 57; Brison, supra note 64; Alex J. Berkman, Comment, Speech as a Weapon: Planned Parenthood v. American Coalition of Life Activists and the Need for a Reasonable Listener Standard, 29 Touro L. Rev. 485 (2013); Jennifer E. Rothman, Freedom of Speech and True Threats, 25 Harv. J.L. & Pub. Pol’y 283 (2001). Some advocates have even suggested that a defendant’s speech must be both objectively and subjectively intended to threaten for a conviction to be constitutional. See, e.g., Petition for a Writ of Certiorari at 3, Knox v. Pennsylvania, 139 S. Ct. 1547 (2019) (No. 18-949).

But while consistency itself is critical, the formalistic choice between a subjective- or objective-intent standard may be less impactful in the long run.81See Amicus Curiae Brief of the ACLU et al. in Support of Petitioner at 21, Elonis, 575 U.S. 723 (No. 13-983), 2014 WL 4215752 (“Requiring the government to demonstrate subjective intent to threaten in true threat cases would not substantially hinder its ability to prosecute actually intended threats.”); Crane, supra note 57, at 1276. Even if instructed to use a subjective-intent standard, triers of fact cannot actually know what the defendant was thinking and therefore will likely use some baseline “reasonable person” lens to guess at the defendant’s intent.82See generally R. George Wright, Objective and Subjective Tests in the Law, 16 U.N.H. L. Rev. 121, 124 (2017) (“What is thought by the law to be subjective actually pervades and informs, in multiple ways, what is thought to be objective, and vice versa. . . . The law’s attempts, in various contexts, to differentiate or combine objective and subjective tests are thus inevitably fruitless.”). Likewise, testimony or other evidence indicating what the defendant was actually thinking will undoubtedly have some relevance in a trier of fact’s assessment of intent, even under an objective analysis.83See id. Additionally, the intent standard is only a threshold question of law. A judge’s instruction to the jury on that issue will never be the final word on whether a defendant is guilty.84See Megan R. Murphy, Comment, Context, Content, Intent: Social Media’s Role in True Threat Prosecutions, 168 U. Pa. L. Rev. 733, 746–47 (2020) (“The requirement of subjective and objective mens rea standards will not resolve the other question left open by Elonis: what evidence will suffice to prove a subjective intent to put some audience member in fear of serious harm.”). Rather than adding another voice to the existing—and largely formalistic—debate over intent, this Note focuses primarily on what courts should do once a case passes that threshold.

B.     Standard of Review

That judges struggle to separate true threats from protected speech is unsurprising. As Black made clear, whether a defendant truly “intended” to threaten is a deeply fact-intensive question,85See supra Section I.B. and questions of fact are usually best left to juries.86See Stephen E. Arthur & Robert S. Hunter, 1 Federal Trial Handbook: Criminal §§ 10:1–:2 (2020–2021 ed. 2020). See generally Ramos v. Louisiana, 140 S. Ct. 1390 (2020) (discussing the origins and meaning of the Sixth Amendment right to a jury trial). A criminal defendant’s mental state, in particular, is an issue that must be decided by the trier of fact based on all direct and circumstantial evidence produced at trial.8775A Am. Jur. 2d Trial § 669 (2018). It is generally not a trial judge’s job to peer into the mind of the defendant.88See id. Even on appeal, a judge should traditionally defer to the factfinder below as much as possible by viewing the evidence in the light most favorable to the verdict and asking only “whether any rational trier of fact” could have found the defendant guilty “beyond a reasonable doubt.”895 Am. Jur. 2d Appellate Review § 583 (2018) (emphasis added). This division of labor is well justified,90Amanda Reid, Safeguarding Fair Use Through First Amendment’s Asymmetric Constitutional Fact Review, 28 Wm. & Mary Bill Rts. J. 23, 27 (2019) (“Deference accords greater finality to fact-finding, and enhances judicial economy by reducing the frequency of appeals. Deference promotes efficiency and stability by recognizing the superior institutional competence of the lower court to engage in fact-finding. . . . Lack of deference undermines the legitimacy and finality of the trial process. Lack of deference raises distributive concerns because often only the wealthy can afford two bites of the apple. And de novo review ultimately renders the jury a nullity because, without deference, the jury’s role is little more than a dry run.”). and it helps explain why courts in true threat cases focus mostly on the subjective/objective dimension of intent: it is strictly a question of law, fit for judicial resolution rather than jury consideration.

But the usual degree of deference to the finder of fact may not adequately protect a defendant’s First Amendment rights. The Supreme Court has held that an appellate court considering First Amendment issues must “make an independent examination of the whole record” to ensure that “the judgment does not constitute a forbidden intrusion on the field of free expression.”91N.Y. Times Co. v. Sullivan, 376 U.S. 254, 285 (1964) (quoting Edwards v. South Carolina, 372 U.S. 229, 232 (1963)); see also Bose Corp. v. Consumers Union of U.S., Inc., 466 U.S. 485, 499 (1984). The requirement that courts conduct an independent de novo review of facts with special constitutional relevance, known as the “constitutional fact doctrine,”92Henry P. Monaghan, Constitutional Fact Review, 85 Colum. L. Rev. 229, 231 (1985). is necessitated by the frequent blurriness of the line separating protected speech from unprotected speech. In Bose Corp. v. Consumers Union of United States, the Court explained that giving juries a mere “general description” of speech that is unprotected would not sufficiently guide them in understanding what is and is not punishable, creating too great a danger that their verdicts would “inhibit the expression of protected ideas.”93Bose, 466 U.S. at 505. The Court has also reasoned that jurors are “unlikely to be neutral” when evaluating speech and ideas that they might find highly offensive but which are nonetheless protected by the Constitution.94Monitor Patriot Co. v. Roy, 401 U.S. 265, 277 (1971).

The Court has applied the constitutional fact doctrine in First Amendment contexts, including defamation lawsuits,95See, e.g., Sullivan, 376 U.S. 254; Bose, 466 U.S. 485. obscenity prosecutions,96See, e.g., Jacobellis v. Ohio, 378 U.S. 184, 190 (1964) (opinion of Brennan, J.). and more,97See, e.g., Hurley v. Irish-Am. Gay, Lesbian & Bisexual Grp. of Bos., Inc., 515 U.S. 557, 567 (1995); Street v. New York, 394 U.S. 576, 592 (1969); Pennekamp v. Florida, 328 U.S. 331, 335 (1946). but it has yet to definitively extend it to true threat cases. Federal courts of appeals disagree on whether to conduct the special review required by the constitutional fact doctrine in appeals of true threat convictions, creating yet another source of inconsistency in this area of law. While the Fourth and Ninth Circuits have held that the doctrine does apply to true threat cases,98United States v. Bly, 510 F.3d 453, 457–58 (4th Cir. 2007); Planned Parenthood of the Columbia/Willamette, Inc. v. Am. Coal. of Life Activists, 290 F.3d 1058, 1070 (9th Cir. 2002) (en banc). the other circuits addressing the issue have generally applied traditional deferential review.99United States v. Wheeler, 776 F.3d 736 (10th Cir. 2015); United States v. Jeffries, 692 F.3d 473, 481 (6th Cir. 2012); United States v. Parr, 545 F.3d 491, 497 (7th Cir. 2008); United States v. Schiefen, 139 F.3d 638, 639 (8th Cir. 1998). The Second Circuit has not weighed in definitively, but it has noted the uncertainty around the issue and floated the possibility that an appellate court might have the discretion, but not the duty, to conduct an independent review of constitutional facts. United States v. Turner, 720 F.3d 411, 419 (2d Cir. 2013). State appellate courts are similarly split.100Compare People v. Stanley, 170 P.3d 782, 790 (Colo. App. 2007) (applying constitutional fact doctrine), with Commonwealth v. Sinnott, 30 A.3d 1105, 1110 (Pa. 2011) (applying traditional sufficiency-of-evidence standard of review).

Given the near-routine application of the constitutional fact doctrine to other First Amendment cases, the argument in favor of extending it to true threat cases is fairly intuitive. As an initial matter, the question of whether a particular statement is a true threat is itself a “constitutional fact.” The First Amendment prohibits punishing an individual purely for the content of her speech (however offensive it may be) unless the speech fits into a category that the Supreme Court has deemed unprotected, such as true threats. Thus, the constitutionality of a conviction is inseparable from whether the speech counts as a true threat.101See supra and infra this whole Note.

Analogies to other First Amendment applications of the constitutional fact doctrine also support its extension to true threats. For example, a threat charge implicates speech rights to the same extent as an obscenity charge because pure speech serves as grounds for criminal punishment in both situations. In fact, compared with civil defamation lawsuits—another area where the constitutional fact doctrine is applied as a matter of course—threat prosecutions actually risk greater burdens on speech. In the true threat case, removing the defendant’s speech from the First Amendment’s protected sphere would result in her imprisonment or other criminal sanction, whereas a defamation defendant usually risks only monetary loss.

Another reason for appellate courts to independently review the constitutional fact of a true threat is the extraordinary ambiguity about what counts as a “true threat.” The Supreme Court did not even conjure a definition for such threats until 2003,102See supra Section I.B. yet we expect jurors to know one when they see one. The task is simply too hard and too important; another layer of protection is necessary. This is not to say the jury’s verdict does not matter. Appellate judges are well aware of the traditional sanctity of a jury verdict and will hesitate to vacate it even when reviewing the record de novo.103Kenneth L. Karst, Threats and Meanings: How the Facts Govern First Amendment Doctrine, 58 Stan. L. Rev. 1337, 1370 (2006). But judges can nonetheless look at the facts underlying the conviction “with the commands of the First Amendment clearly in mind”104Watts v. United States, 394 U.S. 705, 707 (1969). in a way that a jury simply cannot. In true threat cases, they should.

The argument against applying the constitutional fact doctrine to true threats is unclear, as courts that have declined to do so have not provided detailed reasons for their decisions. For example, in United States v. Wheeler, the Tenth Circuit simply said that “whether statements amount to true threats ‘is a question generally best left to a jury,’” unless there is an “unusual set of facts.”105United States v. Wheeler, 776 F.3d 736, 742 (10th Cir. 2015) (quoting United States v. Malik, 16 F.3d 45, 51 (2d Cir. 1994)). The court did not explain why the constitutional fact doctrine should not apply to true threat cases, stating merely that it would follow Tenth Circuit precedent in which the court had declined to conduct independent review.106Id. (citing United States v. Viefhaus, 168 F.3d 392, 397 (10th Cir. 1999)). The Sixth, Seventh, and Eighth Circuits also used the sufficiency-of-evidence standard without even mentioning the possibility of de novo review of the constitutional fact of a true threat.107United States v. Jeffries, 692 F.3d 473, 481 (6th Cir. 2012); United States v. Parr, 545 F.3d 491, 497 (7th Cir. 2008); United States v. Schiefen, 139 F.3d 638, 639 (8th Cir. 1998).

De novo review of the fact finder’s determination that a statement constitutes a true threat maximizes protection of a defendant’s speech rights and aligns with Supreme Court precedent regarding when to apply the constitutional fact doctrine. But, like the subjective/objective intent question, the question of the appropriate standard of review on appeal is a formalistic and preliminary one that will not solve the true threat doctrine’s most important problems. The degree to which an appellate court scrutinizes the record below hardly matters if the court does not know which facts are relevant to the true threat designation. Thus, courts have to face the next frontier of judicial confusion within this ungainly doctrine: context.

C.     Contextual Analysis

The last major problem area within the true threat doctrine—and the one that most significantly impacts whether or not a statement qualifies as a true threat—is the contextual analysis called for by the Supreme Court’s muddled true threat precedent. In Watts, Black, and Elonis, the Court emphasized that “context matters.”108Elonis v. United States, 575 U.S. 723, 747 (2015) (Alito, J., concurring in part and dissenting in part). By and large, courts have properly recited this broad instruction to consider contextual factors in true threat cases.109United States v. Syring, 522 F. Supp. 2d 125, 130 (D.D.C. 2007) (discussing how several circuits analyze threats in their “entire factual context”). But which context matters, and how much does it matter? This Section compares opinions from various federal courts of appeals to determine which contextual factors, if any, recur in their evaluations of true threat defendants’ intent.

Before surveying courts’ approaches to context, a word on the scope of this analysis. Due to the sheer ubiquity of statutes that criminalize threats, it is not feasible to collect every case in which a defendant has been convicted for making a true threat. Indeed, the abundance of threat convictions in the United States underscores the true threat doctrine’s extraordinary power and the government’s reliance on it to punish criminal behavior. Looking at federal law alone, the true threat doctrine has enabled convictions for threats against the president,110See Watts v. United States, 394 U.S. 705 (1969) (affirming conviction under 18 U.S.C. § 871). threats to kidnap or injure,111See United States v. Khan, 937 F.3d 1042 (7th Cir. 2019) (affirming conviction under 18 U.S.C. § 875). stalking,112See United States v. Ackell, 907 F.3d 67 (1st Cir. 2018) (affirming conviction under 18 U.S.C. § 2261A). blocking access to reproductive health clinics,113See Planned Parenthood of the Columbia/Willamette, Inc. v. Am. Coal. of Life Activists, 290 F.3d 1058 (9th Cir. 2002) (en banc) (affirming conviction under 18 U.S.C. § 248). and more.114See, e.g., United States v. Coss, 677 F.3d 278 (6th Cir. 2012) (affirming conviction under federal statute banning threats to injure the reputation of another with intent to extort money); United States v. Fullmer, 584 F.3d 132 (3d Cir. 2009) (affirming conviction under federal statute banning threats involving animal enterprises). Because state statutes criminalizing different forms of threats are even more numerous,115See, e.g., John P. Ludington, Annotation, Validity and Construction of Terroristic Threat Statutes, 45 A.L.R.4th 949 (1986); Erin Masson Wirth, Annotation, Imposition of State or Local Penalties for Threatening to Use Explosive Devices at Schools or Other Buildings, 79 A.L.R.5th 1 (2000). this Section focuses only on contextual factors repeatedly identified by federal appellate court case law.

The Supreme Court’s opinions in Watts and Black highlight a few specific factors relevant to the contextual analysis. In holding that the defendant’s speech threatening to shoot LBJ was not a true threat, the Watts Court noted the location of the defendant’s statement in the “political arena,” the “expressly conditional” nature of his statement, and the fact that listeners laughed in reaction to his statement.116Watts v. United States, 394 U.S. 705, 708 (1969). The Black Court highlighted factors such as the defendant’s stated purpose, the size of the defendant’s audience, and the location where the speech occurred.117See supra text accompanying note 54. These factors are a starting point, but nothing in Watts or Black suggests that courts should be limited to them. To the contrary, Black seemed to encourage as much consideration of context as possible.118See supra note 53 and accompanying text. Justice Alito’s Elonis opinion heeded that recommendation, as it mentioned several factors that do not fit perfectly in the Watts and Black boxes.119See supra notes 71–72 and accompanying text.

The federal circuit courts have also identified contextual factors relevant to the true threat analysis. The Eighth Circuit, in United States v. Dinwiddie, provided a comprehensive—though explicitly nonexhaustive and nondispositive—list of factors to be considered in reviewing a true threat conviction.12076 F.3d 913, 925 (8th Cir. 1996). These factors include how the threat recipient and other listeners reacted, whether the threat was conditional, whether the threat was made directly to its target, whether the speaker made similar statements to the recipient in the past, and whether the recipient had reason to believe that the speaker was prone to violence.121Dinwiddie, 76 F.3d at 925. In upholding the constitutionality of the defendant’s conviction, the Dinwiddie court emphasized that she had directly repeated her threats about fifty times to her intended target.122Id.

Dinwiddie’s list of factors frequently appears in other opinions analyzing true threats.123See, e.g., United States v. Syring, 522 F. Supp. 2d 125, 130 (D.D.C. 2007); Planned Parenthood of the Columbia/Willamette, Inc. v. Am. Coal. of Life Activists, 290 F.3d 1058, 1078 (9th Cir. 2002) (en banc); State v. Taylor, 841 S.E.2d 776 (N.C. Ct. App.), rev’d, No. 156PA20, 2021 WL 5984471 (N.C. 2021). The Third Circuit’s rehearing of Elonis following the Supreme Court’s remand is one example.124United States v. Elonis, 841 F.3d 589 (3d Cir. 2016), cert. denied, 138 S. Ct. 67 (2017). There, the court highlighted the efforts of Elonis’s ex-wife to seek a restraining order against him after he made the alleged threats,125Id. at 598. exemplifying the relevance of the “reaction of the recipient of the threat.”126Dinwiddie, 76 F.3d at 925. It also noted the history of Elonis’s violent statements toward his ex-wife,127Elonis, 841 F.3d at 598. mirroring the category of “similar statements to the victim in the past” from Dinwiddie.128Dinwiddie, 76 F.3d at 925. These considerations, among others, led the Third Circuit to find only harmless error in the jury instruction and to uphold Elonis’s conviction.129Elonis, 841 F.3d at 598. Similarly, in United States v. Clemens, the First Circuit upheld a true threat conviction and noted that the threat recipients “did in fact feel fear” that the defendant would harm them and reacted “to protect themselves.”130United States v. Clemens, 738 F.3d 1, 14 (1st Cir. 2013). Clemens, along with Watts and Dinwiddie, thus exemplifies judicial reliance on the listeners’ reaction as a key contextual factor.

To be clear, courts do not limit themselves to the precise Dinwiddie list or any similar iteration of factors. Instead, they generally stress the need to consider “the whole factual context and all the circumstances bearing on a threat.”131Planned Parenthood of the Columbia/Willamette, Inc. v. Am. Coal. of Life Activists, 290 F.3d 1058, 1080 (9th Cir. 2002) (en banc). In practice, this seems to mean accounting for any factor that feels relevant to the court on a case-by-case basis. For example, in United States v. Wheeler, the defendant was charged for posting on social media and urging his “religious operatives” to “commit a massacre” at a preschool and day care.132United States v. Wheeler, 776 F.3d 736, 745 (10th Cir. 2015). The Tenth Circuit found that these statements were punishable true threats based partly on a contextual factor it referred to as “the collective consciousness,” which in modern times “includes recent massacres at educational and other institutions by active shooters.”133Id. Another example of a court looking at non-Dinwiddie context is United States v. Turner, in which the defendant was charged for a blog post declaring that three Seventh Circuit judges deserved to die for their recent holding on a Second Amendment issue.134United States v. Turner, 720 F.3d 411, 413 (2d Cir. 2013). In upholding that defendant’s conviction as constitutional, the Second Circuit emphasized his publication of photographs, work addresses, and room numbers of the targeted judges.135Id. at 423. The court in Turner seemed to view the specificity of the threat as important context and thus accounted for it in its true threat analysis,136Id. though this did not fall cleanly into any of the Dinwiddie factors.

This free-for-all approach to considering context is common, yet it is far from ideal. Even when judges look at the same factual record, there is substantial space for disagreement about which factors matter and what they mean. United States v. Bagdasarian exemplifies this issue.137652 F.3d 1113 (9th Cir. 2011). Defendant Walter Bagdasarian, “an especially unpleasant fellow,” was convicted in the trial court for viciously racist statements about Barack Obama anonymously posted to an online message board.138Bagdasarian, 652 F.3d at 1115–16. One message said Obama would “have a 50 cal in the head soon” and another called for Obama to be shot.139Id. at 1115. These posts were reported to the Secret Service, which investigated and eventually found six guns (including a .50 caliber rifle) at Bagdasarian’s home.140Id. at 1115–16. They also discovered that on the day Obama was elected, the defendant sent emails referring to shooting Obama’s car.141Id. at 1116.

The Ninth Circuit panel in Bagdasarian agreed on the pure legal questions about intent and standard of review, but Judge Wardlaw dissented from the majority’s interpretation of the case’s key context. The majority was ultimately unpersuaded that Bagdasarian’s posts rose to the level of a true threat.142Id. at 1115. It acknowledged the violence of the posts but determined that they were merely exhortations or predictions.143See id. at 1123 (“Given that Bagdasarian’s statements . . . fail to express any intent on his part to take any action, the fact that he possessed the weapons is not sufficient to establish that he intended to threaten Obama himself. [The Election Day emails] simply provide additional information—weblinks to a video of debris and two junked cars being blown up and to an advertisement for assault rifles available for purchase online—that Bagdasarian may have believed would tend to encourage the email’s recipient to take violent action against Obama. But . . . incitement to kill or injure a presidential candidate does not qualify as an offense under § 879(a)(3).”). Similarly, the majority thought that neither Bagdasarian’s possession of a gun nor his anonymity in posting were enough to prove his specific intent to threaten Obama, particularly because the board to which Bagdasarian posted was “a non-violent discussion forum that would tend to blunt any perception that statements made there were serious expressions of intended violence.”144Id. at 1121, 1123. Judge Wardlaw took the opposite view. She would have held Bagdasarian’s statements were true threats and upheld his conviction based on the very contextual factors that the majority had dismissed: the readers’ perception of the posts as threatening, Bagdasarian’s access to .50 caliber firearms when he made the post, and Bagdasarian’s choice to hide behind a “cloak of anonymity” until the Secret Service found him.145Id. at 1128–31 (Wardlaw, J., concurring in part and dissenting in part).

The source of disagreement in Bagdasarian was not whether to consider context or even what context to consider. Instead, the judges differed in their evaluation of what the context meant and whether it proved a true threat. To some extent, differing interpretations of a set of facts are inevitable. But the dearth of guidance in how judges should weigh key contextual factors makes such disagreements infinitely more likely, which is unacceptable when the defendant’s liberty and speech rights are at stake.

D.     A Case Study: Commonwealth v. Knox

Because every true threat case requires intensive analysis of context, a granular look at a single case may be more enlightening than a zoomed-out observation of the factors that courts consider. Accordingly, this Section dissects Commonwealth v. Knox, a case that the Pennsylvania Supreme Court decided in 2018 and that the U.S. Supreme Court declined to hear in 2019.146190 A.3d 1146 (Pa. 2018). Point-by-point analysis of the Knox opinion reveals the practical impact of the true threat doctrine’s overall inscrutability. The doctrine’s confusions are exacerbated by factual records involving things like social media, rap music, and political controversies—all of which are overwhelmingly common in modern American society, yet exceedingly difficult for judges to decode.

1.       Facts and Procedural Posture

In the summer of 2012, teenager Jamal Knox recorded a rap song entitled “F–k the Police.”147Knox, 190 A.3d at 1149. The lyrics were his own, though the song’s theme was hardly original.148In 1988, rap group N.W.A. released a song entitled “Fuck Tha Police” that has had widespread cultural impact. See Rich Goldstein, A Brief History of the Phrase ‘F*ck the Police, Daily Beast (Apr. 14, 2017, 3:09 PM), []. Knox had been arrested in April of that year and charged with several unrelated criminal offenses; at the time he recorded the rap, those charges were pending.149Knox, 190 A.3d at 1148–49.

Knox’s rap expressed anger toward police in general, but also called for violence against two particular Pittsburgh officers: Officer Michael Kosko and Detective Daniel Zeltner.150Id. at 1149. Kosko and Zeltner were both slated to testify against Knox at a hearing arising from his pending criminal charges.151Id. Before the hearing, a third party uploaded a video of Knox’s rap to YouTube and posted it to a publicly viewable Facebook page.152Id. A police officer monitoring that Facebook page saw the video and showed it to Zeltner and Kosko.153Id. at 1150. Knox was subsequently arrested and charged with witness intimidation and making “terroristic threats” pursuant to Pennsylvania state law.154Id. At a bench trial, a Pennsylvania judge found Knox guilty of both charges.155Id. at 1151. Knox appealed, and the Pennsylvania Supreme Court granted review to determine whether the rap qualified as a true threat.156Id. at 1152.

2.       Threshold Issues

The Knox court first had to resolve the preliminary issues that remain unsettled within the true threat doctrine: the intent requirement and the standard of appellate review. First, the court relied on Black to hold that the First Amendment requires a state to prove that a defendant subjectively and “specifically intended to terrorize or intimidate” in order to convict a defendant of a true threat.157Id. at 1158. This placed Pennsylvania state courts squarely on the subjective-intent side of the circuit split discussed in Section II.A.

Next, the court opted to conduct an independent review of constitutional facts, though not in so many words. It noted that “whether a statement constitutes a true threat” is a “circumstance-dependent . . . mixed question of fact and law.”158Id. at 1152. Further, the court expressly classified “the scope of the true-threat doctrine” as a legal question that it should review de novo.159Id. Its reasoning for this was rather cursory, though it cited a prior decision that discussed the U.S. Supreme Court’s constitutional fact doctrine in First Amendment contexts at greater length.160Id. (citing In re Condemnation by Urb. Redevelopment Auth. of Pittsburgh, 913 A.2d 178, 183 (Pa. 2006) (“We note that the United States Supreme Court has stated that in reviewing First Amendment cases, appellate court[s] must conduct a review of the entire record.”)). In any case, the court did seek to independently review the facts (nominally, at least) as it turned to the more difficult task of contextual analysis.

3.       Contextual Analysis

The Pennsylvania Supreme Court directly embraced the U.S. Supreme Court’s emphasis on context, examining “contextual circumstances such as those referenced in Watts” in order to determine whether the trial court’s finding of an intent to threaten was supported by competent evidence.161Knox, 190 A.3d at 1158. But first, the court looked at Knox’s actual words.162Id. (“We first review the content of the speech itself, beginning with the lyrics.”). The court recited and examined the lyrics in great detail to conclude that, on their face, the lyrics were threatening and personalized toward the police, and toward Zeltner and Kosko in particular.163Id. at 1158–59.One might wonder whether any judges are really suited to the job of interpreting amateur rap lyrics, and in fact rap scholars and rappers raised this very question in an amicus brief to the cert petition.164Motion for Leave to File Brief as Amici Curiae and Brief of Amici Curiae Michael Render (“Killer Mike”), Erik Nielson & Other Artists & Scholars in Support of Petitioner at 17, Knox v. Pennsylvania, 139 S. Ct. 1547 (2019) (No. 18-949), 2019 WL 1115837 [hereinafter Killer Mike Amicus Brief]. For in-depth criticisms of the use of raps as the bases for convictions (and true threat convictions in particular), see Erin Lutes, James Purdon & Henry F. Fradella, When Music Takes the Stand: A Content Analysis of How Courts Use and Misuse Rap Lyrics in Criminal Cases, 46 Am. J. Crim. L. 77 (2019); Clay Calvert, Emma Morehart & Sarah Papadelias, Rap Music and the True Threats Quagmire: When Does One Man’s Lyric Become Another’s Crime?, 38 Colum. J.L. & Arts 1 (2014); Andrea Dennis, Poetic (In)justice? Rap Music Lyrics as Art, Life, and Criminal Evidence, 31 Colum. J.L. & Arts 1 (2007); and Andrew Jensen Kerr, Art Threats and First Amendment Disruption, 16 Duke J. Const. L. & Pub. Pol’y 173 (2021). Even the court acknowledged the special history of rap music and its tendency to feature violent imagery that is, in many cases, not meant to be taken literally.165Knox, 190 A.3d at 1160 (“We acknowledge that . . . rap music often contains violent imagery that is not necessarily meant to represent an intention on the singer’s part to carry through with the actions described. This follows from the fact that music is a form of art and ‘[a]rtists frequently adopt mythical or real-life characters as alter egos or fictional personas.’ We do not overlook the unique history and social environment from which rap arose, the fact that rap artists (like many other artists) may adopt a stage persona that is distinct from who they are as an individual, or the fact that musical works of various types may include violent references, fictitious or fanciful descriptions of criminal conduct, boasting, exaggeration, and expressions of hatred, bitterness, or a desire for revenge. In many instances, lyrics along such lines cannot reasonably be understood as a sincere expression of the singer’s intent to engage in real-world violence.” (cleaned up) (quoting Dennis, supra note 164, at 23)).

Nonetheless, the court held that Knox’s rap was “of a different nature and quality” than the many violent but non–true threat rap songs out there.166Id. First, the court characterized several of the rap’s lines—including “[l]et’s kill these cops cuz they don’t do us no good,” “that whole department can get it,” and “jam this rusty knife all in [the officer’s] guts”—as “unambiguous threats” to the police, as opposed to expressions of mere satire or irony, generalized animosity, or social commentary.167Id. at 1158. Second, the court argued that the lyrics went “beyond the realm of fantasy or fiction” because Knox rapped that he wanted the city to “believe” him and that the threats would become “real.”168Id. Third, the court highlighted the rap’s explicit naming of two specific officers, as well as references to the time the officers’ shifts ended and the identities of their confidential informants.169Id. at 1159. As an aside, the court also noted that the sounds of gunfire and police sirens in the background of the song exacerbated its threatening effect.170Id.

Only after this extensive textual evaluation of the rap did the court turn to context, to which it devoted only three paragraphs. Such a cursory analysis of context misses the point. That the words themselves were threatening is necessary but not sufficient for finding a true threat—they must be accompanied by context that affirmatively supports classifying the speech as a true threat, rather than just failing to rebut it.171Cf. In re S.W., 45 A.3d 151, 157 (D.C. 2012) (“[A] determination of what a defendant actually said is just the beginning of a threats analysis. Even when words are threatening on their face, careful attention must be paid to the context in which those statements are made to determine if the words may be objectively perceived as threatening.”). Citing to Watts and state-law precedent, the Knox court focused on only four contextual factors: whether the threat was conditional, whether the threat was communicated directly to the victim, whether the victim had reason to believe the speaker had a propensity to engage in violence, and how the listeners reacted to the speech.172Knox, 190 A.3d at 1159.

With respect to the first factor, the court found without explanation that Knox’s threat was “mostly unconditional.”173Id. The court was presumably relying on its earlier discussion of the lyrics themselves, but it is unclear which lines it was focusing on. The court then jumped to an evaluation of the listeners’ reaction to the speech, noting that the officer who first viewed the video immediately shared it with the colleagues Knox had named, who then took protective measures.174Id. Despite the video’s public accessibility, the court mentioned no listeners other than the police officers, particularly the targeted officers, who were hardly disinterested bystanders in this interaction.175The focus on police reactions also exacerbates the potential for racial bias to enter into the court’s decisionmaking. See infra notes 195–201 and accompanying test. This application of the “listeners’ reaction” factor differs from Watts, where the Court focused on listeners who had no stake in the relationship between the speaker and the alleged target (in that case, President Johnson).176See supra Section I.A. Nonetheless, the court in Knox was easily satisfied that the police reactions to the speech supported its categorization as a true threat.

The Knox court next examined the second contextual factor on its list: direct communication of the threat to its target. But, again, the court spent little time on its purported independent review.177See supra Section II.D.1. It acknowledged that Knox did not send or speak the rap directly to Zeltner, Kosko, or any other police officers.178Knox, 190 A.3d at 1160. To the contrary, the video was posted broadly to YouTube by someone other than Knox, then reposted (again, by someone other than Knox) to a public Facebook page that police happened to be monitoring.179See supra Section II.D.2. Still, the court found that Knox may have intended that police officers would hear the rap, pointing to the lower courts’ finding that Knox’s “prior course of conduct suggested [he and his co-writer] either intended for the song to be published or knew publication was inevitable.”180Knox, 190 A.3d at 1160.

The court’s cursory analysis of this key contextual factor demonstrates the inadequacy of a free-for-all approach to contextual analyses. The court seemed to defer substantially to the courts below in assessing the directness of the threat; this makes little sense given the conceded need to independently examine the whole record when speech rights are at stake,181See supra note 160. the conflict between how the superior court and trial court interpreted the evidence on this issue,182Brief for Appellant at 20–21, Knox, 190 A.3d 1146 (No. 3 WAP) (“The Trial Court did not render any findings of fact as to who uploaded the FTP Recording to the YouTube Account, or as to who posted the Link to the Facebook Page. The Trial Court found that Knox authored and created an audio recording of the Song, and that Knox thereby directly or indirectly communicated the Song with the intent to threaten police. The Superior Court’s factual findings are inconsistent with those of the Trial Court in this regard. The Superior Court found that the evidence was sufficient to find that [Knox’s co-writer] Beasley uploaded the FTP Recording to the YouTube Account and posted the Link to the Facebook Page. The Superior Court thereby found that Knox indirectly communicated the Song with the knowledge that the Song would be viewed by police or by a third party who would share it with police.” (citations omitted)). and the new and complex questions raised in this sphere by social media and internet technology.183See Murphy, supra note 84; Matt Kass, Note, Elonis v. United States: At the Crossroads of First Amendment and Criminal Jurisprudence in the Digital Age, 43 Rutgers Comput. & Tech. L.J. 110 (2017). The record itself attests to the complicated background here. At trial, the state failed to prove that Knox himself posted the video,184Knox, 190 A.3d at 1168 (Wecht, J., concurring in part and dissenting in part). and the chain of events that led to the targeted officers watching the video was long and convoluted.185Id. at 1148–50 (majority opinion). That is not to say that a thorough analysis of this factor would have definitively altered the verdict. Indeed, a concurring justice examined the context surrounding this issue in much greater detail and reached the same conclusion.186Id. at 1169–71 (Wecht, J., concurring in part and dissenting in part). It is worth noting that Justice Wecht dissented only on the threshold question of the intent requirement, and it was his analysis of context that led him to the same result. In this way, his opinion exemplifies this Part’s argument that the intent requirement is not the be-all and end-all of the true threat debate. But the fact that the majority set out to conduct an independent review of the whole record, explicitly including this factor of direct communication of the threat, yet failed to thoroughly do so reflects a serious shortcoming in current true threat jurisprudence.

As for the third factor, whether the victim had reason to believe the speaker had a propensity to engage in violence, the court again provided only a brief analysis. The majority noted as “relevant” the fact that Zeltner and Kosko “were aware a loaded firearm” was found in Knox’s car during his initial arrest back in April 2012.187Id. at 1160 (majority opinion). This was all the court had to say on the matter. There was no history of Knox actually being violent toward the police officers, nor was there any mention of prior criminal offenses other than the April 2012 arrest that preceded the rap recording.188See id. at 1148–49. It is also worth noting that Knox was just a teenager at the time of his arrest.189Brief for Appellant, supra note 182, at 13. Still, the court seemed to think Knox’s perceived propensity for violence unambiguously supported its ultimate finding that he had made a true threat, as it went on to affirm his conviction.190See Knox, 190 A.3d at 1161.

III.    Standardizing Contextual Analyses: A New Framework for Identifying True Threats

It is, by now, axiomatic that context matters in true threat prosecutions. Knox and federal appellate case law show, however, that courts have failed to develop a coherent approach to conducting this contextual analysis. Such incoherence is likely to continue to infect the doctrine regardless of whether courts opt for a subjective- or objective-intent requirement, or whether they examine the record below de novo or not. This Wild West of whatever context a particular judge sees as important is not sufficiently protective of the speech interests at stake in true threat cases. This Part proposes a four-part framework to address this problem.

A.     Knox and the Need for Reform

First, it is worth asking whether a solution is even possible or if this amalgam of contextual factors is just the best we can do. After all, totality-of-the-circumstances analyses are quite common in constitutional criminal law.191See, e.g., Dickerson v. United States, 530 U.S. 428, 434 (2000) (holding that due-process test for evaluating voluntariness of defendant’s confession takes into consideration the totality of the surrounding circumstances); Illinois v. Gates, 462 U.S. 213, 238 (1983) (mandating totality-of-the-circumstances test for probable-cause analyses). Professor Kenneth Karst, in his comprehensive overview of the true threat doctrinal landscape as of 2006, concluded that this is simply an area where the law must be beholden to the facts.192Karst, supra note 103, at 1411. He argued that judges and juries with decades of human experience will be able to look at the context and reach a socially acceptable conclusion about what is an unprotected threat and what is not.193Id. at 1368–70. Essentially, Karst told us to trust the process.

This conclusion is hard to bear when the process is hardly a process at all. As shown in Part II, appellate courts simply address whichever factors they view as relevant and give them whatever weight they feel is appropriate. Karst is correct that this is inevitable to some degree; judges must use their judgment. But to have no framework guiding the application of that judgment puts speech rights at too great a risk because, practically speaking, implicit (or explicit) bias is free to enter the equation. The likelihood of such bias in First Amendment cases is the impetus for the constitutional fact doctrine, through which the Supreme Court directs judges to vigilantly prevent the punishment of protected but unpopular speech.194See supra Section II.B. Still, judges are not immune to antipathy toward upsetting speech. And giving parties notice about what an appellate court considers when reviewing true threat convictions will help litigants decide what evidence to enter into the record in the first place, ensuring that courts have full information about the context of the speech at issue.

Viewed cynically, Knox demonstrates the risk of unstructured contextual analyses in true threat cases. Antipolice rhetoric is historically unpopular in America,195Emily Ekins, Cato Inst., Policing in America 8 (2016),‌/sites/ [] (“[M]ost Americans (64%) have favorable attitudes toward their local police department and are confident their local police use the appropriate amount of force (58%), are courteous (57%) and honest (57%), treat all racial groups equally (56%), protect people from violent crime (56%), respond quickly to a call for help (56%), and care about community members (55%).”). That said, attitudes may be changing in the wake of 2020’s backlash against police that followed the murder of George Floyd by Minneapolis police, along with other police killings of Black men and women. See Aimee Ortiz, Confidence in Police Is at Record Low, Gallup Survey Finds, N.Y. Times (Aug. 12, 2020), []. and given the long history of judicial deference to police,196See generally Anna Lvovsky, The Judicial Presumption of Police Expertise, 130 Harv. L. Rev. 1995 (2017). it is not outlandish to expect judges to be similarly unreceptive to speech like Knox’s rap lyrics.197The relative lack of diversity of the judiciary may contribute to this issue as well. See Examining the Demographic Compositions of U.S. Circuit and District Courts, Ctr. for Am. Progress (Feb. 13, 2020, 12:01 AM),‌/2020/02/13/480112/examining-demographic-compositions-u-s-circuit-district-courts [‌/3ASV-CE9C] (noting that people of color make up just 20 percent of all judges sitting on Article III circuit and district courts). When white judges from similar backgrounds dominate the bench, their general attitudes toward a particular viewpoint are likely to dominate as well. Furthermore, the fact that Knox is a young Black man who was participating in a musical genre culturally associated with young Black men should not be overlooked.198See Killer Mike Amicus Brief, supra note 164; see also Ekins, supra note 195 (noting that Black Americans disproportionately experience verbal and physical misconduct by police). In addition to making it less likely that the lyrics were meant literally, the rap context may trigger racial bias and “enduring stereotypes about the criminality of young [B]lack men” to an extent that other forms of expression may not.199Killer Mike Amicus Brief, supra note 164, at 19–20 (“Research tells us that listeners unfamiliar with hip hop culture may have difficulty being reasonable when it comes to rap music because it often primes enduring stereotypes about the criminality of young [B]lack men, its primary creators. In the criminal justice system, the results of this racial bias are evident in the disparate treatment that people of color face at virtually every phase of the criminal justice process. When it comes to rap, research reveals similar disparities.”). Indeed, police only found the video at issue because they were monitoring the Facebook page of Knox’s rap alias,200See supra Section II.D.1. a virtual version of the physical surveillance that Black Americans have been subjected to for centuries.201See, e.g., Alvaro M. Bedoya, The Color of Surveillance, Slate (Jan. 18, 2016, 5:55 AM), []. This history leads one to question the fairness of heavily weighing the police reaction to Knox’s speech as a factor in determining whether his speech was protected. Yet the Pennsylvania Supreme Court did so uncritically. Part of the problem may have been that the briefs did not discuss these deeply divisive issues at length, but this is likely due to uncertainty about whether they would count as “context” for true threat purposes.

Knox also shows how judges may underestimate the importance of new technology and online communication. The majority apparently did not find it significant that there was no clear evidence as to whether Knox actually posted the video or even knew of its posting. The majority also failed to extensively discuss the indirectness of the alleged threat. The rap was recorded, posted to Facebook by a third party, viewed by a police officer, then shown to the threat’s targets.202See supra Section II.D.1. Though not dispositive by any means, the presence of online media as the vehicle for the threat at least warranted greater judicial attention.

Given that the contextual inquiry is so crucial to protecting constitutional speech rights in true threat cases, this Note proposes a way to mitigate the potential bias and inexpertness hampering the contextual inquiry and to give notice to parties about what evidence is relevant.

B.     Proposed Four-Part Framework

Scholars have proposed various solutions to reduce the difficulty and unpredictability of asking courts to analyze context in true threat cases. Most commonly, these proposals involve identifying specific factors that courts should consider at some point in their contextual analyses.203See, e.g., Lyrissa Barnett Lidsky & Linda Riedemann Norbut, #I🔫U: Considering the Context of Online Threats, 106 Calif. L. Rev. 1885, 1923–25 (2018); Jordan Strauss, Context Is Everything: Towards a More Flexible Rule for Evaluating True Threats Under the First Amendment, 32 Sw. U. L. Rev. 231 (2003); Scott Hammack, Note, The Internet Loophole: Why Threatening Speech On-line Requires a Modification of the Courts’ Approach to True Threats and Incitement, 36 Colum. J.L. & Soc. Probs. 65, 98–99 (2002); Rothman, supra note 80, at 333–35. But a list of discretionary factors would do little to solve the inconsistency across cases. Cherry-picking from such lists is essentially the approach courts currently use, and it is far too messy.204See supra Section II.C. Moreover, there will inevitably be cases where some of the major factors are completely irrelevant and others where the record involves such atypical facts that a conventional list of factors would be utterly unhelpful. More drastic proposals, such as adding a new category of unprotected threatening speech205See, e.g., Jeremy C. Martin, Comment, Deconstructing “Constructive Threats”: Classification and Analysis of Threatening Speech After Watts and Planned Parenthood, 31 St. Mary’s L.J. 751 (2000). or enabling defendants to assert context as an affirmative defense,206Lidsky & Norbut, supra note 203, at 1925 (“The defense would work as follows. Once the accused is charged with making a criminal threat, the accused may invoke context as a defense to liability.”). have the potential to clarify this convoluted doctrine, but none have garnered consensus. Thus, in the spirit of a marketplace of ideas, this Note offers a new alternative: borrowing an analytical framework (rather than a list of factors) from another important First Amendment doctrine, namely, defamation.

Defamation, or false speech that tends to harm the reputation of its subject, is a well-established tort at common law.207See 8A Stuart M. Speiser, Charles F. Krause & Alfred W. Gans, The American Law of Torts § 29:2 (2020). Defamatory statements remain an unprotected category of speech, and the modern Supreme Court has devoted significant effort to aligning its boundaries with First Amendment principles.208See, e.g., N.Y. Times Co. v. Sullivan, 376 U.S. 254 (1964). The Court generally deems defamation penalties constitutional only when a plaintiff affirmatively proves that the speech at issue was verifiably false.209See, e.g., Phila. Newspapers, Inc. v. Hepps, 475 U.S. 767, 777 (1986) (“[A] private-figure plaintiff must bear the burden of showing that the speech at issue is false before recovering damages for defamation from a media defendant. To do otherwise could ‘only result in a deterrence of speech which the Constitution makes free.’” (quoting Speiser v. Randall, 357 U.S. 513, 526 (1958))). But, as it turns out, determining whether a particular statement is true or false is often easier said than done. For this reason, courts’ efforts to determine whether a defamatory statement is “false” provides a path for courts analogously seeking to determine whether a threat is “true.”

In the 1974 case Gertz v. Robert Welch, Inc., the Court made a sweeping declaration that would haunt it for decades: “Under the First Amendment there is no such thing as a false idea. However pernicious an opinion may seem, we depend for its correction not on the conscience of judges and juries but on the competition of other ideas.”210418 U.S. 323, 339–40 (1974). Lower courts interpreted this language to mean that statements of “opinion” were distinguishable from statements of fact, and so were categorically protected from any defamation action.211See Len Niehoff & Ashley Messenger, Milkovich v. Lorain Journal Twenty-Five Years Later: The Slow, Quiet, and Troubled Demise of Liar Libel, 49 U. Mich. J.L. Reform 467, 467 (2016) (noting “lower courts’ longstanding interpretation of Gertz v. Robert Welch as creating a separate constitutional privilege for expressions of opinion” (footnote omitted)); Robert D. Sack, Protection of Opinion Under the First Amendment: Reflections on Alfred Hill, “Defamation and Privacy Under the First Amendment, 100 Colum. L. Rev. 294, 315 (2000). Simply put, if a court understood a defendant’s speech to be an opinion rather than a factual assertion, the defendant could not be liable as a matter of law. While the Court repudiated this interpretation of Gertz sixteen years later in Milkovich v. Lorain Journal Co.,212See 497 U.S. 1, 18 (1990) (“[W]e do not think this passage from Gertz was intended to create a wholesale defamation exemption for anything that might be labeled ‘opinion.’”). That said, the actual effect of Milkovich on the Gertz fact/opinion distinction is still unclear and is a topic of scholarly debate in its own right. See Niehoff & Messenger, supra note 211. the efforts of lower courts to draw the line between fact and opinion during those intervening years are nonetheless relevant points of comparison for the purposes of this Note.

1.       Ollman v. Evans: The Template Test

During the time between Gertz and Milkovich, the D.C. Circuit issued a decision that was, for a time, the leading approach to distinguishing protected statements of opinion from unprotected false statements of fact. Ollman v. Evans involved a university professor who sued two newspaper columnists for publishing a nationally syndicated column suggesting that he was using the classroom to politically indoctrinate students with his Marxist beliefs.213Ollman v. Evans, 750 F.2d 970, 971–73 (D.C. Cir. 1984) (en banc). The columnists argued that they could not be held liable because the column reflected only their “opinions” about the professor’s work.214Id. In his plurality opinion for the en banc court, Judge Starr stressed the impossibility of identifying “a bright-line or mechanical distinction” between facts and opinions while also emphasizing that ad hoc decisions and overly complex rules would be useless in protecting the speech rights at stake.215Id. at 977–78. The same can be said for true threat analyses.

Given the impossibility of a simple bright-line rule, Judge Starr set out a four-part test for evaluating the totality of the circumstances surrounding an allegedly defamatory statement and assessing whether an average reader would regard the statement as fact or opinion.216Id. at 979. The first step was to “analyze the common usage or meaning of the specific language of the challenged statement itself.”217Id. Under this step, statements that could be interpreted many different ways would be unlikely to support an action for defamation.218Id. The second step was whether the statement was “capable of being objectively characterized as true or false.”219Id. The third step was to look at the “unchallenged language surrounding the allegedly defamatory statement,” such as cautionary or interrogative language, to examine how it could “influence the average reader’s readiness to infer that a particular statement has factual content.”220Id. The final step was to “consider the broader context or setting in which the statement appears”—for example, whether it appeared in the opinion section of a newspaper as opposed to the front page.221Id. at 979, 984. Several other courts embraced this formulation, sometimes short-handing the four steps as (1) specificity, (2) verifiability, (3) literary context, and (4) public or social context.222See, e.g., Price v. Viking Penguin, Inc., 881 F.2d 1426, 1432 (8th Cir. 1989); Henderson v. Times Mirror Co., 669 F. Supp. 356, 359 (D. Colo. 1987), aff’d, 876 F.2d 108 (10th Cir. 1989); Steinhilber v. Alphonse, 501 N.E.2d 550, 555–56 (N.Y. 1986).

2.       Tweaking the Ollman Test for True Threats

The goal of the Ollman test was to organize an all-encompassing analysis of the factual record into a logical process to answer a single question: Is the challenged speech protected by the First Amendment? Any true threat test must address the same question. Thus, the Ollman test—carefully crafted by an en banc D.C. Circuit plurality and further refined by other courts following its lead—is a more solid foundation for true threat analyses than any ad hoc list of factors that appears in true threat precedent. But adjustments to the Ollman test will of course be necessary given that its ultimate distinction between “fact” and “opinion” differs from the distinction of “true threat” versus—well, everything else.

The first Ollman step of “specificity” translates quite well to true threat analysis. If the common meaning of the statement does not facially reflect a specific threat to harm someone or something, then the inquiry should stop there and the defendant should not be convicted. Watts, the case that started it all, can serve as an example. The key statement at issue there was “[i]f they ever make me carry a rifle the first man I want to get in my sights is L. B. J.”223Watts v. United States, 394 U.S. 705, 706 (1969). This statement was, at least on its face, a specific threat to shoot President Lyndon B. Johnson. If Watts had instead said “if they ever make me carry a rifle, I’ll go berserk,” there would have been no specific threat.224Putting aside the fact that the statute used to charge Watts criminalized only threats to the president. Knox, too, exemplifies the importance of specificity. If Knox had not used the individual officers’ names in the rap, it is unlikely he could have been constitutionally punished for the speech, violent and angry as it may have been.

Second in the Ollman framework is “verifiability.” In literal terms, this step does not make much sense in the context of threats, as a threat is by definition predictive of something that has not yet occurred rather than descriptive of something that has already occurred. One possible way to adapt “verifiability” would be to ask whether the defendant actually intended or made efforts to carry out the threat. Evidence of such an intent or effort would theoretically render the seriousness of the threat more “verifiable.” But the Supreme Court explicitly rejected the relevance of the defendant’s intent to carry out the threat in Black.225Virginia v. Black, 538 U.S. 343, 359–60 (2003). Instead, “feasibility” could fill in for verifiability in threat analyses, with courts asking whether the threat would seem feasible to a reasonable person. This inquiry would allow a court to incorporate something like the defendant’s known possession of a firearm (as in Knox) into the analysis, because such a fact would make the threat seem more feasible to a reasonable person. Likewise, the commonly recurring factor of listeners’ reactions could be used as evidence to show whether reasonable people understood the threat to be feasible. For instance, the fact that Watts’s listeners laughed might serve as evidence of their understanding that, even if he were drafted, he was extremely unlikely to get President Johnson “in [his] sights.”226Watts, 394 U.S. at 706–07.

Third, Ollman looked to “literary context.” This step essentially requires looking beyond the few words that are allegedly threatening (or, in the Ollman situation, allegedly defamatory) to everything else the defendant said. Again, the importance of taking this step in an analysis of the context surrounding a threat is fairly obvious. Indeed, interpreting speech based on the words around it is the plainest meaning of the Supreme Court’s instruction to interpret threats in context. Many courts already employ some form of inquiry into other statements by the defendant. For example, the Watts factor of whether speech is “conditional” would come in under this step. The court in Knox likewise applied something like this step when it emphasized the overall violence of the words surrounding the actual threats in the rap, and held that this language supported classifying the threat as “true.”227Commonwealth v. Knox, 190 A.3d 1146, 1158 (Pa. 2018). That said, the court also could have looked at the generalized nature of the rap’s other lyrics as evidence that Knox’s statements expressed anger about the police as a group, not threats to harm two individual officers.

Fourth, Ollman considered social context, which is of paramount importance for identifying true threats. This step most squarely incorporates the key consideration of a statement’s location in the “political arena” noted by the Supreme Court in Watts.228Watts, 394 U.S. at 708. Political speech traditionally receives maximal First Amendment protection,229See, e.g., Citizens United v. Fed. Election Comm’n, 558 U.S. 310, 340 (2010) (“[P]olitical speech must prevail against laws that would suppress it, whether by design or inadvertence.”). so it is especially crucial that courts consider any link between the alleged threat and politics. In Knox, the majority dismissed out of hand any possibility that the rap’s antipathy toward police might be related to political issues,230Knox, 190 A.3d at 1158. despite the fact that rappers often imbue their songs with political meaning.231Cf. Kerr, supra note 164, at 195 (“Rap, much like your typical political speech, functions as a First Amendment ‘safety valve,’ allowing rappers to vent anti-authority aggression and to lament systemic racism and police brutality. We can hear rap, even rap made of violent, inchoate lyrics, about ‘the system’ and law enforcement to be a covered form of political speech. For example, NWA and Ice-T have been valorized by the Art World. But even if they had self-published their anti-police raps their message could be conceptualized as political speech. In short, without the historical context, a judge might not be able to identify or understand a rap’s subtextual political message.” (footnote omitted)). A more extensive discussion of the political context may not have altered the result in Knox,232Indeed, Justice Wecht directly addressed the possible link between the rap and political issues such as police brutality, yet he still voted to uphold the conviction. Knox, 190 A.3d at 1167 (Wecht, J., concurring in part and dissenting in part). but it still would have been valuable to ensuring the maintenance of Knox’s speech rights.

This fourth step should also account for “social” features of the speech, such as the medium of communication, the relationship between the defendant and the alleged victim, and its artistic value. The medium of communication will be especially salient in the modern world of social media, where the particular forum in which a threat is posted may have major repercussions on how threatening it really is. In Knox, for instance, the court would have needed to grapple more extensively with the characteristics of YouTube and Facebook that made it less likely that Knox recorded the rap intending for the police officers to see it.

Analysis of the social relationship between Knox and the threatened officers would also come in under this step. It is true that Knox had prior interactions with the officers, but the court said little about the nature of those interactions other than that the officers believed that Knox had a firearm with him during his first arrest.233Id. at 1160 (majority opinion). The history of police brutality against the Black community in Pittsburgh would also have been relevant to the analysis of social context,234See Jody DiPerna & Elaine Frantz, Historical Context: Violence Occurring Against Black Pittsburghers Today Has Been Happening for More Than a Century-and-a-Half, Pitt. Current (Sept. 16, 2020), []. painting a fuller picture of what may have motivated Knox to speak so violently against police other than an intent to induce fear.

This fourth step would also have given the Knox court an explicit mandate to directly address arguments raised by amici about the social context of rap as an artistic genre and Knox’s place within it.235Cf. Kerr, supra note 164, at 194–99 (discussing rap, and Knox’s rap in particular, as art). The court in Knox made some effort to do this,236Knox, 190 A.3d at 1160. but it did not give any real credence to the idea that it might take some extra work to distinguish true threats from protected speech when the speech is disseminated in a rap song. The court’s conclusion that couching threats in rap or musical lyrics cannot immunize them from prosecution does not, on its own, negate the relevance of the fact that Knox expressed himself through an artistic medium.

In sum, remaking the Ollman test to apply to true threats provides an administrable four-part test for appellate courts conducting an independent review of the record—including all relevant context—to determine whether a given statement constitutes a “true threat.” Put simply, this Note contends that appellate courts asked to identify true threats should examine four criteria: (1) the specificity of the threat, based on the actual words used; (2) the facial feasibility of the threat; (3) the “literary context,” or the language surrounding the alleged threat; and (4) the broader social context in which the threat was made, including the medium through which it was communicated and any political connotations it may have had.

Even applying this test, and assuming resolution of the threshold questions of intent and standard of review, true threat cases will never become categorically “easy,” nor will application of the test necessarily lead to a different outcome in close cases. But the fact that the job is difficult does not mean courts should not do it rigorously; to the contrary, they must take extra care to prevent threat prosecutions from making an end-run around the First Amendment. The test would standardize courts’ approaches to analyzing context that may dictate the constitutionality of a true threat conviction, giving judges and speakers alike crucial guidance about what makes some speech punishable and other speech protected.


Though the First Amendment does not categorically forbid the government from criminalizing threats, it does call for tight limits on the reach of such laws, which by their very nature punish “pure speech.” Unfortunately, the Supreme Court has been vague about what those limits are. To date, the major battleground of true threat jurisprudence has been whether a subjective- or objective-intent standard is proper, but the doctrine’s problems are not fully resolved once a lower court picks a side in that fight. As Knox illustrated, neither test will adequately protect speech unless appellate courts carefully and independently analyze the contextual factors that indicate whether the defendant’s speech fits into the unprotected “true threat” category.

To safeguard free speech rights while still sanctioning dangerously threatening speech, the Supreme Court should endorse a four-part, Ollman-like framework. Such a framework would give much-needed guidance to lower courts on how to interpret the context that is vital to separating true threats from protected speech. As is, the slapdash state of the true threat doctrine belies America’s “profound national commitment” to free speech.

*  J.D., May 2021, University of Michigan Law School. Thank you to Professor Len Niehoff for his invaluable feedback, mentorship, and guidance on all things First Amendment. I’m also grateful to the Volume 120 Notes Office for their thoughtful edits. And of course, my deepest thanks to fellow Volume 119 Notes Editors Kate Markey, Conor Bradley, Veronica Portillo-Heap, Brian Remlinger, and Ingrid Yin for making me a better writer and for being all-around brilliant people.