Opinions of the Colorado Supreme Court are available to the public and can be accessed through the Judicial Branch’s homepage at http://www.courts.state.co.us. Opinions are also posted on the Colorado Bar Association’s homepage at http://www.cobar.org.

ADVANCE SHEET HEADNOTE June 1, 2020

No. 17SC116, People in Interest of R.D.—First Amendment—True Threats— Social Media.

The supreme court reviews whether the court of appeals erred in

determining that threatening messages the juvenile defendant posted on Twitter

were protected speech under the First Amendment. In so doing, the court refines

its earlier statements of the general framework for distinguishing a true threat

from constitutionally protected speech and offers specific guidance for applying

that test to statements communicated online.

The court holds that a true threat is a statement that, considered in context

and under the totality of the circumstances, an intended or foreseeable recipient

would reasonably perceive as a serious expression of intent to commit an act of

unlawful violence. In determining whether a statement is a true threat, a

reviewing court must examine the words used, but it must also consider the

context in which the statement was made. Particularly where the alleged threat is

communicated online, the contextual factors courts should consider include, but are not limited to (1) the statement’s role in a broader exchange, if any, including

surrounding events; (2) the medium or platform through which the statement was

communicated, including any distinctive conventions or architectural features;

(3) the manner in which the statement was conveyed (e.g., anonymously or not,

privately or publicly); (4) the relationship between the speaker and recipient(s);

and (5) the subjective reaction of the statement’s intended or foreseeable

recipient(s).

The court reverses the judgment of the court of appeals and remands with

instructions to return the case to the juvenile court to reconsider the adjudication

under the refined framework. The Supreme Court of the State of Colorado 2 East 14th Avenue • Denver, Colorado 80203

Supreme Court Case No. 17SC116 Certiorari to the Colorado Court of Appeals Court of Appeals Case No. 14CA1800

Petitioner:

The People of the State of Colorado,

In the Interest of

Respondent:

R.D.

Judgment Reversed en banc June 1, 2020

Attorneys for Petitioner: Philip J. Weiser, Attorney General Joseph G. Michaels, Assistant Attorney General Denver, Colorado

Attorneys for Respondent: Megan A. Ring, Public Defender James S. Hardy, Deputy Public Defender Denver, Colorado

JUSTICE MÁRQUEZ delivered the Opinion of the Court. ¶1 The First Amendment’s protection of speech is robust, but not absolute: it

does not, for example, safeguard the utterance of a “true threat.” The task of

identifying a true threat has been complicated by the advent of social media. At

the same time, the proliferation of online expression has amplified the potential

for threatening words to cause harm. This case, which stems from a late-night

argument on Twitter among several high school students, requires us to confront

this changed communication landscape and to refine the applicable framework for

distinguishing a true threat from constitutionally protected speech where that

communication occurs in the cyber arena.

¶2 R.D., a juvenile, was adjudicated delinquent for harassment by

communication under section 18-9-111(1)(e), C.R.S. (2013), based on tweets he

directed at another student during a heated exchange that took place in the wake

of a local school shooting. The narrow question before us is whether R.D.’s

statements were “true threats.” If so, then application of the statute to his conduct

did not violate his First Amendment right to free speech.

¶3 In light of U.S. Supreme Court case law, we refine our earlier statements of

the framework for distinguishing a true threat from constitutionally protected

speech. In addition, we take the opportunity to provide guidance for applying

that test to statements communicated online.

2 ¶4 We hold that a true threat is a statement that, considered in context and

under the totality of the circumstances, an intended or foreseeable recipient would

reasonably perceive as a serious expression of intent to commit an act of unlawful

violence.1 In determining whether a statement is a true threat, a reviewing court

must examine the words used, but it must also consider the context in which the

statement was made. Particularly where the alleged threat is communicated

online, the contextual factors courts should consider include, but are not limited

to (1) the statement’s role in a broader exchange, if any, including surrounding

events; (2) the medium or platform through which the statement was

communicated, including any distinctive conventions or architectural features;

(3) the manner in which the statement was conveyed (e.g., anonymously or not,

privately or publicly); (4) the relationship between the speaker and recipient(s);

and (5) the subjective reaction of the statement’s intended or foreseeable

recipient(s).

1 We need not resolve today whether the test for true threats under the First Amendment also requires consideration of the speaker’s subjective intent to threaten the victim(s). But even assuming it does, the statutory provision at issue required the State to prove beyond a reasonable doubt that the communication here was made “in a manner intended to . . . threaten bodily injury.” § 18-9-111(1)(e). 3 ¶5 Because neither the juvenile court nor the court of appeals had the benefit

of the framework we adopt today, we reverse the judgment of the court of appeals

and remand with instructions to return the case to the juvenile court to reconsider

the adjudication applying this refined test.

4 I. Facts and Procedural History

¶6 In December 2013, a shooting took place at Arapahoe High School that took

the life of a female student and the male student shooter. A few days later, a

student from Thomas Jefferson High School (“TJ”), a school in a neighboring

district, posted on Twitter2 a photo of a banner conveying TJ’s support for

Arapahoe. A student from Littleton High School, which is in the same school

district as Arapahoe, tweeted3 in response that kids from TJ did not care about the

shooting because it happened outside their district. A.C., another TJ student and

a friend of the original poster, soon got involved because he believed the Littleton

student was disrespecting his friend. J.W., A.C.’s friend and fellow TJ student,

also got involved, and the group conversation eventually took on a “TJ versus

Littleton” character. The Littleton student “mentioned”4 the handles, or

2Twitter is a “real-time information network that lets people share and discuss what is happening at a particular moment in time through the use of ‘tweets.’” Dimas-Martinez v. State,

, 243 n.3 (Ark. 2011). 3 A tweet is a message posted to Twitter that might contain text or other media. A tweet appears on the sender’s profile page and may appear on the feed, or timeline, of anyone following the sender. About Different Types of Tweets, Twitter, https://help.twitter.com/en/using-twitter/types-of-tweets [https://perma.cc/ 8ZBR-H79E]. The word “tweet” is also used as a verb to describe the act of posting a message on Twitter. See, e.g., How to Tweet, Twitter, https://help.twitter.com/ en/using-twitter/how-to-tweet [https://perma.cc/9CQ6-3BYE]. 4A “mention” is a tweet that contains another account’s Twitter username, or “handle,” preceded by the “@” symbol. When a user’s handle is mentioned, the

5 usernames, of R.D. and another friend from his school, bringing them into the

exchange.

¶7 As we discuss further below, the record provides an incomplete picture of

the students’ back-and-forth. But it does reveal that R.D. posted the following

messages:

• @[A.C.]5 you a bitch, ill come to Tgay and kill you nigga.6

• @[A.C.] I don’t people who aren’t worth my time. If I see your bitch ass outside of school you catching a bullet bitch.

• @[A.C.] nigga you don’t even know me. Mf I don’t even know were tf your lame bitch ass school is. You a bitch talking shit on here

• @[A.C.] all you fuck niggas will get your ass beat real shit.

• You fuck with the wrong person leave you ass in a body bag.

• @[J.W.] @[A.C.] don’t give af bruh. Don’t even know you niggas and you talking shit.

• @[A.C.] you think this shit a game, I’m not playing. I don’t fight fuck boys and I don’t twitter beef.

user receives notification of the tweet, but the tweet does not appear on the user’s public profile. 5 For purposes of this opinion, we have replaced the students’ Twitter handles with

their initials. 6 We reluctantly reproduce this racial slur and other pejorative terms from the record to give an uncensored account of the facts.

6 R.D. also posted a photograph of a handgun resting beside approximately fifty

cartridges, along with the message, “@[A.C.] this all I’m saying[.] We don’t want

another incident like Arapahoe. My 9 never on vacation.”

¶8 After this, the record reveals that R.D. and A.C. tweeted as follows:7

• A.C.: @[R.D.] you ain’t never shot no one so sit down and get off google images bruh

• R.D.: @[J.W.] @[A.C.] idgaf my @ name should have not been in this shit. You fucked with the wrong one.

• R.D.: @[A.C.] I don’t even know where tf your school at. I’m not even from Colorado. Trust me I’m not afraid to shoot.

• R.D.: @[A.C.] fuck you and your gay ass school. Don’t worry nigga, I’ll see you little hoes tomorrow.

• A.C.: @[R.D.] shoot then pussy.

• A.C.: @[R.D.] you are all talk so go the fuck to bed come up to TJ and get slept.8 Fuck boy.

• R.D.: @[A.C.] haha alright hoe, we’ll see whose a bitch tomorrow.

• R.D.: @[A.C.] I’m not about to fight you broke bitch. Let me catch you away from school you is a dead man. Goodnight hoe.

• A.C.: @[R.D.] 3950 S. Holly street. I’ll see u tomorrow fuck boy

7 These messages are set out in the order in which R.D. either tweeted or “retweeted”—i.e., shared—them. It is therefore possible that messages authored by A.C. and retweeted by R.D. were originally posted by A.C. in a somewhat different order than represented here. 8 A.C. later testified that “get slept” usually means to fight or get knocked out.

7 ¶9 Based on these tweets, the People filed a petition in delinquency charging

R.D. with harassment under section 18-9-111(1)(e), C.R.S. (2013). As relevant here,

that provision prohibits harassment through certain forms of digital

communication:

(1) A person commits harassment if, with intent to harass, annoy, or alarm another person, he or she . . .

(e) [i]nitiates communication with a person, anonymously or otherwise, by . . . text message, instant message, computer, computer network, or computer system in a manner intended to . . . threaten bodily injury . . . .

§ 18-9-111(1)(e). R.D. moved to dismiss the charge, contending that his statements

were protected by the First Amendment to the U.S. Constitution and article II,

section 10 of the Colorado Constitution.9

¶10 At a hearing on the motion, the prosecutor argued that under Virginia v.

Black,

(2003), and Watts v. United States,

(1969) (per

curiam), the government may constitutionally regulate speech that constitutes a

“true threat.” Speech constitutes a true threat, she contended, “when an

individual is intending to threaten bodily harm.”

9In this initial motion to dismiss, R.D. contended that his statements could not be constitutionally regulated because (1) they were made in a public forum and (2) they could not be considered “fighting words.” Because R.D.’s constitutional claim in this case now focuses on whether R.D.’s tweets constituted “true threats,” we restrict our description of the facts and procedural history to this claim.

8 ¶11 In arguing that R.D.’s tweets fell into this category of unprotected speech,

the prosecutor emphasized “the social context” of the statements, noting they were

sent four days after the Arapahoe High School shooting. She stated that police

officers contacted the students who had read the tweets, all of whom said they

were afraid. She further observed that such fear made sense given that a student

had so recently been shot. Finally, she posited that true threats such as R.D.’s need

to be regulated to “protect people’s feeling of safety.”

¶12 The trial court denied R.D.’s motion to dismiss. In a bench ruling, the court

concluded that R.D.’s “particular type of speech is not protected under the First

Amendment.” The court noted it had “consider[ed] the argument of counsel and

the factors the [c]ourt is to consider,” but did not identify what those factors were.

Based on its conclusion that R.D.’s tweets were not protected speech, the court also

found that section 18-9-111(1)(e) is not unconstitutional as applied to R.D.10

A. Trial

¶13 At trial, A.C.’s and J.W.’s testimony revealed that the screenshots and

printouts submitted in evidence of R.D.’s tweets painted an incomplete picture of

the conversation as it occurred on Twitter. The prosecution’s exhibits consisted of

10R.D. later filed a supplement to his motion to dismiss alleging for the first time that section 18-9-111(e) is facially overbroad. The juvenile court never ruled on this supplementary motion.

9 screenshots of some of the messages R.D. authored, but captured almost none of

the other students’ comments. The defense supplemented this one-sided view of

the conversation by introducing a print-out of R.D.’s Twitter profile, which

documented both R.D.’s own tweets and messages by others in the conversation

that R.D. retweeted.11

¶14 Aside from explaining to the adults in the room how Twitter works, A.C.

and J.W. testified to their reactions to the exchange. A.C. testified that he

construed R.D.’s tweets directed at him as threats; that he did not think R.D. was

kidding; and that he understood R.D.’s post containing the picture of a handgun

to convey a threat to his life. Yet on cross-examination, A.C. acknowledged that

he responded to that post by tweeting that R.D. should “get off google images”

because he thought the picture of the handgun was one R.D. had merely

downloaded from the internet. He also admitted that he tweeted the address of

his school.

11 Even this document, however, did not include tweets or private messages by other users that may have been part of the conversation, but that R.D. did not share on his public feed. Indeed, many of R.D.’s tweets in the print-out displayed a link to “view conversation,” which, if clicked in electronic form, would have revealed other replies in the same exchange. But whatever those other messages said, they are not part of the record.

10 ¶15 J.W.’s testimony was similarly inconsistent. When asked on direct about

R.D.’s “threat to kill,” J.W. testified that he did not take it “as a joke” but that he

also did not “take it serious.” He said the message was “a little intimidating.” He

also testified that students on “both sides” were throwing around insults and

talking about physically fighting and that he “didn’t really take anything as being

very serious.”

¶16 In a bench ruling, the juvenile court adjudicated R.D. delinquent. The court

acknowledged that the early part of the Twitter exchange did not establish an

intent to alarm under the harassment statute “because [the students] were both

engaging in that type of conversation.” The court analogized this early portion of

the exchange to a schoolyard fight where everyone is trying to prove they are

“bigger[,] better[,] and meaner” than their peers.

¶17 But the conversation crossed the line, the court found, when R.D. posted the

picture of the handgun. Referring again to the schoolyard brawl analogy, the court

likened that moment to R.D. lifting his shirt to show that he was armed. The court

reasoned that a brawl and its attendant displays of bravado usually cause “no

harm, no foul” if broken up. But when someone in such a face-to-face interaction

says they have a gun, the dynamic becomes menacing. The court found that R.D.’s

message was similar to such a display, although it acknowledged that the tweet

was “different” because the students were not face-to-face.

11 ¶18 Turning to the statute, the court found beyond a reasonable doubt that R.D.

“initiate[d] communication” over a computer network or computer system. It

then considered whether R.D. “inten[ded] to threaten” anyone. The court

observed that J.W. and A.C. testified that they didn’t want to joke about guns, but

that they didn’t take R.D.’s messages seriously, and pointed out that A.C. told R.D.

to “get off google images” and volunteered to him his school’s address. But the

court concluded that the boys’ subjective reactions were irrelevant because the

statute does not require that the recipient actually feel threatened or that actual

bodily injury occur.

¶19 Relying specifically on R.D.’s post of the picture of the handgun, the court

concluded beyond a reasonable doubt that R.D. “initiate[d] communication with

a person by computer network, data network, or computer system in a manner

intended . . . to threaten bodily injury or property damage to [A.C.] and [J.W.], in

violation of [section] 18-9-111(1)(e).” In its ruling, the court made no mention of

the First Amendment and did not opine on the statute’s alleged

unconstitutionality, either facially or as applied to R.D.’s tweets.

¶20 The court sentenced R.D. to write an essay demonstrating that he

understood the challenges of online communication. R.D. submitted the essay to

the court’s satisfaction.

12 B. Appeal

¶21 R.D. appealed, arguing, as relevant here, that his adjudication should be

vacated because application of section 18-9-111(1)(e) to his speech on Twitter

violated his First Amendment right to free speech.12 The People responded that

R.D.’s tweets were true threats and therefore unprotected speech.

¶22 The court of appeals agreed with R.D. and reversed and remanded with

directions to vacate the adjudication and dismiss the proceeding. People in Interest

of R.D.,

, ¶¶ 1, 6, __ P.3d __.

¶23 The court began by acknowledging that the government may regulate

certain unprotected categories of speech, such as true threats. Id. at ¶ 9. The court

defined a “threat” as a “statement of purpose or intent to cause injury or harm to

the person, property, or rights of another, by committing an unlawful act.” Id. at

¶ 10 (citing People v. McIntier,

(Colo. App. 2005)). But a “true

threat” for purposes of the First Amendment, the court explained, “is not merely

talk or jest.”

It is evaluated “in the context in which [it was] spoken or written,”

12 R.D. also challenged the sufficiency of the evidence that he “initiate[d]” communication under 18-9-111(1)(e) and argued that statements he made to his school’s resource officer and vice principal should have been suppressed. Neither of these bases for appeal is before us today.

13 and “by whether those who hear or read the threat reasonably consider that an

actual threat has been made.”

(quoting McIntier,

).

¶24 To determine whether R.D.’s statements constituted true threats, the court

of appeals considered both the plain import of the words used and the context in

which the statements were made, including (1) to whom the statements were

communicated; (2) the manner in which the statements were communicated; and

(3) the subjective reactions of those whom the statements concerned.

(citing

People v. Stanley,

(Colo. App. 2007), and Watts,

).

¶25 Applying that framework, the court concluded that R.D.’s tweets were not

true threats because they did not constitute “a serious expression of an intent to

commit an act of unlawful violence to a particular individual or group of

individuals.”

at ¶ 11 (quoting Black,

). The court reasoned that

although the language of R.D.’s tweets was violent and explicit, the context in

which the statements were made mitigated their tone in three ways.

¶26 First, the court observed that R.D.’s tweets made clear that he did not know

A.C. personally and did not know where A.C.’s school was located. Id. at ¶ 12. In

addition, R.D. never referred to A.C. by name, instead addressing him only by his

Twitter handle. Id.

14 ¶27 Second, the court observed that while R.D. used “@” to direct many of his

tweets to A.C., he did not send those messages privately to A.C. alone; instead,

they were in public view on R.D.’s profile page. Id. at ¶ 13.

¶28 Finally, the court observed that A.C.’s reactions to R.D.’s tweets showed that

“he did not view the statements as true threats when they were received.” Id. at

¶ 15. In particular, the court found significant that when R.D. indicated he did not

know where TJ was located, A.C. volunteered the address and tweeted, “I’ll see

you tomorrow fuck boy”; “you are all talk so go the fuck to bed come up to TJ and

get slept”; and “shoot then pussy.” Id. at ¶ 14. The court also took note of A.C.’s

response to the photo of the handgun: “you ain’t never shot no one so sit down

and get off google images bruh.” Id. The court gleaned from these tweets that

A.C. did not appear threatened and did not take precautionary measures to protect

himself from R.D. Id.

¶29 Based on these contextual factors, the court concluded that R.D.’s tweets did

not constitute true threats. Id. at ¶ 16. Accordingly, it held that application of

section 18-9-111(1)(e) to R.D.’s conduct violated his First Amendment rights. Id. at

¶ 21. We granted the People’s petition for a writ of certiorari to review whether

the court of appeals erred in determining that R.D.’s online statements are

15 protected by the First Amendment.13 In so doing, we must address the applicable

legal standard, which, as the parties’ briefs acknowledge, has been the subject of

some debate.

II. Analysis ¶30 The government’s power to regulate speech is constrained by the First

Amendment, which provides that “Congress shall make no law . . . abridging the

freedom of speech.” U.S. Const. amend. I; see Gitlow v. New York,

(1925) (incorporating First Amendment against the states).14 In this case, R.D. was

adjudicated delinquent for conduct that, if committed by an adult, would violate

section 18-9-111(1)(e). At the time, that provision stated as follows:

(1) A person commits harassment if, with intent to harass, annoy, or alarm another person, he or she . . .

(e) [i]nitiates communication with a person, anonymously or otherwise, by . . . text message, instant message, computer,

13 We granted the People’s petition for a writ of certiorari to review the following issue: Whether the court of appeals erred in determining that the defendant’s comments, made on Twitter, were protected by the First Amendment. 14As previously noted, R.D. also raised a claim under Colorado’s counterpart to the First Amendment. See Colo. Const. art. II, § 10. But because he does not argue that a different analysis applies under the state constitution, we discuss only his First Amendment claim.

16 computer network, or computer system in a manner intended to . . . threaten bodily injury . . . .

§ 18-9-111(1)(e).15

¶31 Because this provision regulates pure speech, we must review the

constitutionality of its application to R.D.’s tweets “with the commands of the First

Amendment clearly in mind.” Watts,

.

15 The provision was amended effective July 1, 2015, and now reads as follows: (1) A person commits harassment if, with intent to harass, annoy, or alarm another person, he or she . . . (e) Directly or indirectly initiates communication with a person or directs language toward another person, anonymously or otherwise, by telephone, telephone network, data network, text message, instant message, computer, computer network, computer system, or other interactive electronic medium in a manner intended to harass or threaten bodily injury or property damage, or makes any comment, request, suggestion, or proposal by telephone, computer, computer network, computer system, or other interactive electronic medium that is obscene . . . . § 18-9-111(1)(e), C.R.S. (2019). The Act amending subsection (1)(e) also added a new subsection (8), which provides that “[section 18-9-111] is not intended to infringe upon any right guaranteed to any person by the first amendment to the United States constitution or to prevent the constitutionally protected expression of any religious, political, or philosophical views.” § 18-9-111(8). Because R.D.’s charged conduct predated these alterations to the statutory language, we do not consider them in our analysis.

17 ¶32 The narrow question before us is whether R.D.’s tweets were “true

threats.”16

¶33 To begin, we clarify that the protection of free speech does not vary by

medium of expression and, accordingly, we set forth background First

Amendment principles and true threats jurisprudence to guide our analysis. Next,

we acknowledge that the advent of social media has complicated the constitutional

inquiry. To respond to today’s changed communication landscape, we both refine

our earlier statements of the general framework for distinguishing a true threat

from constitutionally protected speech and offer specific guidance for applying

that test to statements communicated online.

¶34 Having clarified the test for determining whether R.D.’s online statements

constitute true threats, we reverse the judgment of the court of appeals and

remand with instructions to return the case to the juvenile court to reconsider the

adjudication applying the framework we adopt today.

16 Although the parties’ briefs also discuss whether section 18-9-111(1)(e) is facially overbroad, we decline to address this issue, which neither the juvenile court nor the court of appeals addressed or ruled upon, and which, in any event, is beyond the scope of the question on which we granted certiorari review. In other words, we assume for purposes of this opinion that section 18-9-111(1)(e) proscribes only conduct that constitutes a true threat, at least insofar as it criminalizes what R.D. is charged with here. Accordingly, we limit our analysis to whether R.D.’s tweets constituted true threats.

18 A. The True Threats Exception

¶35 “There are certain well-defined and narrowly limited classes of speech, the

prevention and punishment of which have never been thought to raise any

Constitutional problem.” Chaplinsky v. New Hampshire,

, 571–72 (1942).

“[T]hese areas of speech can, consistently with the First Amendment, be regulated

because of their constitutionally proscribable content.” R.A.V. v. City of St. Paul,

(1992). One such category of unprotected speech is a “true

threat.”

1. Watts v. United States

¶36 The true threats doctrine originated in 1969 with Watts v. United States. In

that case, which arose during the Vietnam War, the eighteen-year-old defendant

was convicted under a federal statute forbidding any person from “knowingly and

willfully” making “any threat to take the life of or to inflict bodily harm upon the

President of the United States.”

. Watts had attended a public

anti-war rally on the grounds of the Washington Monument in Washington, D.C.,

where he joined a scheduled discussion group of young people who were mostly

in their teens and early twenties.

at 705–06. Watts told the group that he had

been drafted but would not report for his physical, and an Army Counter

Intelligence Corps investigator who was present overheard Watts state, “If they

ever make me carry a rifle the first man I want to get in my sights is L.B.J.”

at

19 706. Watts and others in the crowd reacted to this statement with laughter.

.

¶37 The Supreme Court held that the statute under which Watts was convicted

was “[c]ertainly . . . constitutional on its face,” given the government’s

overwhelming interest in protecting the safety of the President and allowing him

to perform his duties without interference from threats of physical violence.

But the Court also explained that because the statute “ma[de] criminal a form of

pure speech,” it had to be “interpreted with the commands of the First

Amendment clearly in mind.”

In particular, “a threat must be distinguished

from . . . constitutionally protected speech.”

The Court concluded that Watts’s

statement, “[t]aken in context,” including its “expressly conditional nature . . . and

the reaction of the listeners,” was mere political hyperbole that could not be

interpreted as a “true ‘threat’” under the statute.

.

2. Virginia v. Black

¶38 Thirty-four years later, in Virginia v. Black, the Supreme Court reaffirmed

that the First Amendment permits states to ban “true threats,” which it defined to

“encompass those statements 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.”

. The Court also clarified that

“[t]he speaker need not actually intend to carry out the threat,” because the true

20 threats exception exists to “protect[] individuals from the fear of violence,” “from

the disruption that fear engenders,” and from “the possibility that the threatened

violence will occur.”

at 359–60 (quoting R.A.V.,

).

¶39 At issue in Black was the constitutionality of a Virginia statute banning cross

burning done with intent to intimidate a person or group of persons. Id. at 347.

One provision of the statute treated cross burning as prima facie evidence of intent

to intimidate. Id. at 347–48. The Court explained that “[i]ntimidation in the

constitutionally proscribable sense of the word is a type of true threat, where a

speaker directs a threat to a person or group of persons with the intent of placing

the victim in fear of bodily harm or death.” Id. at 360. The Court held that the

statute did “not run afoul of the First Amendment insofar as it ban[ned] cross

burning with intent to intimidate,” id. at 361, but a plurality concluded that the

provision treating cross burning as prima facie evidence of such intent was

overbroad, reasoning that cross burning is sometimes protected expression, see id.

at 364–67 (plurality opinion).

3. The Post-Black Debate

¶40 Though Watts and Black made clear that the First Amendment does not

protect a “true threat,” the decisions resulted in a split of authority over how to

discern whether a particular statement amounts to one.

21 ¶41 A majority of jurisdictions have interpreted Black’s definition of a true

threat—a statement where the speaker “means to communicate a serious

expression of an intent to commit an act of unlawful violence”—to require only

that the speaker intended to make the statement. Under this reading, “means to”

modifies only the word “communicate.” See, e.g., United States v. Clemens,

(1st Cir. 2013) (requiring the speaker to intend to make the communication,

but not the threat). Courts adopting this view judge whether a statement

constitutes a true threat using an objective standard, asking how a reasonable

person would interpret the words.17 Proponents of an objective standard have

reasoned that a speaker’s lack of intent to threaten does nothing to reduce the

harms identified in Black that justify the exception of true threats from First

17The objective test has several variations, with some courts asking whether the statement is one a reasonable speaker would foresee would be interpreted as a serious expression of intention to inflict bodily harm, see, e.g., State v. Trey M.,

(Wash. 2016), some asking how a reasonable listener would construe the speech in context, see, e.g., United States v. White,

(4th Cir. 2012), and some considering both perspectives, see, e.g., Haughwout v. Tordenti,

(Conn. 2019) (requiring that “a reasonable person would foresee that the statement would be interpreted by those to whom the maker communicates the statement as a serious expression of intent to harm or assault” and that “a reasonable listener, familiar with the entire factual context of the defendant’s statements, would be highly likely to interpret them as communicating a genuine threat of violence rather than protected expression, however offensive or repugnant” (quoting State v. Krijger,

, 957, 963 (Conn. 2014))).

22 Amendment protection. Black,

; see also, e.g., United States v. Jeffries,

(6th Cir. 2012) (“Much like their cousins libel, obscenity, and

fighting words, true threats ‘by their very utterance inflict injury’ on the recipient.”

(quoting Chaplinsky,

)).

¶42 On the other hand, some courts have interpreted Black to require the speaker

to have the subjective intent to threaten. Under this reading, “means to” modifies

the entire phrase, “communicate a serious expression of an intent to commit an act

of unlawful violence.”18 Proponents of a subjective intent requirement have

tended to posit that a purely objective listener test would chill protected speech.

See Rogers v. United States,

, 47–48 (1975) (Marshall, J., concurring)

(arguing that “charging the defendant with responsibility for the effect of his

statements on his listeners . . . would have substantial costs in discouraging the

‘uninhibited, robust, and wide-open’ debate that the First Amendment is intended

18 See, e.g., United States v. Heineman,

(10th Cir. 2014); United States v. Cassel,

, 631–33 (9th Cir. 2005); State v. Boettger,

, 813–15 (Kan. 2019); see also Perez v. Florida,

(2017) (Sotomayor, J., concurring in denial of petition for writ of certiorari) (“Together, Watts and Black make clear that to sustain a threat conviction without encroaching upon the First Amendment, States must prove more than the mere utterance of threatening words—some level of intent is required. . . . 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.”).

23 to protect” (quoting N.Y. Times Co. v. Sullivan,

(1964))); United

States v. White,

(4th Cir. 2012) (Floyd, J., concurring in part and

dissenting in part) (“Under a purely objective test, speakers whose ideas or views

occupy the fringes of our society have more to fear, for their violent and extreme

rhetoric, even if intended simply to convey an idea or express displeasure, is more

likely to strike a reasonable person as threatening.”).19

4. Elonis v. United States

¶43 The U.S. Supreme Court seemed positioned to settle this debate in Elonis v.

United States,

(2015), where the Court had its first opportunity to

apply the true threats doctrine to statements communicated over social media,

specifically, posts the petitioner made on Facebook. There, the petitioner was

convicted under a federal statute that makes it a crime to transmit in interstate

commerce “any communication containing any threat . . . to injure the person of

another.”

at 2004 (quoting

(c) (2018)). The statute makes no

reference to a required mental state. Id. at 2008. The jury was instructed that to

convict Elonis, it had to find that he intentionally communicated a statement that

19Some have also reasoned that it would be unfair to penalize a speaker for the unintended consequences of their communication. See Leslie Kendrick, Free Speech and Guilty Minds,

, 1282 (2014).

24 a reasonable person would foresee would be regarded by the listener as a threat.20

Id. at 2004, 2007. The question before the Court was “whether the statute also

requires that the defendant be aware of the threatening nature of the

communication, and—if not—whether the First Amendment requires such a

showing.” Id. at 2004.

¶44 Ultimately, the Court resolved the case on statutory grounds and did not

consider any First Amendment issues. It concluded that reading in only a

“reasonable person” standard where a federal criminal statute is silent on the

required mental state would be inconsistent with the principle that “wrongdoing

must be conscious to be criminal.” Id. at 2012 (quoting Morissette v. United States,

(1952)). The Court held that a defendant’s purpose or knowledge

would satisfy this requirement but did not address whether recklessness would

also be sufficient.

Justices Alito and Thomas each wrote separately, criticizing

the majority’s failure to resolve the split in the circuit courts regarding the requisite

20 Specifically, the jury was instructed that [a] statement is a true threat when a defendant intentionally makes a statement in a context or under such circumstances wherein a reasonable person would foresee that the statement would be interpreted by those to whom the maker communicates the statement as a serious expression of an intention to inflict bodily injury or take the life of an individual. Elonis,

.

25 level of intent. See

(Alito, J., concurring in part and dissenting in part)

(“Did the jury need to find that Elonis had the purpose of conveying a true threat?

Was it enough if he knew that his words conveyed such a threat? Would

recklessness suffice? The Court declines to say. Attorneys and judges are left to

guess.”); see

(Thomas, J., dissenting) (“[The majority’s] failure to decide

throws everyone from appellate judges to everyday Facebook users into a state of

uncertainty.”). Thus, after Elonis, the proper test for true threats remains an

unsolved doctrinal puzzle.

¶45 A definitive framework for discerning a true threat has been similarly

elusive in Colorado, though our appellate courts have tended to embrace some

form of an objective test. For example, in People v. Baer,

(Colo. 1999),

this court appeared in passing to endorse a reasonable speaker test, parenthetically

describing a true threat as “one which a reasonable person would foresee would

be interpreted by the recipient as a serious threat to inflict death or bodily injury.”

. And in an earlier, widely cited special concurrence in People v. Janousek,

(Colo. 1994), then-Justice Mullarkey described the “critical inquiry”

under true threats jurisprudence as more of a reasonable listener test: “whether

those who hear or read the threat reasonably consider that an actual threat has

been made.”

(Mullarkey, J., specially concurring); see also R.D., ¶ 10

(reciting Janousek concurrence formulation); Stanley,

(same);

26 McIntier,

(same). More recently, the court of appeals division in

Stanley specifically rejected the contention that Black required more than an

objective test. See 170 P.3d at 786–89.

B. Distinguishing True Threats from Protected Speech in the Age of Social Media

¶46 This court has not had occasion to revisit the framework for assessing

whether a statement is a true threat since the U.S. Supreme Court issued its 2003

decision in Black. And as this case demonstrates, the ways in which technology

has transformed our everyday communication complicates the constitutional

inquiry. We take this opportunity to refine our test for discerning whether a

statement is a true threat, taking into account this altered communication

landscape.

¶47 First, it is foundational that the “‘basic principle[] of freedom of speech,

. . . like the First Amendment’s command, do[es] not vary’ when a new and

different medium for communication appears.” Brown v. Entm’t Merchs. Ass’n,

(2011) (quoting Joseph Burstyn, Inc. v. Wilson,

(1952)). That said, “[e]very medium of expression presents special First

Amendment problems which must be examined in the light of the circumstances

which are interwoven with the speech in issue.” People v. Weeks,

(Colo. 1979) (citing Joseph Burstyn, Inc., 343 U.S. at 502–03, and Kovacs v. Cooper,

(1949) (Jackson, J., concurring)). In this case, we are alert to the

27 competing concerns that “[s]ocial media make hateful and threatening speech

more common but also magnify the potential for a speaker’s innocent words to be

misunderstood.” Lyrissa Barnett Lidsky & Linda Riedemann Norbut, #I U:

Considering the Context of Online Threats,

, 1885 (2018).

¶48 Words communicated online and without the interpretive aid of body

language are easily misconstrued. Indeed, our reliance on nonverbal cues was

implicit in Chaplinksy, where the U.S. Supreme Court first articulated the “fighting

words” doctrine. There, the Court recognized that “[t]he English language has a

number of words and expressions which by general consent are ‘fighting words’

when said without a disarming smile.”

(emphasis added) (quoting

State v. Chaplinsky,

(N.H. 1941)). Modern replacements for such

cues, like emojis and gifs, often lack standard meaning and can be difficult to

interpret. Complicating things further, emojis may look different depending on

the sender’s or recipient’s operating system. For one example, an emoji that

resembles a toy squirt gun in a message sent on one platform may appear as a

revolver on a recipient device. Cf. Lidsky & Norbut, supra, at 1908 (explaining that

the gun emoji in the article’s title “looks like a space pistol on some platforms and

like a revolver on others”).

¶49 The chance of meaning being lost in translation is heightened by the

potential for online speech to be read far outside its original context. These days,

28 one needs no more than a whim and a smartphone to broadcast to a massive

audience. A message posted in Denver can reach New York, Tokyo, or Munich in

an instant. Indeed, the term “viral” is apt for the rapidity with which an online

statement can spread. A recipient might retransmit a message to audiences not

foreseeable to the original speaker. A message might be recirculated after an

intervening event that alters its impact. And online speech transmitted in the heat

of the moment—which, if uttered verbally, would not linger beyond the speaker’s

apology—might be archived and subjected to scrutiny years after the fact.

¶50 The risk of mistaking protected speech for a true threat is high. But so are

the stakes of leaving true threats unregulated. With the click of a button or tap of

a screen, a threat made online can inflict fear on a wide audience.

See, e.g., Julie Turkewitz & Jack Healy, ‘Infatuated’ with Columbine:

Threats and Fear, 20 Years After a Massacre, N.Y. Times (Apr. 17, 2019),

https://www.nytimes.com/2019/04/17/us/columbine-shooting-sol-pais.html

(reporting that “millions of parents, students, and educators across Colorado”

awoke on Columbine’s 20th anniversary to news of an individual’s alarming social

media posts and threats to friends and family, and that hundreds of schools across

the state closed in response). Indeed, a single online post can trigger the diversion

of significant law enforcement resources. See, e.g., United States v. Bradbury,

(7th Cir. 2017) (observing that defendant’s Facebook post

29 precipitated an extensive police investigation). Or such a threat may be directed

to a known and vulnerable victim in the privacy of their home. See Elonis,

(Alito, J., concurring in part and dissenting in part) (“Threats of

violence and intimidation are among the most favored weapons of domestic

abusers, and the rise of social media has only made those tactics more

commonplace.”). Online communication—in particular, the ability to

communicate anonymously—enables unusually disinhibited communication,

magnifying the danger and potentially destructive impact of threatening language

on victims. See Reno v. ACLU,

(1997) (O’Connor, J., concurring)

(“[C]yberspace allows speakers and listeners to mask their identities.”). In short,

technological innovation has provided apparent license and a ready platform to

those wishing to provoke terror.

¶51 Given this changed landscape, we are convinced that the various objective

tests previously articulated by this court and the court of appeals are insufficient

to distinguish “what is a [true] threat . . . from what is constitutionally protected

speech.” Watts,

. Judging a statement from the vantage point of a

“reasonable speaker” or “reasonable listener,” in our view, inadequately accounts

for potentially vast differences in speakers’, listeners’, and disinterested

fact-finders’ frames of reference. We therefore hold that a true threat is a statement

that, considered in context and under the totality of the circumstances, an intended

30 or foreseeable recipient would reasonably perceive as a serious expression of

intent to commit an act of unlawful violence.21 We believe that this refinement of

the objective standard strikes a better balance between giving breathing room to

free expression and protecting against the harms that true threats inflict.

¶52 In determining whether a statement is a true threat, a reviewing court must

examine the words used, but it must also consider the context in which the

statement was made. Particularly where the alleged threat is communicated

online, the contextual factors courts should consider include, but are not limited

to (1) the statement’s role in a broader exchange, if any, including surrounding

events; (2) the medium or platform through which the statement was

communicated, including any distinctive conventions or architectural features;

(3) the manner in which the statement was conveyed (e.g., anonymously or not,

privately or publicly); (4) the relationship between the speaker and recipient(s);

and (5) the subjective reaction of the statement’s intended or foreseeable

recipient(s).

21In the absence of additional guidance from the U.S. Supreme Court, we decline today to say that a speaker’s subjective intent to threaten is necessary for a statement to constitute a true threat for First Amendment purposes. But even assuming that the First Amendment requires proof of such subjective intent, the statute here required the government to show beyond a reasonable doubt that R.D. “initiate[d] communication . . . in a manner intended to . . . threaten bodily injury.” § 18-9-111(1)(e).

31 ¶53 Courts should start, of course, with the words themselves, along with any

accompanying symbols, images, and other similar cues to the words’ meaning.

Cf. United States v. Edwards, No. 2:17-CR-170,

, at *2 (S.D. Ohio Jan.

17, 2018) (in witness retaliation case, analyzing Facebook post that called

confidential informant a snitch and included laughing faces and a skull emoji).

This inquiry should include whether the threat contains accurate details tending

to heighten its credibility. See, e.g., Elonis, 135 S. Ct. at 2005–06 (noting the accuracy

of the details in defendant’s Facebook post conveying a threat against his wife,

including a diagram of her house and directions to “fire a mortar launcher . . . from

the cornfield behind it because of easy access to a getaway road” and “a clear line

of sight through the sun room”). It should also examine whether the speaker said

or did anything to undermine the credibility of the threat. See, e.g., Watts, 394 U.S.

at 707–08 (noting that petitioner’s threat to kill the President was made conditional

upon induction into the Armed Forces, an event petitioner vowed would never

occur).

¶54 Importantly, “what a defendant actually said is just the beginning of a

threats analysis.” Haughwout v. Tordenti,

(Conn. 2019). For example,

a veiled statement may carry a true threat. See, e.g., Jeffries,

(“[O]ne

cannot duck [a threats prosecution] merely by delivering the threat in verse or by

dressing it up with political (and protected) attacks on the legal system.”); Planned

32 Parenthood of Columbia/Willamette, Inc. v. Am. Coal. of Life Activists,

,

1062–63 (9th Cir. 2002) (en banc) (concluding that, viewed in context,

“Wanted”-style posters listing the names of doctors who had performed abortions

could be true threats); cf. Elonis,

(Alito, J., concurring in part and

dissenting in part) (“To hold otherwise would grant 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.”). On the other hand, words that are threatening on their face

may actually be just creative expression, jest, or hyperbole. See, e.g., Jeffries,

(“[A] song, a poem, a comedy routine or a music video is the kind

of context that may undermine the notion that the threat was real.”); Burge v.

Colton Sch. Dist. 53,

(D. Or. 2015) (concluding

eighth grader’s comment on Facebook that a teacher at school “need[ed] to be

shot” was reasonably understood to be merely a critique of the teacher’s skills);

State v. Boettger,

, 818 (Kan. 2019) (imagining a police protester

standing near police officers and quoting the lyrics of N.W.A.’s “Fuck tha Police,”

(Straight Outta Compton (Ruthless/Priority 1989)), “[t]ak[e] out a cop or two”).

In short, words matter. But so does context.

¶55 Particularly when evaluating online communication, courts should consider

whether the statement was part of a larger exchange, including surrounding

events. If so, the court should take note of the overall tone of that conversation, as

33 well as the origin of the allegedly threatening language—for example, whether it

was spontaneous or responsive to some other communication. It should also

consider how surrounding events may impact the statement’s tenor. United

States v. Voneida,

(3d Cir. 2009) (concluding that recency of

Virginia Tech shooting supported finding that student’s posts to his MySpace

page, including that “[s]omeday [he would] make the Virginia Tech incident look

like a trip to an amusement park,” were true threats). But see Watts,

(Douglas, J., concurring) (noting danger of policing alleged threats “under

circumstances when intolerance for free speech [is] much greater than it normally

might be” (quoting Note, Threatening the President: Protected Dissenter or Political

Assassin, Geo. L. J. 553, 570 (1969))).

¶56 Relatedly, the court should consider the medium or platform used to

communicate the alleged threat. First, the choice of medium itself may be

revealing. See, e.g., United States v. Bagdasarian,

, 1120–21 (9th Cir.

2011) (reasoning that posting violent messages about the President on financial

message board blunts the perception that the statements are true threats). And

evidence regarding prevailing norms in a particular genre or even internet

subforum may also help recast violent language in a less threatening light.

See, e.g., Bell v. Itawamba Cty. Sch. Bd.,

(5th Cir. 2014) (noting that

“hyperbolic and violent language is a commonly used narrative device in rap,

34 which functions to convey emotion and meaning—not to make real threats of

violence”). In the context of social media, the court should also consider the

platform’s distinctive architectural features, cf. Unsworth v. Musk,

No. 2:18-CV-08048-SVW-JC,

, at *6–7 (C.D. Cal. May 10, 2019) (in

defamation case, reasoning that Twitter’s 280-character limit rendered dubious the

notion that short-hand supports an inference that text in question was opinion

rather than fact), and conventions, see, e.g., Matter of Welfare of A.J.B.,

(Minn. 2019) (distinguishing direct messages from mentions on Twitter).

¶57 The manner in which the statement was conveyed may also provide insight.

For example, “a speaker’s anonymity could influence a listener’s perception of

danger.” Bagdasarian, 652 F.3d at 1120–21 (but concluding there was no reason in

that case to think the speaker’s anonymity made it more, rather than less, likely

that a violent post regarding the President was a serious threat). The directness of

the message may also be revealing. See, e.g., Elonis,

(Alito, J.,

concurring in part and dissenting in part) (“‘Taken in context,’ lyrics in songs that

are performed for an audience or sold in recorded form are unlikely to be

interpreted as a real threat to a real person,” whereas “[s]tatements on social media

that are pointedly directed at their victims . . . are much more likely to be taken

seriously.”); A.J.B.,

(Chutich, J., concurring in part and

dissenting in part) (reasoning that accused’s having posted a “tweet storm of 40

35 posts, all of which specifically tagged [the target’s] Twitter handle,” supported a

finding of malicious intent).

¶58 Courts should also consider the speaker’s familiarity with the recipients or

targets of the threat and the nature of the relevant parties’ personal history. For

example, in Elonis, the defendant’s alleged threats included lyrics posted to

Facebook that threatened violence against his wife soon after she left him and took

with her their two children.

. Relatedly, courts should consider

whether a threat’s intended recipient or target is particularly vulnerable, whether

because of personal characteristics or the parties’ relationship. See, e.g., A.J.B.,

(considering “an unrelenting torrent of cruel tweets at . . . an

individual diagnosed with autism and Attention Deficit Hyperactivity Disorder”

encouraging the target to commit suicide).

¶59 Finally, the subjective reaction of a statement’s target or foreseeable

recipients will be an important clue as to whether the message is a true threat.

See, e.g., Watts,

(reasoning that in part because of listeners’ laughing

response, defendant’s statement could not be interpreted as true threat). This

inquiry need not be limited to the recipient’s immediate reaction. See, e.g., D.J.M. v.

Hannibal Pub. Sch. Dist. No. 60,

(8th Cir. 2011) (teenage

recipient of threats via instant message initially responded “lol”—shorthand for

“laughing out loud”—but was concerned enough to tell trusted adult); Haughwout,

36

(observing that some students initially “elected to treat [the remarks

at issue] as made in jest,” but that “some of those same students nevertheless were

sufficiently perturbed to contact the university police”).

¶60 That said, courts should be wary of placing significant weight on the

subjective reaction of a statement’s unintended recipients. To do so risks punishing

a speaker for the content of a message that has been decoupled from its context.

This is of heightened concern given the vast temporal, geographic, and cultural

distance current technology permits speech to travel. We are mindful that

someone who stumbles upon a message he perceives as threatening may

experience sincere fear and anxiety. But to construe the true threats exception to

protect every passive internet user from the risk of such harms gives the doctrine

too wide a scope.

¶61 Moreover, a listener’s subjective reaction, without more, should not be

dispositive of whether a statement is a true threat. We acknowledge that the true

threats exception serves to protect individuals from “the fear of violence,” and

“from the disruption that fear engenders.” R.A.V.,

. But whether a

particular reader or listener will react with fear to particular words is far too

unpredictable a metric for First Amendment protection. Such a rule would not

give sufficient “breathing space” to the freedom of speech. Cf. Chaplinsky,

(“The word ‘offensive’ is not to be defined in terms of what a particular

addressee thinks.” (quoting Chaplinsky,

)).

¶62 The factors discussed here are not meant to constitute an exhaustive list.

Depending on the facts and circumstances, other considerations may be relevant

to the overarching goal of examining a statement in all its context to discern

whether it is a true threat or protected expression. Relatedly, the fact-finder has

discretion to weigh each factor in the balance, and to decide whether a particular

factor cuts for or against finding a true threat. Finally, in considering each factor,

courts may find it helpful to admit expert testimony to help illuminate coded

meanings, explain community norms and conventions, or bridge other contextual

gaps.

III. Application

A. Standard of Review

¶63 Whether a particular statement constitutes a true threat is an issue of fact to

be determined by the fact finder in the first instance. People v. Chase,

,

¶ 70,

; State v. Johnston,

(Wash. 2006). But in First

Amendment speech cases, an appellate court must make an independent

examination of the record to assure itself that the judgment does not impermissibly

intrude on the field of free expression. Chase, ¶ 70,

. Thus, whether

38 a statement constitutes a true threat is a matter subject to independent review.

Johnston,

.

B. R.D.’s As-Applied Challenge

¶64 It is unclear from the record what standard the trial court applied in

concluding that R.D.’s “particular type of speech is not protected under the First

Amendment.” The court heard argument from counsel but took no evidence on

that question. Moreover, the trial transcript reveals that the court did not

reconsider R.D.’s constitutional argument at the close of the prosecution’s case or

in the final ruling adjudicating R.D. delinquent. And in judging R.D.’s tweets

against the elements of section 18-9-111(1)(e), the trial court actively disregarded

testimony suggesting that A.C. and J.W. did not take R.D.’s messages seriously,

considering their reaction irrelevant under the statute. As stated above, their

reaction was a relevant factor to consider under the First Amendment.

¶65 Because we have clarified the test to be used when evaluating whether a

statement constitutes a true threat, the trial court is in the best position to review

the record, to take further evidence in its discretion, and to reach a conclusion on

the matter.

IV. Conclusion

¶66 We hold that a true threat is a statement that, considered in context and

under the totality of the circumstances, an intended or foreseeable recipient would

39 reasonably perceive as a serious expression of intent to commit an act of unlawful

violence. In determining whether a statement is a true threat, a reviewing court

must examine the words used, but it must also consider the context in which the

statement was made. Particularly where the alleged threat is communicated

online, the contextual factors courts should consider include, but are not limited

to (1) the statement’s role in a broader exchange, if any, including surrounding

events; (2) the medium or platform through which the statement was

communicated, including any distinctive conventions or architectural features;

(3) the manner in which the statement was conveyed (e.g., anonymously or not,

privately or publicly); (4) the relationship between the speaker and recipient(s);

and (5) the subjective reaction of the statement’s intended or foreseeable

recipient(s).

¶67 We agree with the parties that in this case, the government must also prove

that R.D. had the subjective intent to threaten. We need not decide today whether

the First Amendment requires that showing in every threats prosecution.

¶68 Because neither the juvenile court nor the court of appeals had the benefit

of the framework we adopt today, we reverse the judgment of the court of appeals

and remand with instructions to return the case to the juvenile court to reconsider

the adjudication applying this refined test.

40 41