Deahl, Associate Judge

Matthew B. Kaplan, Arlington, VA, for appellant.

Governor E. Jackson, III, Towson, MD, for appellee.

Ashwin P. Phatak, Principal Deputy Solicitor General, with whom Karl A. Racine, Attorney General for the District of Columbia at the time, Caroline S. Van Zile, Solicitor General, Graham E. Phillips, Deputy Solicitor General, and Jeremy R. Girton, Assistant Attorney General, were on the brief, for intervenor.

Eugene Volokh, of the bar of the State of California, pro hac vice, by special leave of court, with whom Gene C. Schaerr, Erik S. Jaffe, Washington, DC, and Joshua J. Prince, Philadelphia, PA, were on the brief, for Professor Eugene Volokh and Protect the First Foundation as amicus curiae in support of appellant.

Arthur B. Spitzer and Scott Michelman, Washington, DC, were on the brief for American Civil Liberties Union of the District of Columbia as amicus curiae in support of appellant.

Before Blackburne-Rigsby, Chief Judge, and Beckwith, Easterly, McLeese, Deahl, Howard, AliKhan, and Shanker, Associate Judges.

Opinion for the court by Associate Judge Deahl, with whom Chief Judge Blackburne-Rigsby and Associate Judges Beckwith, Easterly, and Shanker join in full, and Associate Judge Howard joins as to all but Part III.C.

Opinion by Associate Judge McLeese, with whom Associate Judge AliKhan joins, dissenting in part and concurring in the judgment, at page 1171.

Deahl, Associate Judge:

The First Amendment restricts the government's ability to prohibit and punish speech. Its protections extend to speech that is offensive, outrageous, and emotionally distressing. As just one example, the Supreme Court has recognized that the First Amendment protects the right of protesters at military funerals to proclaim "Thank God for IEDs," "Thank God for 9/11," "Thank God for Dead Soldiers," and "You're Going to Hell." Snyder v. Phelps , 562 U.S. 443, 448, 131 S.Ct. 1207, 179 L.Ed.2d 172 (2011). When it comes to expression, the First Amendment precludes the government from policing the bounds of polite society.

This appeal requires us to consider whether the District's stalking statute can be reconciled with the First Amendment. That statute makes it a criminal offense to engage in a course of conduct—including two or more "communicat[ions] to or about another individual"—that one knows or should know would reasonably cause another to suffer emotional distress. D.C. Code §§ 22-3132(8)(A), - 3133. By its terms, it restricts all manner of speech, without regard to its truth or falsity, and without regard to whether it is of public or purely private concern. The constitutional problems with the statute are glaring. See Forsyth County, Ga. v. Nationalist Movement , 505 U.S. 123, 134, 112 S.Ct. 2395, 120 L.Ed.2d 101 (1992) ("Listeners’ reaction to speech is not a content-neutral basis for regulation."); R.A.V. v. City of St. Paul , 505 U.S. 377, 382, 112 S.Ct. 2538, 120 L.Ed.2d 305 (1992) ("Content-based regulations are presumptively invalid."); Garrison v. Louisiana , 379 U.S. 64, 74, 85 S.Ct. 209, 13 L.Ed.2d 125 (1964) ("Truth may not be the subject of either civil or criminal sanctions where discussion of public affairs is concerned.").

People are generally allowed to say things that they know or should know will cause others emotional distress. Such speech is frequently socially valuable or, at the very least, accepted. Doctors tell patients that they likely have only months to live, or more distressing yet, that their children do. Spouses may knowingly inflict emotional distress by revealing a longstanding paramour and demanding a divorce. Police officers bring news of loved ones having been killed. Judges pronounce death sentences. All of those messages undoubtedly trigger extraordinary distress, and as a result, they are prohibited by the stalking statute's plain terms (at least if the messages need repeating). The statute is unconstitutional if read in that straightforward fashion, as "[s]peech may not be banned on the ground that it expresses ideas that offend." Matal v. Tam , 582 U.S. 218, 223, 137 S.Ct. 1744, 198 L.Ed.2d 366 (2017).

But the stalking statute has a savings clause, and we granted en banc review in this case to resolve its meaning. The stalking statute provides that "[t]his section does not apply to constitutionally protected activity." D.C. Code § 22-3133(b). To save the District's stalking statute from unconstitutionality, we interpret this clause to mean that, when speech is at issue, the statute covers only speech that fits within the "well-defined and narrowly limited classes of speech, the prevention and punishment of which have never been thought to raise any Constitutional problem." United States v. Stevens , 559 U.S. 460, 468-69, 130 S.Ct. 1577, 176 L.Ed.2d 435 (2010) (quoting Chaplinsky v. New Hampshire , 315 U.S. 568, 571-72, 62 S.Ct. 766, 86 L.Ed. 1031 (1942) ). That includes threats, obscenity, defamation, fraud, incitement, and speech integral to criminal conduct. Id. at 468, 130 S.Ct. 1577. Outside of those narrow categories, speech is constitutionally protected activity that the statute does not apply to.

In this case, the Superior Court found that there was good cause to believe that Lauren Mashaud, a married man, stalked Christopher Boone when he truthfully revealed to Boone's family, friends, and colleagues that Boone had an affair with Mashaud's wife. Mashaud now appeals. Because Mashaud's speech was constitutionally protected activity—i.e., it does not fit within any of the categories outside of the First Amendment's protections—we conclude that he did not stalk Boone. We therefore reverse.

I.

Boone's Affair and Mashaud's Exposure of It

This case stems from an extramarital affair between Boone and Mashaud's wife, Kimberly Weatherspoon. They met while working together at a consulting firm, where Weatherspoon was an intern and Boone was a vice president. After attending a company happy hour together, the two struck up a months-long affair. Boone testified that when the affair started, he believed Weatherspoon was in a "rocky relationship" with a boyfriend. According to Boone, when Weatherspoon finally disclosed that she was in fact married, he was "in a complete state of shock and disbelief" and told her that he could not knowingly engage in an extramarital affair. The next day, Weatherspoon told Boone that her husband, Mashaud, had found out about the affair. Mashaud was so upset that "he did not sleep for 48 continuous hours." While Weatherspoon told Mashaud that the affair was over, it in fact continued for about two more months.

A couple of months after Mashaud first learned of the affair, and around the same time it in fact came to an end, Mashaud sent an email to three employees of Boone's firm. At least one of those employees was a member of the firm's human resources department, the others were described only as "select directors" of the firm, and Boone was cc'd on the email as well. The email's subject line was, "Christopher Boone – No Contact/Harassment," and its stated purpose was "to bring a matter to [their] attention that may be a violation of [the firm's] Code of Conduct and/or other policies, procedures, business ethics, and character or standard of the Company." The email went on to state that:

Christopher Boone and my wife, Kimberly Weatherspoon, are involved in an extramarital affair that took place, primarily, in the workplace. Aside from the potential sexual harassment claims the situation presents, it also involves the inappropriate use of company resources and assets. Christopher Boone and Kimberly Weatherspoon have used company time and company resources to further their affair. ... I will anticipate a response from you once you have investigated these concerns and taken appropriate corrective action.

Boone testified that the email left him feeling violated, threatened, and embarrassed. Boone replied via email, telling Mashaud that he "would prefer that you leave me and my personal/business relationships out of a situation that is clearly between you and your wife." Boone added an apparent invitation to fisticuffs: "If you are a man ... I am more than happy to talk, meet, or whatever any time you're ready." Mashaud did not respond.

About three months later, Mashaud tracked down some of Boone's family and friends on Facebook and messaged them about the affair. As Mashaud explained, the privacy settings of Boone's Facebook account were configured so that anybody could see a list of people who had "liked" his photographs. And at the time, Facebook allowed users to pay to send messages to any other user, even if the two had no connection. Using this feature, Mashaud individually sent messages to at least fifteen people who had liked Boone's photos. Each message began by telling the recipient that they "should know the kind of person Christopher Boone really is." It went on to explain that Boone "had a sexual affair with [Mashaud's] wife" and that Boone showed he lacked "integrity and respect for himself" by "fail[ing] to respect the boundaries of a married woman." Mashaud attached a photo of Boone with Weatherspoon to his messages, which he closed with a Bible verse: "There is nothing covered that won't be uncovered, nothing hidden that won't be made known," citing Luke 12:2-3. Several recipients alerted Boone to these messages, and he testified that they again left him feeling "violated [and] threatened." He further testified to his confusion as to how Mashaud had obtained his friends’ and family members’ contact information, and he described his growing concern about his personal safety.

Mashaud also created a blog called "The Power of Light and Truth" discussing the affair and its aftermath. The publicly accessible blog repeatedly mentioned Boone by name, included at least five photographs of him, and identified his employer. Mashaud also "tagged" Boone by name in his posts so that internet searches for Boone's name would return hits to the blog detailing Boone's affair with Weatherspoon. For example, one post—titled "His name is Christopher Boone"—purported to "put a name and a face" to Boone and included links to his social media accounts and work profile. When Boone saw the blog, several months after he learned of the Facebook messages, "he was taken aback." Believing that Mashaud meant him physical harm, he notified his employer and apartment complex, both of which implemented additional security protocols. He also filed a police report and "started to see a mental health therapist for fairly aggressive treatment ... in a 12-week treatment program." Finally, Boone filed a petition in D.C. Superior Court for a civil protection order, or CPO.

Superior Court and Appellate Proceedings

Boone sought a CPO on the basis that Mashaud had committed a criminal offense against him, namely, stalking. The trial court could issue a CPO so long as it concluded by a preponderance of the evidence that Mashaud had "committed or threatened to commit a criminal offense against" Boone. D.C. Code § 16-1005(c).

Boone, Mashaud, and Weatherspoon all testified at a bench trial. Mashaud did not deny sending the various messages or authoring the blogposts, but instead argued (1) that his communications to and about Boone were protected by the First Amendment, and (2) that even if they were not protected, the messages would not have caused a reasonable person in Boone's shoes to fear for their safety, feel seriously alarmed, or suffer emotional distress. See D.C. Code § 22-3133.

The trial court disagreed as to both points. It reasoned that Mashaud's communications were not constitutionally protected because they were not "regarding matters of public concern" but "of a purely private nature." The court then found by a preponderance of evidence that Mashaud had stalked Boone, concluding that a reasonable person in Boone's shoes would be "seriously alarmed, annoyed, frightened, or tormented" by Mashaud's email message, Facebook messages, and blogposts. The court therefore issued a CPO. The CPO ordered both Mashaud and Weatherspoon to stay away from Boone's person, home, workplace, and vehicle; not to contact Boone in any manner; and not to "communicate about [Boone] by name or by implication on the internet or social media." The court further ordered Mashaud to enroll in a domestic violence intervention program and to pay Boone $1,800 as reimbursement for his therapy.

Mashaud appealed, arguing (1) that his statements were protected by the First Amendment and (2) that the trial court had erroneously applied a repealed version of the stalking statute, which had since been amended to omit conduct that merely "annoys" or "torments" another. We agreed with Mashaud on the second ground while reserving judgment as to Mashaud's First Amendment challenge. We reversed the CPO and remanded for the trial court to make findings tethered to the revised statute. Mashaud v. Boone , No. 14-FM-894, Mem. Op. & J. at 2 n.2, 4-5, 133 A.3d 1004 (Feb. 29, 2016).

On remand, the trial court again rejected Mashaud's constitutional challenge. The court drew a bright line between speech regarding matters of public concern and matters of private concern, describing the former as "usually deemed constitutionally protected" and the latter as "less protected, if protected at all." Relying on this distinction, the court found—contrary to its earlier ruling—that all of the blogposts were constitutionally protected speech because they touched on matters of public concern (i.e., marriage and extramarital affairs). Conversely, it determined that neither the email nor the Facebook messages were constitutionally protected speech, as they involved matters of purely private concern. It then applied the current version of the stalking statute to the email and Facebook messages and once again found that Mashaud had stalked Boone. Describing Mashaud's conduct as "clearly demonstrat[ing] a vindictive motive," the court concluded that he had "engaged in this behavior repeatedly ... with the intent to cause [Boone] to feel seriously alarmed, disturbed, or frightened or suffer emotional distress." It also found that "any reasonable person" in Boone's shoes "would feel seriously alarmed, disturbed, or frightened," or would "suffer[ ] emotional distress" in response to Mashaud's conduct. The court thus reissued the CPO.

Mashaud appealed for a second time, once again arguing that the District's stalking statute infringed upon his First Amendment rights. A division of this court again did not reach that question, instead ruling that the trial court had "misapprehend[ed] the extent of protection provided by the First Amendment" when it suggested that speech of purely private concern was not constitutionally protected. Mashaud v. Boone , 256 A.3d 235, 239 (D.C. 2021). After correcting that misstep, the division opted to remand back to the trial court "to consider in the first instance whether the evidence proved stalking given that speech about matters of private concern may enjoy constitutional protection." Id. Judge Beckwith dissented, arguing that Mashaud's constitutional arguments were "fully briefed and conducive to resolution by this court on appeal," so it did not make sense to "leave the trial court with the confusing task of determining for a third time" whether Mashaud's speech was covered by the stalking statute. Id. at 246. Judge Beckwith concluded that the most sensible reading of the stalking statute was that it applies only to those "categories of speech the Supreme Court has recognized as unprotected by the First Amendment." Id. at 243.

Boone petitioned for en banc review and Mashaud joined his request, as did the American Civil Liberties Union of D.C. as amicus. In framing the issue that they agreed merited this court's en banc review, they asked us to determine how best to interpret the stalking statute's savings clause: "This section does not apply to constitutionally protected activity." D.C. Code § 22-3133(b). We granted en banc rehearing and vacated the division's opinion. Mashaud v. Boone , 269 A.3d 1022 (D.C. 2021) (order).

II.

Before considering what the stalking statute's savings clause means when it exempts constitutionally protected activity, we first consider whether the evidence was sufficient to find that Mashaud stalked Boone independent of that clause. While Mashaud raises this argument, it is the District, as intervenor, that most forcefully presses the view that the evidence was insufficient to find that Mashaud had stalked Boone. No member of the division endorsed that view, though if it were correct, it would obviate the need to interpret the savings clause.

The District's sufficiency challenge falters on two independent grounds. First, we have no basis to reject the trial court's findings that Mashaud intended both his initial email and various Facebook posts to cause Boone emotional distress, D.C. Code § 22-3133(a)(1), a finding we owe considerable deference to. Second, the evidence was sufficient to support the trial court's conclusion that a reasonable person in Boone's shoes would have been emotionally distressed by the email, the Facebook messages, and Mashaud's blogposts. Id. § 22-3133(a)(3). We discuss those points, respectively, in Parts II.A and II.B below. But first, we lay some legal groundwork.

The District's stalking statute makes it a criminal offense to "engage in a course of conduct directed at a specific individual" with the intent to cause them to "(A) Fear for his or her safety or the safety of another person; (B) Feel seriously alarmed, disturbed, or frightened; or (C) Suffer emotional distress." D.C. Code § 22-3133(a)(1). For concision's sake, we refer to those emotional states collectively as "emotional distress," the broadest of the three categories. Even where a defendant lacks the intent to inflict emotional distress, they will violate the stalking statute if they engage in a course of conduct that they know or "should have known would cause a reasonable person" to suffer emotional distress. Id. § 22-3133(a)(3). In either case, the statute does not require any showing that the targeted person in fact suffered emotional distress; it is enough that the defendant intended such distress or should have known that a reasonable person would suffer it.

The statute further defines "emotional distress" as "significant mental suffering or distress that may, but does not necessarily, require medical or other professional treatment or counseling." Id. § 22-3132(4). It specifically covers "communicat[ions] to or about another individual." Id. § 22-3132(8)(A). Finally, to "engage in a course of conduct" requires "2 or more occasions" of distressing conduct, id. , and the defendant must "possess[ ] the requisite mental state" on each of those two (or more) occasions of distressing conduct, Coleman v. United States , 202 A.3d 1127, 1140 (D.C. 2019).

That last part is critical to how the District frames its argument: (1) it argues that the trial court's most recent order found stalking based on the email message and Facebook messages alone, without regard to the blogposts, so we should follow suit; (2) it asserts that the Facebook messages should be viewed as a single unitary act, so that if either the email message or the Facebook messages were not sufficiently distressing (or intended to be sufficiently distressing), then Mashaud did not stalk Boone; and (3) it then takes primary aim at the email message as too innocuous to be emotionally distressing, either in design or in effect. The dissent accepts that framing of the sufficiency argument, and agrees with the District's conclusion (with the aid of an interpretive gloss, discussed in Part III.C). In our view, the framing is faulty and the conclusion unwarranted. The first premise is mistaken, the second is dubious, and the conclusion is wrong in any event.

Contrary to the District's view, we cannot discount the blogposts as irrelevant to the sufficiency equation. The trial court found that those blogposts were part of Mashaud's criminal course of conduct when it initially granted a CPO against him, and that the blogposts would cause a reasonable person in Boone's shoes to suffer emotional distress. While the court did not again rely on those blogposts after our initial remand, that was only because it found they were of "public concern" and thus fell outside of the stalking statute's reach entirely. But nobody—not Mashaud, the District, Boone, the amici, or our dissenting colleagues—defends that reasoning. As explained by the division (unanimous on this point), the trial court was operating on the faulty premise that there is a public/private dichotomy as to speech, with the former enjoying absolute protection, and the latter receiving none. See Mashaud , 256 A.3d at 239. Disabused of that fallacy, there is no reason to ignore the blogposts when evaluating the sufficiency of the evidence.

Our colleagues assert that Boone did not challenge the trial court's decision to discount the blogposts. We disagree. In his en banc brief, Boone highlights the blogposts as the third activity in Mashaud's course of conduct that amounted to stalking, noting that Mashaud acted with the "intent of trying to permanently sully the personal and professional reputation of Mr. Boone" by "includ[ing] tags in the blog posts so that the blog was easily discoverable upon a Google search." He then made a more pointed attack on the trial court's decision to discount the blogposts, noting that the stalking statute's plain terms "encompass communications irrespective of the means—including a ‘blog’ " in which Mashaud "was not generally espousing his views on extramarital affairs in an informational manner. Rather, he was intentionally using the Internet as his weapon of choice." Our colleagues have also rejected the trial court's reasoning that statements of public concern are not covered by the stalking statute—their interpretation contains no such carve-out—so their refusal to apply their own test to those blogposts is puzzling.

It is also far from clear that we should treat the fifteen or more Facebook messages sent separately to distinct users as one collective act, rather than as separate acts, a critical component of our colleagues’ rationale that they never justify. The District's only basis for urging us to treat them as a single act is its assertion, without citation, that "Boone has always characterized Mashaud's Facebook messages as a single act." That's wrong. Boone says exactly the opposite: that Mashaud "did not pay to have one Facebook message sent, but paid in separate, distinct, multiple transactions to have messages sent ... thereby furthering the legitimacy of the trial court's finding that there was an engagement in a ‘course of conduct.’ " The District points to nothing in the trial court record where Boone conceded they should be treated as a unitary act, and we do not detect such a concession ourselves. Nonetheless, because the evidence is sufficient even if we treat the Facebook messages as a single act in Mashaud's course of conduct, we will accept for the sake of argument this premise of the District's position.

As for the District's conclusion that Mashaud's conduct was too innocuous to be emotionally distressing, either in design or in effect, we now address why that is mistaken. We first explain why the trial court was not plainly wrong to find that Mashaud intended both his email message and Facebook messages to be emotionally distressing. We then explain why it did not err in concluding that the email message, Facebook messages, and blogposts would cause a reasonable person in Boone's shoes to suffer emotional distress.

A. The Record Supports the Findings that Mashaud Intended to Cause Boone Emotional Distress.

"In reviewing the sufficiency of evidence in a bench trial, this court will not reverse unless an appellant has established that the trial court's factual findings are plainly wrong or without evidence to support them." Holmon v. District of Columbia , 202 A.3d 512, 521 (D.C. 2019) (citing Jones v. United States , 67 A.3d 547, 549 (D.C. 2013) ). The trial court's first basis for granting a CPO was that Mashaud's email and Facebook messages each "demonstrated a vindictive motive" and that he sent them "with the intent to cause [Boone] to feel seriously alarmed, disturbed, or frightened or suffer emotional distress." See D.C. Code § 22-3133(a)(1). That finding alone establishes a stalking charge under the terms of § 22-3133(a)(1), no matter whether a reasonable person in Boone's shoes would in fact have found those messages emotionally distressing.

Viewing the evidence "in the light most favorable" to the trial court's ruling, as we must on sufficiency review, we cannot say that no reasonable factfinder could have concluded that Mashaud intended to cause Boone emotional distress when he sent the email and Facebook messages. Johnson v. United States , 40 A.3d 1, 14 (D.C. 2012). We owe the trial court's findings about Mashaud's intent considerable deference, particularly where Mashaud testified and questions of subjective intent turn largely on credibility assessments. See Crawford-El v. Britton , 523 U.S. 574, 599, 118 S.Ct. 1584, 140 L.Ed.2d 759 (1998) (disputes about intent "frequently turn on credibility assessments"); Gaulden v. United States , 239 A.3d 592, 597 (D.C. 2020) ("We defer to a judge's reasonable ‘credibility determinations’ because such determinations are ‘the appropriate function of the fact finder.’ "). 1. The email message

Mashaud's initial email to Boone's colleagues, including at least one member of his firm's HR team, appears designed to get Boone fired. It begins by highlighting that the matter at issue, an affair between a vice president and an intern, "may be a violation of [the] Company's Code of Conduct and/or other policies, procedures, business ethics and character or standard of the Company." It then details how the affair "took place, primarily, in the workplace," and notes "the potential sexual harassment claims this situation presents." It closes by demanding a response "once you have investigated these concerns and taken appropriate corrective action." It raises a clear threat to Boone's livelihood, an implied threat of sexual harassment claims, and under the circumstances—where Boone does not dispute the truth of the email's contents—those threats stood a reasonable chance of bearing out. While the statute defines emotional distress as "significant mental suffering or distress that may" (but need not) require counseling, a substantiated threat to one's livelihood, financial wellbeing, and professional relationships clears that threshold. We see no basis for concluding that the trial court was plainly wrong in finding that Mashaud intended to cause Boone emotional distress. It is more than plausible—Mashaud had a clear motive to want Boone to suffer—and has abundant support in the record, most notably the email's contents. The District offers three counterpoints.

First, it asserts that the trial court, in support of its initial CPO that we vacated years ago, credited Mashaud's claim that he "did not mean to cause anyone any harm," and the District asserts that this factual finding cannot be reconciled with the trial court's subsequent finding that Mashaud sent his email with an intent to cause emotional distress. That's wrong. For starters, while the trial court noted in its findings that Mashaud "testified that he did not mean to cause anyone any harm," it never purported to credit that testimony. The trial court consistently noted the aspects of Mashaud's testimony that it found to be credible, and this claim was not among them. Also, this snippet of the trial court's recounting of Mashaud's uncredited testimony came in the context of discussing his blogposts, not his email to Boone's colleagues.

Second, the District claims that the record contained no "direct or indirect evidence" of Mashaud's mental state when sending the email to Boone's employer. We disagree. "Intent ... must ordinarily be proved circumstantially," Abdulshakur v. District of Columbia , 589 A.2d 1258, 1263 (D.C. 1991), as people tend not to admit to culpable states of mind. And here, there is no shortage of evidence that Mashaud intended to cause Boone emotional distress. Start with the fact that Boone had carried on a months-long affair with Mashaud's wife, which is a strong reason for Mashaud to want him to suffer. Then consider the contents of the email, leading with "Christopher Boone – No Contact/Harassment" in the subject line, then alluding to "sexual harassment claims" and demanding a response once "appropriate corrective action" had been taken. Add in the fact that, having alerted Boone's colleagues to the affair, Mashaud did not stop there, but instead went on to message dozens of Boone's Facebook friends and then create a blog littered with Boone's personal information. The trial court's conclusion that Mashaud had "a vindictive motive" for sending his email is not only plausible; it is nearly inescapable.

Third, the District asserts that the stalking statute "is not designed to discourage truthful reporting of potential misconduct." One would hope that is true, but the plain terms of the stalking statute in fact capture such speech, and the District offers no narrowing construction that would keep it from doing so. The narrowing construction that we adopt in Part III would take truthful reports of potential misconduct outside of the stalking statute's reach as constitutionally protected speech. Conversely, the District proposes no narrowing construction at all, but instead would simply have us counterfactually declare that truthful reports of misconduct will never cause, and can never be intended to cause, significant distress in another. That view is remarkably unrealistic. There are few things more emotionally distressing than one's darkest misdeeds coming to light, and no reason to think that the rightly implicated suffer less distress than the falsely accused; the falsely accused can at least take some solace in the truth potentially absolving them. While we doubt anybody intended the stalking statute to prohibit accurate reports of misconduct, we cannot get around the statute's plain text by simply pretending that people do not in fact suffer distress from such reports. Absent some narrowing construction, the statute and the District are saddled with that absurdity.

2. The Facebook messages

The Facebook messages, sent about three months after that initial email, present an even easier case. Recall that Mashaud had tracked down at least fifteen of Boone's Facebook friends—all strangers to Mashaud—and directly messaged them to reveal "the kind of person Christopher Boone really is" and "the dark reality of the[ ] affair." He revealed to them the details of the affair and Boone's "morally reprehensible behavior," including that it lasted for three months. The messages explained that Boone's friends and family should "know of his behavior" and Boone's "lack of integrity and respect for himself," before attaching a picture of Boone with Weatherspoon and closing with a Bible verse suggesting Boone could not keep the affair hidden. The trial court was not plainly wrong to conclude that Mashaud intended to cause Boone emotional distress through those numerous Facebook messages. He had strong reason to want to hurt Boone—the man had carried on an affair with his wife—and no other apparent reason for alerting so many of Boone's friends and family members to the affair.

The District makes a more halfhearted attack on this finding of intentionality, describing it as "a closer call." It makes two pointed attacks on it, each relegated to a single sentence. First, it contends that the trial court's finding of intentionality "cannot be reconciled with" its earlier conclusion that Mashaud did not harbor such an intent, an argument we have already rejected because the trial court never made that asserted finding. Second, it stresses that Mashaud testified that "exposing the affair to others was part of the process of healing his marriage." But even if the trial court credited that testimony, the statute does not require that a stalker be motivated solely by a desire to cause another emotional harm, so the point does not help the District. Revenge and personal betterment are not mutually exclusive motivations. For some, getting revenge on those who have wronged them is part and parcel of the healing process. See generally Jeffrie Murphy, Getting Even: Forgiveness and Its Limits (1st ed. 2003).

After detailing the evidence in the record, the trial court found that Mashaud "intended, and I emphasize intended, to cause [Boone] to feel seriously alarmed, disturbed, or frightened or to suffer emotional distress by sending Facebook messages directly to [Boone's] family and friends discussing personal matters." The record contains plenty of evidence supporting that conclusion.

B. The Record Supports that Mashaud Should Have Known He Would Cause a Reasonable Person in Boone's Shoes Emotional Distress.

Putting Mashaud's intentions aside, the record was also sufficient to support the trial court's findings that Mashaud should have known his actions would cause a reasonable person in Boone's shoes to suffer emotional distress. The District again challenges the evidentiary basis for this finding, but its arguments are similarly unpersuasive.

1. The email message

As we have already explained, Mashaud's email message to Boone's colleagues, including an HR representative and "select directors" of the firm, was a potent threat to Boone's livelihood. A reasonable person in Boone's shoes, having been outed for carrying out an affair "primarily, in the workplace," with a married subordinate and intern, might reasonably suffer serious distress about the consequences. They might reasonably suffer distress not only about losing their immediate job, but also about their long-term professional prospects, particularly given Mashaud's reference to "sexual harassment claims" that are capable of sinking careers.

The District disagrees, and asserts that having one's extramarital affair involuntarily exposed "would not cause a reasonable person to experience fear, serious alarm, or emotional distress." It seems implausible on its face to say that such a revelation would not cause a reasonable person emotional distress. But even if it were true, that grossly understates what Mashaud's email exposed. He did not merely reveal Boone's affair; he plainly implied that Boone had sexually harassed an intern under his supervision, that he did so using company time and resources, that he would not leave her alone despite Weatherspoon's requests, and that he should face professional consequences for his misconduct. As the trial court found, "a reasonable person would feel alarmed, disturbed, or frightened or suffer emotional distress if an email was sent to their place of employment suggesting that the person may have engaged in sexual harassment of a coworker." We agree with that assessment.

2. The Facebook messages

The same is true of the Facebook messages. The trial court made a finding that Mashaud "knew or should have known that ... sending Facebook messages directly to [Boone's] family and friends discussing personal matters ... would cause a reasonable person in [Boone's] circumstances to feel seriously alarmed, disturbed, or frightened or to suffer emotional distress." That finding is well-supported by the evidence.

The District counters that no reasonable person would have suffered emotional distress as that term is properly understood, highlighting cases in which we have described emotional distress as requiring "something markedly greater than the level of uneasiness, nervousness, unhappiness or the like which is commonly experienced in day to day living." Coleman , 202 A.3d at 1145 (quoting Wallace v. Van Pelt , 969 S.W.2d 380, 386 (Mo. Ct. App. 1998) ). But who could doubt that a reasonable person in Boone's shoes would suffer distress markedly greater than day-to-day worries? Having one's family and friends alerted to an extramarital workplace affair via Facebook message from a stranger, complete with pictorial evidence of the "dark reality of their affair" and accusations of refusing to abide by a no-contact request, is hardly just your ordinary day's hardship.

3. The blogposts

Finally, the District does not question that Mashaud's blogposts would cause a reasonable person in Boone's shoes to suffer emotional distress. And for good reason. As the District acknowledges, publicizing another's private information is a quintessential example of how some stalkers "use indirect means to threaten or monitor victims." The Nat'l Ctr. for Victims of Crime, Model Stalking Code Revisited 31 (2007). Mashaud's blog, "The Power of Light and Truth," repeatedly identified Boone by name, included at least five photographs of him, and specifically identified his employer. Boone's discovery of the blogposts prompted him to file a police report—thinking Mashaud meant him physical harm—and to notify his employer so that it could implement additional security measures. Also, in tandem with the email and Facebook messages, it caused Boone to start seeing "a mental health therapist for [a] fairly aggressive ... 12-week treatment program." We do not think that reaction was unreasonable, and it constitutes "significant mental suffering ... that may," and in this case apparently did, "require medical or other professional treatment or counseling." D.C. Code § 22-3132(4).

Rather than disputing that the blogposts would cause a reasonable person emotional distress, the District instead tries to take them out of the equation on the ground that the trial court did not rely on them in its most recent CPO ruling. As we explained above, unless the trial court's purely legal basis for discounting the blogposts was correct—and neither the District nor our dissenting colleagues opine that it was—that is a mistake. The trial court never renounced or contradicted its earlier findings that the blogposts would cause a reasonable person emotional distress. Those public blogposts, viewed by some "25 to 30 thousand people" in the first two months of the blog's existence, were at least as distressing as the email and Facebook posts. For all the same reasons the email and Facebook messages were emotionally distressing, the blogposts were as well.

III.

That brings us to Mashaud's primary argument: that by criminalizing "communicat[ions] to or about another individual" that would reasonably inflict emotional distress, the District has impermissibly restricted free speech, absent some narrowing construction of the statute. He is correct about that. Amici in support of Mashaud (the ACLU of D.C., Professor Eugene Volokh, and the Protect the First Foundation) further argue that the statute is unconstitutionally overbroad and would need to be struck down if it is not susceptible to a narrowing construction. Again, we agree. The stalking statute's plain terms prohibit a vast amount of speech—regardless of its truth, social value, or relation to public debate—based on its content. See Forsyth , 505 U.S. at 134, 112 S.Ct. 2395 ("Listeners’ reaction to speech is not a content-neutral basis for regulation."). It is substantially overbroad and plainly contravenes the First Amendment because it is not narrowly tailored to advance a compelling state interest. See, e.g. , Reed v. Town of Gilbert , 576 U.S. 155, 163, 135 S.Ct. 2218, 192 L.Ed.2d 236 (2015) ("Content-based laws—those that target speech based on its communicative content—are presumptively unconstitutional and may be justified only if the government proves that they are narrowly tailored to serve compelling state interests.").

Yet we will not strike down a statute as substantially overbroad if it is "readily susceptible to a narrowing construction that would make it constitutional," Virginia v. Am. Booksellers Ass'n, Inc. , 484 U.S. 383, 397, 108 S.Ct. 636, 98 L.Ed.2d 782 (1988), and that is where the savings clause comes in. It exempts "constitutionally protected activity" from the stalking statute's reach. D.C. Code § 22-3133(b). For the reasons that follow, we interpret this provision as narrowing the stalking statute's scope, in relation to communications to or about another person, to just those discrete categories of speech that fall outside the scope of the First Amendment's protections.

A. The Stalking Statute Is Substantially Overbroad

We begin by explaining why the stalking statute is substantially overbroad in the absence of some narrowing construction. As a court we are unanimous that serious concerns about the statute's constitutionality require us to interpret the statute narrowly. Unlike the dissent, we conclude that substantial overbreadth is not just a serious concern, but a certainty, absent some narrowing construction.

To begin, the stalking statute is a content-based restriction on speech. If a law requires " ‘enforcement authorities’ to ‘examine the content of the message that is conveyed to determine whether’ a violation has occurred," it is a content-based restriction. McCullen v. Coakley , 573 U.S. 464, 479, 134 S.Ct. 2518, 189 L.Ed.2d 502 (2014). And here, to determine whether a given communication would reasonably cause its target to suffer emotional distress requires an examination of the content of the communication itself. In short, the statute "draws distinctions based on the message a speaker conveys." Reed , 576 U.S. at 163, 135 S.Ct. 2218. Communications that could reasonably cause emotional distress are prohibited, whereas more innocuous statements are permitted.

Only Boone disputes this point, arguing that the stalking statute is merely a content-neutral regulation of how "speech is used by an alleged stalker." He misunderstands what it means for something to be a content-based regulation of speech. The statute applies to speech based on its content regardless of the manner in which the speech is conveyed. Presumably, Boone would not have sought a CPO had Mashaud emailed the consulting firm to praise him as a wonderful mentor to Weatherspoon or messaged his Facebook friends with similar praise without the affair in the background, and there is no reason to think that such messages would have caused a reasonable person emotional distress. Whether the statute prohibits Mashaud's speech requires some examination of the content of his speech, and it is therefore a content-based regulation. McCullen , 573 U.S. at 479, 134 S.Ct. 2518.

To be sure, not all content-based regulations are unconstitutional. For instance, the First Amendment "permits a State to ban a ‘true threat,’ " see Virginia v. Black , 538 U.S. 343, 359, 123 S.Ct. 1536, 155 L.Ed.2d 535 (2003), and to determine whether something is a true threat requires an examination of a statement's content. But the stalking statute is far broader than that, as it prohibits any speech that one should know would cause another to feel "seriously alarmed, disturbed, or frightened" or suffer "emotional distress." D.C. Code § 22-3133(a). That does not merely creep over the line into constitutionally protected areas; it sets up camp there.

Consider the perfectly ordinary and socially valuable speech that may put others in serious emotional distress, but that the stalking statute's plain terms nonetheless prohibit (if repeated). Doctors deliver life-shattering prognoses that surely send reasonable people to suffer emotional tailspins of distress. Spouses knowingly inflict emotional distress by revealing longstanding paramours and demanding divorces. Police officers deliver news of loved ones having been killed. Judges pronounce death sentences. Employers tell staff that they are fired. They all know, or should know, the extraordinary distress their messages bring, and so fall within the statute's prohibitions. Distressing speech is an important and often valuable part of life. A statute that prohibits speech indiscriminately based solely on its propensity for causing such distress is a constitutional nonstarter.

What's more, the statute's broad scope applies not only to speech concerning purely private matters, but also to communications on matters of public concern: speech that "occupies the highest rung of the hierarchy of First Amendment values." Snyder , 562 U.S. at 452, 131 S.Ct. 1207 (quoting Connick v. Myers , 461 U.S. 138, 145, 103 S.Ct. 1684, 75 L.Ed.2d 708 (1983) ). The desire to elicit an extreme emotional reaction is "not an uncommon purpose, nor one held only by a few evil people." Eugene Volokh, One-to-One Speech vs. One-to-Many Speech, Criminal Harassment Laws, and "Cyberstalking" , 107 Nw. U.L. Rev. 731, 773 (2013). Consider the abortion rights opponent who tells clinic staff that they are murderers, or the animal rights activist whose messaging includes graphic images of slaughtered, mangled, and deformed infant livestock. Both speak on issues of public concern and are therefore entitled to the strongest First Amendment protections. But both invariably intend to evoke deep emotional distress with the hope of making their targets feel "socially ostracized ... and emotionally racked with guilt, regret, and a perception of social condemnation." Id. at 773. Yet both would be in violation of the plain text of the District's stalking statute, which includes no special carve-out for speech on matters of public concern.

It is a foundational principle of the First Amendment that "speech cannot be restricted simply because it is upsetting or arouses contempt." Snyder , 562 U.S. at 458, 131 S.Ct. 1207. "The First Amendment protects lots of speech that is substantially emotionally distressing." United States v. Yung , 37 F.4th 70, 77 (3d Cir. 2022) ; see also United States v. Sryniawski , 48 F.4th 583, 587 (8th Cir. 2022) ("The Free Speech Clause protects a variety of speech that is intended to ... make another timid or fearful."). "If there is a bedrock principle underlying the First Amendment, it is that the government may not prohibit the expression of an idea simply because society finds the idea itself offensive or disagreeable." Texas v. Johnson , 491 U.S. 397, 414, 109 S.Ct. 2533, 105 L.Ed.2d 342 (1989).

The District's stalking statute runs headlong into that bedrock prohibition. Absent some narrowing construction, the District's statute would criminalize speech at the heart of many touchstone First Amendment cases. Protesters carrying signs reading "Thank God for IEDs" and "Thank God for Dead Soldiers" outside a deceased service member's funeral, Snyder , 562 U.S. at 460, 131 S.Ct. 1207, would likely be stalkers in the District, at least if their protests were repeated. The same is true for individuals who display burning crosses, swastikas, and other hateful symbols despite knowing that this expressive conduct will arouse "anger, alarm or resentment in others on the basis of race, color, creed, religion or gender." R.A.V. , 505 U.S. at 380, 112 S.Ct. 2538. A repeat neo-Nazi march through a neighborhood that is home to numerous Holocaust survivors, Nat'l Socialist Party of Am. v. Village of Skokie , 432 U.S. 43, 44, 97 S.Ct. 2205, 53 L.Ed.2d 96 (1977), would likewise be considered stalking. And when a national publication satirized an interview with Minister Jerry Falwell, indicating that his first time having sex was "a drunken incestuous rendezvous with his mother in an outhouse," that too would seem sufficiently distressing to be prohibited by the District's stalking statute. Hustler Mag., Inc. v. Falwell , 485 U.S. 46, 50-52, 108 S.Ct. 876, 99 L.Ed.2d 41 (1988).

As the Supreme Court has long recognized, "the First Amendment needs breathing space." Broadrick v. Oklahoma , 413 U.S. 601, 611, 93 S.Ct. 2908, 37 L.Ed.2d 830 (1973). We are particularly vigilant when the law at issue imposes criminal penalties, which create an especially high risk that an overbroad law "may deter or ‘chill’ constitutionally protected speech." Virginia v. Hicks , 539 U.S. 113, 119, 123 S.Ct. 2191, 156 L.Ed.2d 148 (2003). Accordingly, even a statute with many permissible applications must nonetheless be struck down when "a substantial number of its applications are unconstitutional, judged in relation to the statute's plainly legitimate sweep." Stevens , 559 U.S. at 473, 130 S.Ct. 1577 (quoting Wash. State Grange v. Wash. State Republican Party , 552 U.S. 442, 449 n.6, 128 S.Ct. 1184, 170 L.Ed.2d 151 (2008) ). If its scope were not somehow limited, the District's stalking statute would clearly be substantially overbroad.

The District offers several arguments in response, but they are not persuasive. First, citing to Coleman , it argues that speech must go beyond "mere distress or annoyance" to constitute stalking under the statute. That is undisputed, but that fact does not bring the statute into alignment with the First Amendment. Even very distressing speech is constitutionally protected, as Snyder , R.A.V. , Skokie , and Hustler make vividly clear.

Second, the District faults Mashaud for not demonstrating "from actual fact" that a substantial number of instances exist in which the law cannot be constitutionally applied. See N.Y. State Club Ass'n, Inc. v. City of New York , 487 U.S. 1, 14, 108 S.Ct. 2225, 101 L.Ed.2d 1 (1988). That is not a requirement of an overbreadth challenge, and the case the District cites to suggest otherwise is inapposite. New York State Club was a case brought by a plaintiff who could not identify a single, non-hypothetical entity whose free speech would be impacted by the challenged statute. Id. at 13-14, 108 S.Ct. 2225. Not surprisingly, the Supreme Court declined to presume the existence of such an entity when none was before it. Here, in contrast, Mashaud has identified at least one case in which the law has been unconstitutionally applied: his own. And we have already catalogued a swath of everyday speech that the stalking statute would similarly prohibit if its plain terms are read without narrowing, including truthful disclosures of serious misconduct or other crimes (where the outed party would reasonably feel distress as a result).

Finally, the District argues that even if there are some potential scenarios where the stalking statute's prohibitions apply to constitutionally protected conduct, the law's "legitimate reach dwarfs its arguably impermissible applications." See New York v. Ferber , 458 U.S. 747, 773, 102 S.Ct. 3348, 73 L.Ed.2d 1113 (1982). Again, the authority it relies on for this proposition is inapposite. Ferber was a challenge to New York's ban on the production and distribution of child pornography—material the Supreme Court held was entitled to no First Amendment protection. Id. at 765, 102 S.Ct. 3348. While the court agreed that New York's statute could conceivably apply to instances of protected expression (e.g., "pictorials in the National Geographic"), it voiced serious doubt that these cases "amount to more than a tiny fraction of the materials within the statute's reach." Id. at 773, 102 S.Ct. 3348.

Here, we face the opposite scenario. By its plain language, the District's stalking statute criminalizes any "communicat[ions] to or about an individual" that would reasonably cause emotional distress. A substantial amount of upsetting, disturbing, and alarming speech is protected by the First Amendment, not a merely conceivable iota of it. The better analogy is to Stevens , a case involving a federal statute that prohibited the creation, sale, or possession of a "depiction of animal cruelty." 559 U.S. at 464, 130 S.Ct. 1577. While Congress's intent was evidently to prohibit so-called "crush videos"—videos "depict[ing] women slowly crushing animals to death," catering to those with "a very specific sexual fetish," id. at 465-66, 130 S.Ct. 1577 —the defendant himself was charged with distributing videos of illegal dog fights, id. at 466, 130 S.Ct. 1577. The Supreme Court struck down the statute after determining that its plain meaning could extend to depictions of conduct as innocuous as hunting (and it did not require empirical evidence that the statute had been so applied, contrary to the District's argument above). Id. at 476-77, 130 S.Ct. 1577. Such a sweeping prohibition, the court held, was "substantially overbroad, and therefore invalid under the First Amendment," even if a more narrowly tailored statute could rightly ban crush videos. Id. at 482, 130 S.Ct. 1577.

The same reasoning applies here. We do not doubt that there is a compelling interest in prohibiting stalking, and that a narrowly tailored statute might prohibit at least some of the communications at issue here (probably not the email, though). As written, however, the statute is unconstitutionally overbroad and would need to be struck down unless some narrowing construction can save it from wholesale invalidation. Am. Booksellers , 484 U.S. at 397, 108 S.Ct. 636.

B. The Stalking Statute Can Be Saved by Limiting Its Application to Speech that is Beyond the First Amendment's Protections.

Because we may strike down a constitutionally overbroad statute only "as a last resort," Broadrick , 413 U.S. at 613, 93 S.Ct. 2908, we are obliged to determine if the stalking statute is " ‘readily susceptible’ to a narrowing construction that would make it constitutional," Am. Booksellers , 484 U.S. at 397, 108 S.Ct. 636.

The stalking statute gives us a natural entry point for that inquiry. It provides that it "does not apply to constitutionally protected activity." D.C. Code § 22-3133(b). Amici argue that we should read the savings clause as a categorical limitation on the statute's scope, as Judge Beckwith advocated in her dissent from the division's opinion. When it comes to communications to or about another person, they assert, the clause limits the definition of stalking to just those discrete, well-defined categories of speech that lack First Amendment protections: threats, defamation, and the like. The District counters that the savings clause states no more than a truism, leaving us to scrutinize on a case-by-case basis which of its particular applications run afoul of the Constitution.

For the reasons that follow, we conclude that the first approach is not only the superior of the two, but the only one that would save the statute from invalidation.

1. The savings clause limits the statute's scope so that it applies only to speech that is beyond the First Amendment's protections.

Amici argue that we should read the savings clause as exempting entire categories of constitutionally protected speech from the stalking statute's reach. Under this reading, when the savings clause says it does not apply to "constitutionally protected" speech, it exempts all speech except that which fits within the "narrowly limited classes of speech, the prevention and punishment of which have never been thought to raise any Constitutional problem." Stevens , 559 U.S. at 468-69, 130 S.Ct. 1577. That would leave threats, obscenity, defamation, fraud, incitement, and speech integral to criminal conduct within the stalking statute's reach, id. at 468, 130 S.Ct. 1577, while leaving the balance of speech beyond it.

There are substantial advantages to this interpretation of the statute. Most importantly, it would eliminate the statute's unconstitutional overbreadth. The statute would proscribe only that speech that lacks First Amendment protections. It would likewise avoid any vagueness problems—which the District's interpretation suffers from, as discussed below—as a would-be perpetrator would not need to perform a full constitutional analysis to determine if their conduct falls within the statute's scope. See Volokh, supra , at 764 (noting that, if interpreted in this way, "the statute would be a combination of a criminal libel statute, a threat statute, a fighting words statute, and the like").

Giving meaning to the statute's savings clause would also be consistent with our normal approach to interpreting statutes. Reading the savings clause in the District's preferred way would render it pointless, creating a surplusage problem. One could write it out of the statute entirely without affecting its meaning, because it goes without saying that the statute cannot apply to constitutionally protected activity. See Animal Legal Def. Fund v. Hormel Foods Corp. , 258 A.3d 174, 183 (D.C. 2021) (describing the "basic principle of statutory interpretation that each provision of the statute should be construed so as to give effect to all of its provisions, not rendering any provision superfluous" (cleaned up)); United States v. Menasche , 348 U.S. 528, 538-39, 75 S.Ct. 513, 99 L.Ed. 615 (1955) ("It is our duty to give effect, if possible, to every clause and word of a statute rather than to emasculate an entire section, as the Government's interpretation requires." (citation omitted)).

The District argues that our reading of the savings clause is inconsistent with the Council's intent, and that this construction "could immunize dangerous behavior that the Council has a compelling interest in prohibiting." Assuming the Council has a compelling interest in prohibiting speech that alarms, frightens, disturbs, or causes emotional distress, it must do so through legislation that is narrowly tailored to achieve that interest. Reed , 576 U.S. at 171, 135 S.Ct. 2218. To say, as the District does, that the Council intended to pass a statute that would criminalize all dangerous speech that it might otherwise prohibit through a more narrowly tailored statute is just to say the Council intended to pass an unconstitutional statute. If that were truly the only permissible reading of the statute, we would have to invalidate it wholesale. The District's concerns are also overstated. It highlights cases involving "harassing but non-threatening and non-obscene phone calls" that it worries would fall outside the stalking statute's reach under our reading, citing to Atkinson v. United States , 121 A.3d 780 (D.C. 2015), and Whylie v. United States , 98 A.3d 156 (D.C. 2014). But its worries are unfounded, as neither case depended on the sort of content-based speech restriction that would be excluded from the stalking statute's scope under the construction we adopt.

In Atkinson , for example, if we entirely ignore the content of the defendant's communications there is still no doubt that he committed stalking. In fact, by and large, we did ignore the content of his communications in that case. In Atkinson , the defendant's course of conduct included "multiple [phone calls] throughout the evening, often ‘back-to-back,’ " and "into the early morning hours," despite being told that he "needed to stop calling" and "I don't want you calling this house anymore." 121 A.3d at 782-83. When the complainant unplugged her phone, the defendant then called each of her parents and he was again told not to call. Id. A few days later, he showed up at the complainant's apartment building and "stood outside for a couple of minutes," until he was directed to leave by security. Id. Then he returned the following day again, and "security informed him that he was not permitted in the building." Id. Several months later, he again placed "nonstop calls" to the complainant from 1:30 a.m. to 3:00 a.m., and during that time he again showed up to the complainant's apartment and tried calling her from the apartment's call box. Id. at 783-84. He then tried sneaking into the apartment building to get to the complainant, but was again thwarted. Id. at 784.

Notice that we have not said one word about the content of the defendant's speech in Atkinson , because it was immaterial to his stalking conviction. He could have been trying to sell his victim a vacuum cleaner and the above actions would still have amounted to stalking. Our narrowed reading of the statute thus would have no impact on a course of conduct like that; it is only when a stalking charge depends on the content of one's speech that our narrow interpretation of the statute's application to speech comes into play. In Atkinson , the words spoken, and messages conveyed, were entirely beside the point (unlike here).

As for Whylie , not only did the course of conduct involve "more than a thousand calls" to the victim's workplace—so that one could ignore the content of those calls too—many of those calls included explicit and detailed threats against the victim's safety, which fall outside the First Amendment's protections. 98 A.3d at 160 & n.3. Our construction of the savings clause would not affect the statute's application to such constitutionally unprotected true threats. Atkinson and Whylie are the two lead examples the District offers to illustrate its practical concerns with our interpretation of the statute. Far from justifying the concerns, those cases demonstrate why they are unfounded.

A further caution comes from Mashaud, who warns that a broad reading of the savings clause in this case could have implications for District's law criminalizing the distribution of non-consensual pornography (often called "revenge porn"). Specifically, he notes that the revenge porn statute contains an identical exception for "constitutionally protected activity," which he suggests would be impacted by amici's proposed construction of the savings clause in this case. See D.C. Code § 22-3055(a)(1). But we see no risk of upsetting that statutory scheme. As explained, a narrow construction of the stalking statute is necessary only because the most natural reading of its prohibitions would be overbroad. While we have never considered the constitutionality of the District's revenge porn statute, similar prohibitions enacted by other jurisdictions have been consistently upheld as narrowly tailored to serve a compelling government interest. See, e.g. , Indiana v. Katz , 179 N.E.3d 431, 456 (Ind. 2022) ; Minnesota v. Casillas , 952 N.W.2d 629, 643 (Minn. 2020) ; Vermont v. VanBuren , 210 Vt. 293, 214 A.3d 791, 814 (2019) ; see generally Danielle Keats Citron & Mary Anne Franks, Criminalizing Revenge Porn , 49 Wake Forest L. Rev. 345 (2014). Assuming our revenge porn statute is similarly narrowly tailored to achieve a compelling state interest—and it seems to be at first blush—there would simply be no need to read the savings clause in a way that would narrow that statute's scope.

2. The savings clause does not simply state a truism.

The District counters that the savings clause should be read to state a truism: that like all legislative enactments, the stalking statute only applies to "conduct or expression," which we take to include expressive conduct, "that the Constitution does not shield after conducting the relevant constitutional analysis."

If read in this way, then it is technically true (as the District contends) that the statute would no longer be overbroad; by its literal terms, it would forbid nothing that the Constitution protects. But that walks right into a different and equally devastating constitutional problem: it substitutes vagueness for overbreadth, thereby leaving the statute on no firmer constitutional footing. "The void-for-vagueness doctrine requires that a penal statute define the criminal offense with sufficient definiteness that ordinary people can understand what conduct is prohibited." United States v. Smith , 685 A.2d 380, 384 (D.C. 1996) (quoting Kolender v. Lawson , 461 U.S. 352, 357, 103 S.Ct. 1855, 75 L.Ed.2d 903 (1983) ). The doctrine serves two important functions. The first function is that vague laws "undermine the Constitution's separation of powers" by placing judges, rather than legislators, in the position of deciding when to "make an act a crime." United States v. Davis , ––– U.S. ––––, 139 S. Ct. 2319, 2325, 204 L.Ed.2d 757 (2019). The second function is that vague laws "contravene the first essential of due process of law" by failing to give people "fair notice of what the law demands of them." Id. ; see also Sessions v. Dimaya , ––– U.S. ––––, 138 S. Ct. 1204, 1232, 200 L.Ed.2d 549 (2018) (Gorsuch, J., concurring) ("It leaves the people to guess about what the law demands—and leaves judges to make it up."). The stalking statute's savings clause, if interpreted as a truism in the manner proposed by the District, would typify both of these pitfalls.

First, it makes little sense to say (as the District asserts) that the stalking statute applies to all speech except that which the "Council could not permissibly prohibit under the Constitution." Whether speech has been permissibly regulated depends on how narrowly tailored the regulation is to a compelling interest. The Council cannot get around the narrow tailoring requirement by simply declaring that it is prohibiting all distressing speech that, through a narrowly tailored statute, it otherwise might. It must actually do the tailoring. A statute that simply directs judges to determine if speech of a particular type could be constitutionally proscribed lacks the requisite tailoring and is constitutionally infirm on its face. See Stevens , 559 U.S. at 482, 130 S.Ct. 1577 ("We therefore need not and do not decide whether a statute limited to crush videos or other depictions of extreme animal cruelty would be constitutional. We hold only that [this statute] is not so limited but is instead substantially overbroad."). The stalking statute is not narrowly tailored to a compelling state interest if we read the savings clause to state a truism, regardless of whether a variety of the speech it covers might be proscribed through a more targeted statute.

The District thus fundamentally misunderstands what strict scrutiny is when it counters that "[m]ost applications of the stalking law would survive even strict scrutiny" if the savings clause were interpreted as a mere truism. Strict scrutiny takes stock of the statute as a whole and asks whether it is narrowly tailored to achieve a compelling state interest, and a statute will fail that test if a "narrower regulation would suffice" to satisfy the state's interest. See Richard H. Fallon, Jr., Strict Judicial Scrutiny , 54 UCLA L. Rev. 1267, 1328 (2007). It is a category error to say a statute would survive strict scrutiny in "most applications"; " ‘overinclusive’ statutes also fail strict judicial scrutiny." Id.

Second, a statute with terms requiring "application of the appropriate constitutional test" would leave the line between innocent and prohibited conduct a matter of guesswork for most of the District's residents. As one treatise describes the problem, such a statute would be "patently vague":

[T]he Constitution does not, in and of itself, provide a bright enough line to guide primary conduct, and [ ] a law whose reach into the protected sphere is limited only by the background assurance that unconstitutional applications will eventually be set aside is a law that will deter too much that is in fact protected.

Laurence H. Tribe, American Constitutional Law 1031 (2d ed. 1988) (emphasis omitted) (discussing the risk of vagueness in "judicial reconstruction" of unconstitutionally broad statutes). Even for trained attorneys, a statute requiring ex ante constitutional analysis to determine its scope fails to give adequate notice of its requirements. And particularly when that statute imposes criminal liability, the result of this vagueness is inevitably the impermissible chilling of protected speech. See Grayned v. City of Rockford , 408 U.S. 104, 109, 92 S.Ct. 2294, 33 L.Ed.2d 222 (1972) ("[W]here a vague statute abuts upon sensitive areas of basic First Amendment freedoms, it operates to inhibit the exercise of those freedoms." (cleaned up)).

"The prohibition of vagueness in criminal statutes ... guards against arbitrary or discriminatory law enforcement by insisting that a statute provide standards to govern the actions of police officers, prosecutors, juries, and judges." Dimaya , 138 S. Ct. at 1212 (plurality opinion). It requires that the legislature "define what conduct is sanctionable and what is not." Id. "Where the legislature fails to provide such minimal guidelines, a criminal statute may permit ‘a standardless sweep that allows policemen, prosecutors, and juries to pursue their personal predilections.’ " Kolender , 461 U.S. at 358, 103 S.Ct. 1855. The legislature is not permitted to "set a net large enough to catch all possible offenders, and leave it to the courts to step inside and say who could be rightfully detained, and who should be set at large ... , substitut[ing] the judicial [and executive] for the legislative department of government." Id. at 358 n.7, 103 S.Ct. 1855 (quoting United States v. Reese , 92 U.S. 214, 221, 23 L.Ed. 563 (1875) ). Adopting the District's interpretation would mean the Council did exactly that. It would countenance an overbroad statute with the promise that courts would, on a case-by-case basis, pare it back as needed to conform to constitutional requirements, leaving the public to guess as to what the statute actually prohibits and the courts to define the statute's true scope.

C.

1. The dissent's test is unclear, unworkable, and unconstitutional.

Our dissenting colleagues endorse an alternative narrowing construction of the statute in which they effectively say the stalking statute's plain terms do not apply to conduct that we adjudge to be sufficiently worthy. That proposed reading of the stalking statute is atextual and unworkable, and it does not solve the stalking statute's constitutional infirmities.

As best we can discern, our colleagues’ test would carve out righteous speech (or conduct) from the stalking statute's ambit. They object to that characterization of their test, but the shoe fits, and their unwillingness to put a point on their test does not preclude us from doing so. In their own words, the stalking statute does not prohibit speech that is "justifiable" or is "not otherwise wrongful," i.e., that is righteous. Under that reading, justifiable speech simply does not trigger the kind of distress that counts under the stalking statute no matter how severe that distress might reasonably be, despite the statute's plain terms. The stalking statute's application thus turns on the relative worth that prosecutors and judges attach to the speech at issue. There is no textual hook for that interpretation in § 22-3133, which speaks purely in terms of the course of conduct's intended or reasonably anticipated effects on the target, without any concern for the righteousness of the conduct independent of those effects. Our colleagues instead rely on a vague phrase in § 22-3131, a statutory preface setting forth the "legislative intent" behind the stalking statute, which does not purport to define the offense of stalking at all, a task the Council undertook in § 22-3133. They import one particular phrase from that preface—"severe intrusions on the victim's personal privacy and autonomy"—into the statutory elements of stalking set forth in § 22-3133. How they landed on that particular phrase we do not know, but regardless, theirs is not a plausible reading of the statute and it does not remedy the statute's constitutional infirmities.

Their approach is strikingly similar to the one the Supreme Court unanimously rejected in Stevens , 559 U.S. 460, 130 S.Ct. 1577. Recall that Stevens involved a statute targeted at so-called "crush videos," depicting a particular type of animal cruelty that doubles as a sexual fetish. Id. at 465-66, 130 S.Ct. 1577. But the statute was overbroad, as it would have banned any "depiction of animal cruelty ... for commercial gain," including hunting videos or videos depicting such cruelty only to raise funds to combat it. Id. at 464-65, 130 S.Ct. 1577. "Not to worry," the government argued in Stevens , the statute's prohibitions should be read to "reach only ‘extreme’ cruelty" that has no "redeeming social value." Id. at 478-80, 130 S.Ct. 1577. The Supreme Court unanimously—including the lone Justice who dissented on different grounds—rejected that narrow reading of the statute because it had no mooring in the relevant provision's plain terms. The same is true here. Our colleagues do not offer any kind of test for what constitutes stalking beyond whether it is offensive enough to their own sensibilities that they would deem it as such, the very same argument the Supreme Court shot down in Stevens .

In any event, carving out righteous speech does not come close to curing the statute's substantial overbreadth. Again, a bedrock principle of the First Amendment is that it protects even abhorrent speech. Carving out "justifiable" speech does not meaningfully temper the statute's substantial overbreadth because "[t]he First Amendment's guarantee of free speech does not extend only to categories of speech that survive an ad hoc balancing of relative social costs and benefits." Stevens , 559 U.S. at 470, 130 S.Ct. 1577. Speech that most people would describe as completely unjustified and abject remains protected. So while our colleagues claim that their reading "avoid[s] some of the constitutional difficulties" with the statute, in truth it leaves the statute's constitutional problems unabated.

Our colleagues’ approach would also introduce intolerable vagueness into the statute. It is woefully unclear what counts as "a severe intrusion on ... personal privacy and autonomy" under their test. They counter that our interpretation has its own vagueness because some categories of unprotected speech are not so well-defined, highlighting "speech integral to criminal conduct." The Supreme Court seems to differ. Stevens , 559 U.S. at 468-69, 130 S.Ct. 1577 ("speech integral to criminal conduct" is among the "well-defined and narrowly limited classes of speech, the prevention and punishment of which have never been though to raise any Constitutional problem" (emphasis added)). The speech integral to criminal conduct exception has roots that are centuries deep. See Frohwerk v. United States , 249 U.S. 204, 206, 39 S.Ct. 249, 63 L.Ed. 561 (1919) ("[N]either Hamilton nor Madison, nor any other competent person then or later, ever supposed that to make criminal the counselling of a murder ... would be an unconstitutional interference with free speech."). The "severe intrusions on the victim's personal privacy and autonomy" test was coined just today, and our colleagues shed vanishingly little light on what it means in practice. Indeed, they do not even attempt to apply their test to the Facebook messages or blogposts at issue in this case, ignoring the latter based on the mistaken view that Mashaud has waived any reliance on them, and treating the former as a unitary act that was insufficient to establish a course of conduct.

We have no idea whether those communications satisfy our colleagues’ test, not because they raise close questions—close questions are inevitable even when statutes are perfectly clear—but because it is not clear what the relevant question is under their proposed test. Emotional distress is a statutorily defined term used in regular parlance and backed by hundreds of years of precedents and scholarship illuminating the phrase in various legal contexts. But "a severe intrusion on the victim's personal privacy and autonomy" has no statutory definition, no regular usage, and it is not a legal concept in any context. And our colleagues do not exactly flesh it out. All they tell us is that Mashaud's email is not sufficiently intrusive to be covered by the stalking statute, but beyond that, we are left to guess. They would introduce a judicial gut-check into a criminal statute that regulates speech; whether somebody has stalked another comes down to a prosecutor's or judge's sense of whether their conduct was "otherwise wrongful." That extraordinary vagueness is intolerable in a criminal statute that "abuts upon sensitive areas of basic First Amendment freedoms." Grayned , 408 U.S. at 109, 92 S.Ct. 2294 (cleaned up). Finally, as a practical matter, reading a righteousness carve-out (or intrusiveness element) into the stalking statute would weaken the statute's protections even when no speech is at issue. Our statutory reading narrows only the stalking statute's application to speech, limiting its proscriptions to those expressions that fall outside the First Amendment's protections. In contrast, our colleagues’ approach would narrow the stalking statute's application to all forms of conduct that might amount to stalking, such as following or monitoring another. While our own interpretation is that those acts need only be sufficiently emotionally distressing to constitute stalking, our colleagues would require that they also involve "severe intrusion[s] on ... personal privacy and autonomy" to qualify. While we have confessed our uncertainty as to what that test means, we have serious doubts about whether many of the stalking convictions that we have previously upheld would survive that narrowing of the statute.

Consider Beachum v. United States , where we upheld a stalking conviction in which the perpetrator had: (1) "regularly appeared on [the complainant's] block"; (2) rang her "doorbell and bang[ed] on [her] door for about a minute" on one occasion; (3) left a note in her mailbox that read "It's not about being your friend, it's about being your best friend. Can you? Hi."; and (4) once "tried to get [her] to come over and talk to him as he slid his hand down his pants." 197 A.3d 508, 509 (D.C. 2018). Under our colleagues’ proposal, the government in that case would need to prove not only that at least two of these acts would reasonably cause somebody emotional distress, but further that they each amounted to severe intrusions on personal privacy and autonomy. Assuming the last act clears that bar, which of the others might? Banging on a door for about a minute could hardly be called a severe intrusion into another's personal autonomy. Leaving a note like the one above could not possibly be. And it is hard to see how simply being on the complainant's block routinely would qualify either, given the multitude of reasons Beachum might have had for being there. Id. at 509 (complainant first saw Beachum when he was "visiting one of [her] neighbors"). Those acts are no more objectively upsetting than Mashaud's own conduct in this case, which our colleagues find insufficient to trigger liability even if the statute applies to all manner of speech.

Our colleagues say they "very much disagree," and that Beachum did in fact severely intrude upon his victim's privacy and autonomy when he banged on her door for about a minute and left her a note saying that he wanted to be best friends. It is tough to argue the point because their preferred test seems to turn entirely on one's own sensibilities, and there are no authorities or precedents lending meaning to their test that might permit us to counter. But we suspect that if you took a poll and asked people what might be fairly described as a severe intrusion on their privacy—having a stranger expose their extramarital affair to work supervisors, or having their door banged on for about a minute—few would land on our colleagues’ side of that divide. Far fewer still would say, as our colleagues do, that the trial court was plainly wrong in its own assessment of the evidence here. Theirs is a surprising view borne of an unpredictable test.

To put a point on it, our colleagues would weaken the statute's force for the sake of avoiding a First Amendment problem that they do not actually ameliorate.

2. The dissent's critiques of our interpretation are unpersuasive.

Our dissenting colleagues voice a host of qualms with our interpretation of the stalking statute's savings clause, though they have offered no interpretation of it themselves. They do not defend the District's position, that it states a mere truism, which is an untenable view for the reasons covered in Part III.B.2. Nor do they offer any other alternative to our interpretation. And the mismatch between their critique and their proposed remedy is telling.

If the savings clause were truly not susceptible to our narrowing interpretation of it, then the proper remedy would be to strike down the statute entirely as it is unconstitutionally overbroad absent our interpretive gloss. See, e.g. , Stevens , 559 U.S. at 482, 130 S.Ct. 1577. That is just to say that our colleagues’ critiques of our position are no points in favor of their own, which is untenable for the reasons we have described. Their critiques simply cast doubt on whether there can be any effective narrowing of the stalking statute that would save it from wholesale invalidation as overbroad. In any event, while the dissent offers nine critiques of our position, none raises serious concerns.

First, our colleagues decry an imprecision in what we have called constitutionally "unprotected speech." They say the phrase elides a nuance in First Amendment jurisprudence, that no speech is in fact entirely unprotected. That is true only in a pedantic sense, but in no way that matters here. Both this court and the Supreme Court routinely refer to categories of unprotected speech in the exact manner that we have today. See, e.g. , Stevens , 559 U.S. at 470, 130 S.Ct. 1577 ("[T]his Court has often described historically unprotected categories of speech."); Williams-Yulee v. Fla. Bar , 575 U.S. 433, 446, 135 S.Ct. 1656, 191 L.Ed.2d 570 (2015) (discussing "factors in determining whether to recognize ‘new categories of unprotected speech’ "); Brown v. Ent. Merchs. Assn. , 564 U.S. 786, 791-92, 131 S.Ct. 2729, 180 L.Ed.2d 708 (2011) (First Amendment's prohibition on content-based restrictions is "subject to a few limited exceptions for historically unprotected speech, such as obscenity, incitement, and fighting words"); Ferber , 458 U.S. at 764, 102 S.Ct. 3348 ("child pornography... like obscenity, is unprotected by the First Amendment"); Illinois, ex rel. Madigan v. Telemarketing Assocs., Inc. , 538 U.S. 600, 612, 123 S.Ct. 1829, 155 L.Ed.2d 793 (2003) ("fraudulent" speech "is unprotected speech"); In re S.W. , 45 A.3d 151, 156 (D.C. 2012) (describing "true threats" as "unprotected under the Constitution"); Competitive Enter. Inst. v. Mann , 150 A.3d 1213, 1240 (D.C. 2016) (asking whether "the speech in question actually falls within the unprotected category" (quoting Bose Corp. v. Consumers Union of U.S., Inc. , 466 U.S. 485, 505, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984) )). To analogize, describing the several categories of historically unprotected speech as such is like saying the sky is blue—readily understood and perfectly accurate for most purposes, even if one might interject a technical caveat that the sky only appears to be blue because of the optical phenomenon known as Rayleigh scattering. The caveat is not wrong, it is just usually omitted and typically unnecessary.

Second, our colleagues discern a "large disconnect" between our view and the statutory text because we have not surveyed the field of other constitutional amendments and what activities might qualify as "constitutionally protected" under them. We have not suggested that the savings clause exempts only activity that is constitutionally protected by the First Amendment. But the only constitutional right implicated by this case is the right to free speech, and that is clearly the right that the stalking statute most obviously and palpably inhibits—few other constitutional rights can be easily "directed at a specific individual," a requirement of the stalking statute. And our colleagues, who under the banner of judicial restraint will not say anything at all about the savings clause's meaning, are hardly in a position to critique us for not explaining how the clause operates when other constitutional rights are implicated. We suspect that if a homeowner causes a soldier emotional distress by twice denying them shelter, as is their right under the Third Amendment, that too is exempted from the stalking statute's reach. The question is simply not implicated here.

Third, our colleagues allege a lack of clarity in our opinion. We have illustrated through multiple examples above (such as our discussions of Atkinson and Whylie ) what it means to say that a stalking charge cannot depend on the content of one's speech, unless the speech fits within those historically recognized categories of unprotected speech. To the extent our colleagues have residual confusions, future cases can address those. In our view, we have gone to far greater lengths than our colleagues have in explaining how our test works. And while they counter that "both opinions decline to decide a number of issues," ours is the only opinion that decides the question we were asked to grant en banc review to resolve—the proper interpretation of the savings clause. Their own view in no way obviates the need to answer that question, it simply puts it off for yet another day.

Fourth, they claim our view introduces excessive complexity into the statute, because it is sometimes tricky to discern whether speech fits into one of the historically recognized categories of unprotected speech. But that same complexity exists in every statute regulating threats, fraud, incitement, and the like, so this particular complexity is nothing new. And the complexity inheres in our colleagues’ view as well, because even if they are of the view that the savings clause states a truism (they will not say if they are), they would still have to address as-applied constitutional challenges to the stalking statute, and tackle questions about the degree to which speech is constitutionally protected. Those questions can occasionally be difficult, to be sure, but our colleagues will have to answer them under their own test. And on top of that complexity, they have layered their opaque "severe intrusion on personal privacy and autonomy" test. If that test lacks complexity it is solely because it is entirely unmoored from any precedents or other authoritative guidance, leaving our colleagues to consult only their own sensibilities. That permits them to avoid the important questions we granted en banc review to resolve, but for the citizen who has to conform their behavior to their new test, and to the trial judge who would be left to guess what it means, it is hopelessly vague.

Fifth, our colleagues argue that our interpretation of the stalking statute may need additional narrowing to avoid an absurdity it posits. Perhaps, but the same is more obviously true of their preferred interpretation, which does not come close to resolving the stalking statute's substantial overbreadth (as discussed in the prior section). They do not even claim that it does. This critique, like the bulk of theirs, is more forceful when lodged against their own view.

Sixth, they assert that whether speech is a severe intrusion on the victim's personal privacy and autonomy under their test is no more vague than, say, whether speech is integral to criminal conduct or fits within other categories of unprotected speech. They are deeply mistaken; there is a huge difference between the occasional difficult legal question and vagueness. The speech integral to criminal conduct exception is a "well-defined and narrowly limited" category of unprotected speech, Stevens , 559 U.S. at 468-69, 130 S.Ct. 1577, with a historic pedigree going back hundreds of years, Frohwerk , 249 U.S. at 206, 39 S.Ct. 249, with many Supreme Court cases expounding upon its contours, see, e.g. , id. ; Giboney v. Empire Storage & Ice Co. , 336 U.S. 490, 498, 69 S.Ct. 684, 93 L.Ed. 834 (1949), including one set to be decided this term, see United States v. Hansen , No. 22-179 (argued March 27, 2023). While it can sometimes be hard to discern whether a communication is integral to criminal conduct—or a threat, or defamatory, or obscene—these are questions we and courts across the country have grappled with in countless cases for decades, and sometimes centuries, to obviate any vagueness concerns. See, e.g. , Carrell v. United States , 165 A.3d 314, 319-25 (D.C. 2017) (en banc) ("over the years this court has addressed any vagueness concerns" with the threats statute); Oparaugo v. Watts , 884 A.2d 63, 76 (D.C. 2005) (defamation); Ashcroft v. ACLU , 535 U.S. 564, 574, 122 S.Ct. 1700, 152 L.Ed.2d 771 (2002) (obscenity). By contrast, no principles, precedents, or authorities underlie whether conduct severely intrudes on another's privacy and autonomy; our colleagues have invented that test and will tell us only that Mashaud's email does not amount to it, that Beachum's banging on a door does, and then leave us guessing about what side of the line the universe of other conduct falls. That is a textbook vagueness problem that is intolerable when First Amendment rights are implicated.

Seventh, our colleagues say that our reading will "exempt very disturbing behavior," positing a defendant who sends repeat letters proclaiming they are "glad your child is dead" would not have stalked under our test. They take no position on whether that conduct would be criminally culpable under their own test (so, yet again, their concern rings hollow). It is true that the above example will not be stalking as we have interpreted the statute, but that is because the D.C. Council has not narrowly tailored a statute to proscribe that conduct, not because of our efforts to rescue this statute from its plain overbreadth. The Council could more narrowly tailor a statute to, for instance, prohibit direct communications to another that are intentionally designed to cause, and would reasonably cause, extreme distress after the recipient has clearly stated their objections to the contacts. Maybe such a statute, with a carve-out for political speech, would pass constitutional muster. But the District's stalking statute is a far cry from that, as it instead prohibits all manner of acceptable (and often socially desirable) speech that we have described above, while inviting the courts to sort out which speech really merits criminal culpability on a case-by-case basis. Our colleagues take them up on that invitation when they should decline it.

Eighth, our colleagues say we have not abided by the general preference for constitutional avoidance, but that critique is misplaced. As explained, we have adhered to that very principle by "interpret[ing] ambiguous statutory language to avoid serious constitutional doubts." Iancu v. Brunetti , ––– U.S. ––––, 139 S. Ct. 2294, 2301, 204 L.Ed.2d 714 (2019). We do not strike down the statute due to its overbreadth, but have given it a limiting construction to avoid that end. If our colleagues are faulting us for opining that the statute would be unconstitutionally overbroad absent some narrowing construction—as opposed to merely expressing concerns that it might be, as they do—that constitutional issue does not call for avoidance because it is not difficult. Vt. Agency of Nat. Res. v. U.S. ex rel. Stevens , 529 U.S. 765, 787, 120 S.Ct. 1858, 146 L.Ed.2d 836 (2000) ("[S]tatutes should be construed so as to avoid difficult constitutional questions." (emphasis added)). The avoidance canon does not constrain us to expressing concerns that a spade might be a spade rather than simply calling it one. And our colleagues’ own route of constitutional avoidance is to adopt an implausible and atextual reading of the stalking statute, in the exact way the Supreme Court unanimously rejected in Stevens , 559 U.S. at 481, 130 S.Ct. 1577 ("This Court may impose a limiting construction on a statute only if it is readily susceptible to such a construction." (cleaned up)).

Ninth, our colleagues justify their dismissal of the blogposts on the basis that Boone did not adequately challenge the trial court's decision to do so, so they feel free to disregard them as well. We have already responded to the point: Boone clearly relies on the blogposts in his en banc brief as a component part of Mashaud's stalking. Our colleagues counter that Boone made no legal attack in his en banc brief on the trial court's decision to discount the blogposts, but that is no doubt because nobody defended the trial court's reasoning (which the division unanimously rejected).

* * *

In sum, while the District's stalking statute would be unconstitutionally overbroad absent some narrowing construction, it is susceptible to just one narrowing construction that actually cures its First Amendment problems. When the statute says that it does not apply to "constitutionally protected activity," that must be understood to mean that it applies only to communications when they fall within the narrow categories of speech that lack First Amendment protections.

IV.

The only question remaining is whether Mashaud's speech fits within a narrow category of speech that lacks First Amendment protection. That is a pretty open and shut case: it does not. Mashaud did not defame Boone; it is undisputed that his statements were true. Armstrong v. Thompson , 80 A.3d 177, 183 (D.C. 2013). He did not threaten him. See In re S.W. , 45 A.3d 151, 156 & n.15 (D.C. 2012) (threats require "a serious expression of an intent to cause a present or future harm"). His speech did not incite criminal activity, and it was neither obscene nor fraudulent.

Boone counters that Mashaud's speech fits within one recognized exception to the First Amendment's protections, because it was integral to a criminal act—namely, stalking. This argument is fatally circular. While it is true that the First Amendment does not protect speech integral to criminal conduct, see Giboney , 336 U.S. at 498, 69 S.Ct. 684, "the speech must be integral to conduct that constitutes another offense that does not involve protected speech," Sryniawski , 48 F.4th at 588. The exception is typified by crimes like solicitation or conspiracy, which encourage others to participate in and thereby advance a non-speech offense. It makes no sense as an exception if the speech both constitutes the crime itself and thereby avoids First Amendment protections by being integral to its own commission. By Boone's reasoning, the government could prohibit any disfavored speech by first defining it as a crime and then claiming that the First Amendment no longer applies because that speech is now integral to that same offense. An exception like that would swallow the First Amendment whole and bears no resemblance to what this narrowly drawn exception actually is.

Because the course of conduct identified by the trial court consisted solely of "communications to or about another individual," and those communications did not fall within one of the categories of speech that lack First Amendment protections, the court erred by finding that Mashaud committed the crime of stalking.

V.

For the foregoing reasons, the judgment of the Superior Court is reversed, and the case is remanded with instructions to vacate the CPO and dismiss Boone's petition.

So ordered.

McLeese, Associate Judge, with whom AliKhan, Associate Judge, joins, dissenting in part and concurring in the judgment:

I agree that the civil protection order (CPO) in this case should be reversed. I also agree that the court should interpret the stalking statute narrowly in light of serious concerns about the statute's apparent breadth. I disagree, however, about what narrowing interpretation to adopt in order to decide this case. This opinion focuses on the interpretation of the phrases "[f]eel seriously alarmed, disturbed, or frightened" and "[s]uffer emotional distress." D.C. Code § 22-3133(a)(1)(B)-(C), (2)(B)-(C), (3)(B)-(C). As I explain, this court has already adopted a narrow interpretation of those phrases. Under that interpretation, the evidence in my view is insufficient to support a finding that Dr. Mashaud committed stalking. I therefore dissent in part and concur in the judgment.

I. The Narrowing Interpretation Adopted in Coleman v. United States

Among other things, the stalking statute prohibits communicating to or about another individual on two or more occasions, if the alleged stalker (1) intends to cause that individual to feel seriously alarmed, disturbed, or frightened or to suffer emotional distress; (2) knows that the individual would reasonably feel seriously alarmed, disturbed, or frightened, or suffer emotional distress; or (3) should have known that a reasonable person in the individual's circumstances would feel seriously alarmed, disturbed, or frightened or would suffer emotional distress. D.C. Code § 22-3132(3), (8) ; id. § 22-3133(a)(1)(B)-(C), (2)(B)-(C), (3)(B)-(C). "Emotional distress" is defined as "significant mental suffering or distress that may, but does not necessarily, require medical or other professional treatment or counseling." Id. § 22-3132(4).

In Coleman v. United States , 202 A.3d 1127, 1144-45 (D.C. 2019), this court interpreted the phrases "[f]eel seriously alarmed, disturbed, or frightened" and "[s]uffer emotional distress." D.C. Code § 22-3133(a)(1)(B)-(C), (2)(B)-(C), (3)(B)-(C). We stated generally that the "stalking statute is meant to prohibit seriously troubling conduct, not mere unpleasant or mildly worrying encounters that occur on a regular basis in any community." Coleman , 202 A.3d at 1144. More specifically, we cited with approval the statement that "emotional distress" should be interpreted to be a "feeling of mental anguish, something markedly greater than the level of uneasiness, nervousness, unhappiness or the like which is commonly experienced in day to day living." Id. at 1145 (brackets, emphasis, and internal quotation marks omitted). We explained that the phrase "serious alarm, disturbance, and fright should be understood as mental harms comparable to fear for one's safety or significant emotional distress." Id. We concluded that "to trigger criminal liability" under the stalking statute the conduct at issue must "involve a severe intrusion on the victim's personal privacy and autonomy." Id. (brackets and internal quotation marks omitted).

Our decision in Coleman rested on general considerations of statutory interpretation, and we did not address the potential significance of First Amendment concerns. 202 A.3d at 1144-45. In my view, those concerns confirm the prudence of the interpretation we adopted in Coleman . As a matter of ordinary language, the provisions at issue could be understood to give the stalking statute a troubling reach. For example, consider a case in which a citizen who justifiably believes that a government employee should be discharged for misconduct writes two letters to the employee's supervisor expressing that view. That conduct could certainly cause the employee to seriously fear being discharged and to be seriously alarmed and disturbed at that prospect. The citizen could also be inferred to have known that the employee would reasonably feel that way.

In my view, however, it would be absurd to think that the legislature intended to criminalize the writing of such letters. See, e.g. , Wade v. United States , 173 A.3d 87, 95 (D.C. 2017) ("When interpreting statutes, we assume that the legislature acted logically and rationally and we avoid interpretations of statutes which lead to implausible results.") (internal quotation marks omitted). We also "have a duty to avoid constitutional difficulties by applying an appropriate narrowing construction where possible." Grogan v. United States , 271 A.3d 196, 210 (D.C. 2022) (internal quotation marks omitted). Interpreting the stalking statute to reach the writing of such letters would raise obvious constitutional difficulties. See, e.g. , United States v. Yung , 37 F.4th 70, 78 (3d Cir. 2022) ("The First Amendment protects at least some speech that persistently annoys someone and makes [that person] fearful or timid."). Our duty to adopt an appropriate narrowing interpretation is heightened under the stalking statute, because the legislature directed us to interpret the stalking statute so that it "does not apply to constitutionally protected activity." D.C. Code § 22-3133(b).

The interpretation of the stalking statute that we adopted in Coleman avoids at least some of the constitutional difficulties that would be posed by a broader reading of the statute. For example, the hypothetical citizen who justifiably complained about a government employee's misconduct is not subject to liability under the stalking statute as we interpreted that statute in Coleman . A good-faith complaint about an employee's performance simply does not constitute "a severe intrusion on ... personal privacy and autonomy." Coleman , 202 A.3d at 1145 (brackets and internal quotation marks omitted). Moreover, an employee's concerns about possible job loss as a result of such a complaint cannot plausibly be viewed as "anguish ... markedly greater than the level of uneasiness, nervousness, unhappiness or the like which is commonly experienced in day to day living." Id. (brackets, emphasis, and internal quotation marks omitted).

II. The Insufficiency of the Evidence

In order to obtain a CPO, Mr. Boone bore the burden of proving by a preponderance of the evidence that Dr. Mashaud committed stalking. E.g. , J.O. v. O.E. , 100 A.3d 478, 481 & n.8 (D.C. 2014). This court reviews the trial court's finding on that issue deferentially, "giving full play to the right of the judge, as trier of fact, to determine credibility, weigh the evidence, and draw reasonable inferences." In re K.M. , 75 A.3d 224, 230 (D.C. 2013) (internal quotation marks omitted); see also D.C. Code § 17-305(a) (where trial judge acted as finder of fact, judgment "may not be set aside except for errors of law unless it appears that the judgment is plainly wrong or without evidence to support it").

As I have noted, the stalking statute requires conduct on two or more occasions. D.C. Code § 22-3132(8). In Coleman , we held that the required mental state of the defendant must separately be shown as to each of at least two occasions. 202 A.3d at 1139-42. In the present case, the trial court rested its finding of stalking on two occasions: (1) the sending of an email to Mr. Boone and Mr. Boone's employer and (2) the sending of Facebook messages to a number of Mr. Boone's family members and friends. I turn first to the sending of the email. Because I conclude that the evidence is insufficient as to that occasion, I do not consider the sending of the Facebook messages.

Dr. Mashaud's email accurately revealed that Mr. Boone, who was a company vice president, had had a sexual relationship with Dr. Mashaud's wife, who was an intern at the company. The email raised concerns about possible violations of sexual-harassment policies and inappropriate use of company resources. The trial court did not find that the email was false in any respect, and Mr. Boone's brief to the en banc court did not contend that the email was false in any respect. The email is temperate in tone and is not insulting, abusive, or graphic.

Mr. Boone's testimony about the effect of the email on him was that he felt "violated" and "embarrassed." Mr. Boone initially testified that he also felt threatened by the email, but he later clarified that he did not feel threatened until after subsequent conduct by Dr. Mashaud.

The trial court found that Dr. Mashaud intended the email to cause Mr. Boone to feel seriously alarmed, disturbed, or frightened, or to suffer emotional distress. The trial court also found that Dr. Mashaud knew or should have known that the email would cause a reasonable person in Mr. Boone's circumstances to feel seriously alarmed, disturbed, or frightened, or to suffer emotional distress. In reaching those conclusions, the trial court did not have the benefit of our ruling in Coleman , which was not issued until several years after the trial court's ruling in this case. In explaining its ruling, the trial court emphasized that the affair was a personal matter and that the email raised concerns about possible workplace misconduct by Mr. Boone.

I conclude that the foregoing evidence does not adequately support a finding that Dr. Mashaud intended, knew, or should have known that the email would cause Mr. Boone (or a reasonable person in Mr. Boone's position) to feel "anguish" arising from "a severe intrusion on [Mr. Boone's] personal privacy and autonomy." Coleman , 202 A.3d at 1145 (brackets, citation, and internal quotation marks omitted). Dr. Mashaud may have intended, and certainly should have known, that the email might cause Mr. Boone to feel concern, embarrassment, and unhappiness. For the reasons I have explained, however, the stalking statute requires a substantially higher showing. I conclude that Mr. Boone's brief testimony that he felt "violated" and "embarrassed" by the email provides inadequate support for a finding that the email was a "severe intrusion on personal privacy and autonomy" that foreseeably would cause Mr. Boone (or a reasonable person) to feel that his personal safety was threatened or to feel "anguish ... markedly greater than the level of uneasiness, nervousness, unhappiness or the like which is commonly experienced in day to day living." Id. (brackets, ellipses, emphasis, and internal quotation marks omitted).

Because the evidence does not support the requisite findings as to the sending of the email, the CPO must be reversed.

III. Responses to the Opinion for the Court

The opinion for the court adopts a different approach to the troubling apparent breadth of the stalking statute. D.C. Code § 22-3133(b) provides that the stalking statute "does not apply to constitutionally protected activity." The opinion for the court interprets that language to mean that the stalking statute does not apply to any speech that falls outside certain "narrowly limited classes of speech, the prevention and punishment of which have never been thought to raise any Constitutional problem." Supra at 1144 (quoting United States v. Stevens , 559 U.S. 460, 468-69, 130 S.Ct. 1577, 176 L.Ed.2d 435 (2010) (internal quotation marks omitted)). The opinion for the court appears to indicate that its approach applies not only to pure speech but also to expressive conduct. Supra at 1157, 1162. For ease of reference, I hereinafter use "speech" to refer also to expressive conduct.

In my view, the approach adopted by the court presents many difficulties and uncertainties. I ultimately conclude that the stalking statute is not "readily susceptible" to the narrowing interpretation adopted by the court. Stevens , 559 U.S. at 481, 130 S.Ct. 1577 ("This Court may impose a limiting construction on a statute only if it is readily susceptible to such a construction.") (brackets and internal quotation marks omitted).

A. Categories of "Unprotected" Speech

A basic premise of the court's approach is that there are "categories of speech that lack First Amendment protections." Supra at 1159. That premise does not appear to be sound. It is true that the Supreme Court has often referred to certain categories of speech as "not within the area of constitutionally protected speech." R.A.V. v. City of St. Paul , 505 U.S. 377, 383, 112 S.Ct. 2538, 120 L.Ed.2d 305 (1992) (internal quotation marks omitted). The Court has also explained, however, that such language is a "shorthand" that is not "literally true" and "must be taken in context." Id. In R.A.V. , for example, the Court considered an ordinance criminalizing "fighting words" that insulted or provoked violence on the basis of race, color, creed, religion, or gender. Id. at 391, 112 S.Ct. 2538. The Court acknowledged that it had previously referred to "fighting words" as outside the scope of the First Amendment. Id. at 383-84, 112 S.Ct. 2538. Nevertheless, the Court held that the ordinance was facially invalid under the First Amendment, because the ordinance impermissibly discriminated among fighting words based on the content of the fighting words and the viewpoint of the speaker. Id. at 391-96, 112 S.Ct. 2538.

Speaking more broadly, the Court explained that the First Amendment does apply to categories of speech sometimes referred to as unprotected:

Our cases surely do not establish the proposition that the First Amendment imposes no obstacle whatsoever to regulation of particular instances of such proscribable expression, so that the government may regulate them freely. That would mean that a city council could

enact an ordinance prohibiting only those legally obscene works that contain criticism of the city government or, indeed, that do not include endorsement of the city government. Such a simplistic, all-or-nothing-at-all approach to First Amendment protection is at odds with common sense and with our jurisprudence as well.

R.A.V. , 505 U.S. at 384, 112 S.Ct. 2538 (brackets, citation, and internal quotation marks omitted).

In my view, it is quite undesirable to tie the interpretation of § 22-3133(b) to language that the Supreme Court has acknowledged is not accurate. For example, consider "true threats." The opinion for the court clearly intends that the stalking statute be understood to apply to them. Supra at 1161–62. But the test that the court adopts would, by its own terms, seemingly exclude true threats from the stalking statute. "True threats" are not entirely unprotected by the First Amendment, and the punishment of "true threats" can "raise [a] Constitutional problem." Stevens , 559 U.S. at 468-69, 130 S.Ct. 1577 (internal quotation marks omitted); see, e.g. , R.A.V. , 505 U.S. at 388, 112 S.Ct. 2538 ("[T]he Federal Government may not criminalize only those threats against the President that mention [the President's] policy on aid to inner cities."); State v. Vawter , 136 N.J. 56, 642 A.2d 349, 359 (1994) (even if construed as limited to threats of violence, statute violated First Amendment because statute proscribed only threats "on the basis of race, color, creed or religion" and therefore impermissibly discriminated on basis of viewpoint).

The opinion for the court dismisses this problem as a "pedantic" "nuance." Supra at 1167. As the Supreme Court's decision in R.A.V. illustrates, however, the idea that no speech is wholly unprotected by the First Amendment is a quite important substantive principle of First Amendment law.

The opinion for the court also appears to imply that the D.C. Council, in enacting § 22-3133(b), probably relied on loose language in prior decisions of the Supreme Court and this court describing certain categories of speech as unprotected. Supra at 1167–68. That is a somewhat puzzling implication, given the court's apparent acknowledgment that its reading of § 22-3133(b) "is inconsistent with the Council's intent." Supra at 1160. In other words, the opinion for the court does not actually contend (and in my view could not plausibly contend) that the D.C. Council intended the words "does not apply to constitutionally protected activity" to mean "is wholly inapplicable to all speech except a few categories of speech (loosely) referred to as ‘unprotected’ by the First Amendment." Moreover, as I will now try to explain, the interpretation of § 22-3133(b) adopted by the opinion for the court in my view cannot reasonably be reconciled with § 22-3133(b) ’s text.

B. Statutory Text

The opinion for the court attempts to address the problem just noted by in effect interpreting § 22-3133(b) to carve out all speech except for "categories of speech the Supreme Court has sometimes loosely referred to as unprotected." Supra at 1167–68. In my view, that response worsens another serious problem I see with the court's approach: the large disconnect between the court's interpretation of § 22-3133(b) and the actual wording of that provision.

Section 22-3133(b) refers to the Constitution generally, not the First Amendment, and to "protected activity," not "protected speech." In my view, the court's opinion does not adequately come to grips with those features of § 22-3133(b). I take as a given that § 22-3133(b) must be read, consistent with its plain language, to implicate not only the First Amendment but also the Constitution generally, and to apply not only to speech but also more broadly to "activity" other than speech. I also take as a given that the words "constitutionally protected" must be given one consistent meaning under § 22-3133(b), rather than one meaning in the context of the First Amendment and a different meaning in the context of other constitutional protections, or one meaning when applied to "activity" other than speech and a different meaning when applied to the form of activity that is speech. See, e.g. , Clark v. Martinez , 543 U.S. 371, 378, 125 S.Ct. 716, 160 L.Ed.2d 734 (2005) (when interpreting single statutory phrase that modified three listed categories, Court explains that to give phrase "a different meaning for each category would be to invent a statute rather than interpret one"); Reno v. Bossier Par. Sch. Bd. , 528 U.S. 320, 330, 120 S.Ct. 866, 145 L.Ed.2d 845 (2000) ("As we have in the past, we refuse to adopt a construction that would attribute different meanings to the same phrase in the same sentence, depending on which object it is modifying.").

The opinion for the court does not shed light on how its interpretation of the words "constitutionally protected" for purposes of speech considered under the First Amendment can plausibly function when applied to constitutional protections other than the First Amendment and to activity other than speech. If the terms of § 22-3133(b) are given a consistent meaning, then under the court's approach the stalking statute would apply only to those non-speech activities that "fall within the narrow categories of [activity] that lack [constitutional] protection." Supra at 1170. That would be absurd, and I assume that the court does not intend for § 22-3133(b) to be read in that fashion. I note, however, that at oral argument one of the amici apparently took the position that § 22-3133(b) should be interpreted to preclude a finding of stalking based even in part on the fact that the defendant was standing outside the complaining witness's home displaying a firearm that the defendant lawfully possessed.

A more plausible textual reading of "constitutionally protected activity" is available, of course: the phrase could be read to mean "activity that it would be unconstitutional to consider as the basis of a stalking conviction." The court does not accept that reading in the context of speech and the First Amendment, however, so that reading of the phrase is unavailable to the court for purposes of construing § 22-3133(b) ’s actual language in the context of other forms of conduct and other constitutional provisions. In my view, this problem by itself demonstrates that the court's interpretation of § 22-3133(b) is untenable.

The foregoing problems get worse when one adds the complication that the categories of speech identified by the court are not actually unprotected under the First Amendment. Thus, the court presumably must interpret "constitutionally protected activity" to mean something like "(a) speech that falls outside the categories of speech that the Supreme Court has sometimes loosely referred to as unprotected by the First Amendment and (b) other activities that it would violate the Constitution to consider as the basis of a stalking conviction." I do not view that as a plausible reading of the actual words of § 22-3133(b).

The opinion for the court suggests that these issues can properly be left for another day. Supra at 1167–68. I disagree. As the court acknowledges, supra at 1170, a narrowing construction to avoid constitutional issues is not permissible if the statute's text is not "susceptible to the narrowing construction." Grogan v. United States , 271 A.3d 196, 210 (D.C. 2022) (internal quotation marks omitted). I conclude that the court errs by adopting an interpretation of "constitutionally protected activity" solely for purposes of speech and the First Amendment that cannot plausibly be squared with any general understanding of those terms in § 22-3133(b).

C. Clarity

The court's opinion is unclear about the nature of the limitation it would impose. Although this is not entirely clear (and the court's opinion is not explicit on the point), I assume that the court does not intend to suggest that evidence of "protected" speech is entirely out of bounds in a stalking case. For example, I assume that, even under the approach adopted by court, evidence that the defendant in a stalking case told the complaining witness that the defendant was hoping to make the complaining witness "suffer" would be admissible to prove the defendant's mental state. Other questions remain, though. The court's opinion sometimes seems to indicate that "protected" speech cannot be one of the actions that form the course of conduct constituting a stalking offense. Supra at 1160 (speech that is not unprotected is not part of definition of "stalking"). At other times, however, the court's opinion seems to indicate that "protected" speech can be part of the course of conduct constituting a stalking offense as long as the offense does not "depend[ ] on" on the content of the speech. Supra at 1161. Also, it is unclear what the last formulation means. Does a stalking offense depend on the content of speech only if the content of speech is necessary to the jury's finding of guilt? In other words, may a jury in a stalking case consider the content of speech with no limit other than that the course of conduct found by the jury must include at least two acts that are not "protected" speech considered for its content?

D. Complexity

The court's opinion in my view substantially underestimates the complexity of the offense that would be created by the narrowing interpretation the court adopts. The precise details might depend on how the uncertainties noted in the preceding paragraph were resolved. In any event, a stalking case that involved evidence of speech might present considerable complexity.

1. Identifying "Unprotected" Speech

First, if the prosecution relied on speech as part of the alleged course of conduct, the factfinder would have to determine whether the speech was "unprotected." See Hasty v. United States , 669 A.2d 127, 133 (D.C. 1995) ("Where a narrowing construction or interpretation has been placed upon a statute that, absent the narrowing construction, might otherwise be unconstitutional in some respect, that narrowing construction or interpretation, upon request and where supported by the evidence, must be the subject of proof at trial and should be submitted to the trier of fact for its determination."). Assuming that "unprotected" is understood to refer to categories of speech that the Supreme Court has sometimes loosely referred to as unprotected, juries in such cases might have to determine whether given speech was a true threat, was obscene, was defamatory, was fraudulent, constituted incitement, was integral to criminal conduct, constituted fighting words, was child pornography, or presented some grave and imminent threat. Supra at 1160 & n.14 (acknowledging that under its interpretation, stalking statute "would be a combination of a criminal libel statute, a threat statute, a fighting words statute, and the like") (internal quotation marks omitted).

Some of those categories have relatively well-settled contours, but others decidedly do not. One of the amici, Professor Eugene Volokh, has written a lengthy law-review article on the exception for speech that is integral to criminal conduct. The "Speech Integral to Criminal Conduct" Exception , 101 Cornell L. Rev. 981 (2016). In that article, Professor Volokh describes the exception as "long-dormant" and "little-defined," rooted in a Supreme Court decision that is "broadly and imprecisely written" and "has often been misinterpreted." Id. at 983, 1052. The opinion for the court decides one relatively straightforward issue about the exception: the exception cannot properly be interpreted to be entirely circular, and a legislature therefore could not define speech as a crime and then avoid First Amendment scrutiny on the theory that the speech is integral to that crime. Supra at 1170–71.

Other questions remain. Professor Volokh expresses the view that the exception also applies to speech integral to conduct that is merely "tortious" or "civilly actionable." 101 Cornell L. Rev. at 983 & n.1, 1012. The opinion for the court, however, refers only to criminal conduct. E.g. , supra at 1144, 1170–71. On the former view, the exception might well be quite substantial in scope. Moreover, relying on a single decision from a federal appellate court, the opinion for the court concludes that the exception applies only if the speech at issue is integral to "another offense that does not involve protected speech." Supra at 1171 (internal quotation marks omitted). I am not entirely certain what that means and whether the First Amendment imposes such a limitation. For example, consider a case in which the defendant is charged with stalking for repeatedly knocking on the complainant's door, despite having been asked to stop. I think it quite unclear whether the First Amendment would prevent consideration of evidence that the defendant was yelling racial epithets or sexually explicit insults at the same time, to prove: (a) the defendant's mental state; (b) circumstances that tended to support the victim's claim of subjective fear and/or the objective reasonableness of the victim's fear; or (c) further evidence of the course of conduct constituting stalking, at least as long as the jury did not rest entirely on the content of "protected" speech in finding the elements of stalking. That is true, in my view, even if the defendant's conduct did not itself constitute a true threat or fit within some other category of speech that the Supreme Court has sometimes loosely referred to as "unprotected." More generally, I do not envy the task of trial judges and juries in grappling with such questions and other comparable questions that will arise in applying other exceptions acknowledged by the opinion for the court, such as the exception for speech that poses a "grave and imminent threat." Supra at 1160 n.14 (internal quotation marks omitted).

2. Identifying "Protected" Speech that May Still Constitute Stalking

Juries might also need to be instructed that even "protected" speech can be considered as part of the course of conduct constituting a stalking offense, as long as the offense does not depend on the content of the speech. Supra at 1161. Whether speech is being considered for its content is a question that courts and academics have found quite challenging. See, e.g. , City of Austin v. Reagan Nat'l Advert. of Austin, LLC , ––– U.S. ––––, 142 S. Ct. 1464, 1471-75, 212 L.Ed.2d 418 (2022) (five Justices holding that regulation was not content based); id. at 1479-81 (Alito, J., concurring in part and dissenting in part) (suggesting that regulation was content based); id. at 1481-92 (Thomas, Gorsuch, & Barrett, JJ., dissenting) (concluding that regulation was content based); see generally, e.g. , Leslie Kendrick, Content Discrimination Revisited , 98 Va. L. Rev. 231, 232-33, 233 n.3 (2012) (citing articles reflecting belief "that, whatever the merits of a content-discrimination principle as a conceptual matter, the Supreme Court's application of it has been unprincipled, unpredictable, and deeply incoherent"). The opinion for the court apparently might require, in cases presenting the issue, that trial judges attempt to craft an instruction defining the concept of content-based discrimination for juries and that juries then attempt to apply that concept. Those seem no easy tasks. For example, can the jury consider that the defendant was yelling angry words, or does that depend on whether the words were angry in tone, only angry in content, or both?

In my view, such complexities weigh significantly against the narrowing interpretation the court adopts.

E. Incompleteness

One of the most obvious problems with the stalking statute is that, read broadly, the statute could be interpreted to reach entirely innocent conduct. The narrowing interpretation adopted by the court does not directly address that problem. For example, consider a case in which an employer justifiably suspends an employee without pay and then later discharges the employee. When the terms of the stalking statute are interpreted broadly and out of context, one could reasonably say that the employer "engaged in a course of conduct directed at" the employee that the employer "should have known would cause a reasonable person in the [employee's] situation to ... [s]uffer emotional distress." D.C. Code § 22-3133(a)(3)(C). I think that it would be absurd to interpret the stalking statute to criminalize the employer's conduct in those circumstances. I assume that the court would not take the view that suspending an employee without pay and then discharging the employee is a form of "protected" speech being considered for its content. It follows that the court leaves unremedied a distinct and serious problem with the stalking statute. Thus, the court seems to guarantee that an additional narrowing interpretation of the stalking statute will be necessary.

It is true, as the court notes (supra at 1164–65), that the narrowing interpretation I prefer will not necessarily solve all issues presented by the stalking statute. My point is that the opinion for the court certainly will not.

F. Vagueness

The court states that its approach "avoid[s] any vagueness problems ... [because] a would-be perpetrator would not need to perform a full constitutional analysis to determine if ... conduct falls within the [stalking] statute's scope." Supra at 1160. As discussed, I think it doubtful that the approach reflected in the court's opinion would simplify the resolution of stalking cases or give clear guidance to trial judges and primary actors as to the permissible scope of the stalking statute. Rather, the court's opinion would in some cases require "would-be perpetrators" (and juries) to decide whether speech was a true threat, was obscene, was defamatory, was fraudulent, constituted incitement, was integral to criminal (or perhaps tortious or civilly actionable) conduct, constituted fighting words, was child pornography, or presented some grave and imminent threat. Supra at 1160 & n.14. The court indicates that a statute is unconstitutionally vague if the statute "requir[es] ex ante constitutional analysis to determine its scope." Supra at 1163. That, however, seems to be exactly what the stalking statute would require as interpreted by the court in this case.

G. Magnitude of the Limitation

It is difficult to determine the extent to which the approach reflected in the court's opinion would diminish the reach of the stalking statute, because it is not clear what exact limitation the court contemplates. In any event, however, the court's opinion does seem to exempt very disturbing behavior. Consider a case in which the complaining witness's young child has recently died, and the defendant each day for an extended period sends a letter to the complaining witness saying, "I am glad your child is dead. I hope you suffer every day for the rest of your life." Under the court's opinion, the stalking statute would seemingly not apply to that course of conduct. Moreover, the court excludes that conduct from the stalking statute without explicitly determining whether that conduct is "constitutionally protected activity" in the normal sense of that phrase, i.e., whether the First Amendment would preclude imposition of a criminal sanction for that conduct.

To be clear, I express no firm view as to whether the First Amendment would permit or preclude imposition of criminal liability based on such conduct. Although my leaning is that criminal liability could lawfully be imposed based on such conduct, that seems to me to be an unsettled question of First Amendment law that the court need not answer in this case. My point is that the court's opinion excludes such conduct from the stalking statute without even deciding whether the First Amendment requires that result.

H. The Doctrine of Constitutional Avoidance

This court has said that it "ordinarily will not address constitutional issues ... when the case may be resolved on other grounds. The practice of avoiding constitutional issues if it is reasonably possible to do so is predicated on a fundamental rule of judicial restraint, which is perhaps more deeply rooted than any other doctrine of constitutional adjudication." Gaynor v. United States , 16 A.3d 944, 948 (D.C. 2011) (citation and internal quotation marks omitted). The opinion for the court resolves numerous constitutional issues, including that the stalking statute would be substantially overbroad unless given a narrowing construction, supra at 1155–59; that the First Amendment would be unconstitutionally vague if used as the basis for a criminal prohibition, supra at 1162–63; and that the exception for speech that is integral to a criminal act does not apply unless the criminal act "constitutes another offense that does not involve protected speech," supra at 1171 (internal quotation marks omitted). In contrast, I conclude that this case can be resolved without the need to decide such constitutional issues. The availability of a non-constitutional basis for resolving this case weighs heavily against the narrowing interpretation adopted by the court.

I. Blogposts

The court concludes that, in assessing the sufficiency of the evidence, the court is required to consider the later blogposts Dr. Mashaud sent, even though the trial court held on remand that the blogposts could not constitutionally be considered. Supra at 1149–50. Mr. Boone, however, has not challenged the trial court's ruling on that issue in his briefing before the en banc court. I see no reason to consider that unchallenged ruling. See generally, e.g. , Grimes v. D.C., Bus. Decisions Info. Inc. , 89 A.3d 107, 112 n.2 (D.C. 2014) (declining to address issue that was not briefed on appeal).

The court contends that Mr. Boone did adequately raise this issue, citing two passing references to the blogposts in Mr. Boone's en banc brief. Supra at 1149–50. In my view, those passing references, unaccompanied by any legal analysis, do not suffice to adequately present a challenge to the trial court's constitutional ruling. See, e.g. , Roth v. D.C. Bd. of Zoning Adjustment , 279 A.3d 840, 845 (D.C. 2022) (declining to consider issue mentioned in passing, because petitioner did not provide legal argument in support of passing suggestion). In sum, I agree with the District of Columbia and Dr. Mashaud that Mr. Boone did not adequately present a challenge to the trial court's constitutional ruling excluding consideration of the blogposts.

J. The Court's Critique of the Dissent

The court identifies a number of perceived problems with the narrowing interpretation I prefer. I do not see any easy solution to the troubling apparent breadth of the stalking statute, and I agree that the interpretation I favor presents its own challenges. On balance, however, I am not persuaded that the approach adopted by the court is preferable. In any event, for reasons I have already explained, I do not believe that the approach adopted by the court is permissible. Rather, in my view, the stalking statute is not susceptible to the interpretation that the court adopts.

First, the court expresses uncertainty about the narrowing interpretation I prefer. Supra at 1163–64. To be clear, I do not propose to "ratchet[ ]" up the degree of anticipated harm beyond the level already reflected in our prior decisions. Supra at 1164 n.15. I do note, however, that our prior decisions, including Coleman , have emphasized that the level of anticipated harm under the stalking statute must be "severe." 202 A.3d at 1145. Rather, I conclude that—independent of First Amendment concerns, but also supported by such concerns—the stalking statute should not be interpreted to reach courses of conduct that are not otherwise wrongful but that foreseeably might cause severe alarm or emotional distress. As I have explained, absent such an interpretation, the stalking statute would have an absurd reach, criminalizing much innocent conduct.

Second, the court describes the narrowing interpretation I prefer as "carv[ing] out righteous speech." Supra at 1164. I disagree with that characterization. Under the interpretation adopted by this court in Coleman , the stalking statute requires that the conduct at issue (including speech) constitute "a severe intrusion on the victim's personal privacy and autonomy." Coleman , 202 A.3d at 1145. The conclusion that evidence in this case is insufficient is based on the application of that standard, not based on whether Dr. Mashaud's conduct could in some sense be viewed as "righteous."

Third, the court points out that there is no obvious "textual hook" in § 22-3133 itself for the narrowing interpretation I prefer. Supra at 1163–64. There is, however, such a textual hook in D.C. Code § 22-3131, which expresses the D.C. Council's intent in making stalking a crime. That provision states that "[s]talking involves severe intrusions on the victim's personal privacy and autonomy." I view that description of the offense as a natural place to look when adopting a narrowing interpretation of the offense to avoid absurdity. In fact, our decision in Coleman expressly relied on the D.C. Council's statement of intent in concluding that stalking must involve "a severe intrusion on ... personal privacy and autonomy." 202 A.3d at 1145 (brackets and internal quotation marks omitted). I note, moreover, that the court does not dispute that it would be absurd to interpret the stalking statute to reach entirely innocent conduct, does not propose an alternative solution to address the problem, and does not identify a "textual hook" for such a solution.

In any event, the narrowing interpretation I favor rests on the need to avoid absurdity. In that setting, it is well settled that "basic principles of statutory construction require that the actual language of a statute be ignored or revised to avoid the absurdity that would result if it were read literally." Young v. U-Haul Co. , 11 A.3d 247, 250-51 (D.C. 2011) (brackets and internal quotation marks omitted).

Relatedly, I do not agree with the court's suggestion that adopting the narrowing interpretation I prefer would in any way be inconsistent with the Supreme Court's decision in Stevens , 559 U.S. 460, 130 S.Ct. 1577. Supra at 1164–65. Stevens involved a proposed narrowing interpretation to avoid First Amendment problems, not a narrowing interpretation necessary to avoid absurdity. Stevens , 559 U.S. at 460, 130 S.Ct. 1577.

Fourth, the court correctly points out that the words "severe intrusion on ... personal privacy and autonomy" are not entirely clear. Supra at 1165. The same is true, however, of terms that the court would require juries to decide but the court does not attempt to define, such as "integral to criminal (and perhaps tortious or civilly actionable) conduct," and "grave and imminent threat." To the extent that the court contends that the phrase "severe intrusion on ... personal privacy and autonomy" is "intolerabl[y] vague[ ]" and "woefully unclear," supra , I disagree. The court fails to persuasively explain why those terms are unacceptably murkier than terms such as "integral to criminal (or tortious or civilly actionable) conduct" or many other terms in criminal law, such as "reasonable" or "adequate provocation," see generally, e.g. , Parker v. United States , 155 A.3d 835, 845 (D.C. 2017) (amount of force used in self-defense must be reasonable); Carome v. Carome , 262 A.3d 242, 247 (D.C. 2021) (destruction of property is not malicious if defendant acted based on "adequate provocation of the kind that would cause an ordinary, reasonable person to lose ... self-control") (internal quotation marks omitted).

Fifth, and relatedly, the court seems to suggest that the words "severe intrusion on ... personal privacy and autonomy" are a complete novelty insusceptible of reasonable understanding. Supra at 1165. To the contrary, however, there are preexisting torts that involve the concepts of invasion of privacy and intentional infliction of emotional distress. See, e.g. , Doe v. Bernabei & Wachtel, PLLC , 116 A.3d 1262, 1266 (D.C. 2015) (invasion of privacy); Newmyer v. Sidwell Friends Sch. , 128 A.3d 1023, 1037 (D.C. 2015) (intentional infliction of emotional distress). I do not mean to suggest that the elements of those torts should be imported into the stalking statute wholesale, but their existence indicates that these concepts are not contentless or inadministrable.

Sixth, the court states that the narrowing interpretation I prefer would not solve or even ameliorate the substantial overbreadth of the stalking statute. Supra at 1164–65, 1166–67. It seems clear that the narrowing interpretation that I prefer would at least ameliorate overbreadth concerns, because that interpretation would preclude stalking convictions in many of the hypothetical scenarios discussed by the court, such as when a doctor conveys bad news to a patient. Supra at 1143–44. I express no view as to whether the narrowing interpretation I prefer would completely solve overbreadth concerns, because I see no need to decide any constitutional issue in order to resolve this case. I do note, however, that the First Amendment analysis seems to me to be more complicated than is reflected in the court's opinion. For example, it appears to be unsettled to what extent the First Amendment prohibits the imposition of civil or criminal liability based on malicious speech on matters of private concern. See Snyder v. Phelps , 562 U.S. 443, 452-55, 131 S.Ct. 1207, 179 L.Ed.2d 172 (2011) (acknowledging that "First Amendment protections are often less rigorous" "where matters of purely private significance area at issue"). Similarly, the Supreme Court has held that the First Amendment can in some circumstances permit the criminalization of cross-burning with bad intent. Virginia v. Black , 538 U.S. 343, 348, 352-363, 123 S.Ct. 1536, 155 L.Ed.2d 535 (2003) (upholding Virginia statute making it crime to burn cross with intent to intimidate). As previously noted, the novelty and difficulty of such questions counsel against unnecessarily reaching First Amendment issues in this case.

Seventh, the court states that the narrowing interpretation I prefer would "weaken the statute's protections even when no speech is at issue." Supra at 1166. I disagree. As I have explained (and as the court's opinion does not dispute), some sort of narrowing interpretation is necessary to avoid giving the stalking statute absurd application to innocent conduct. I do not view the avoidance of absurdity as "weaken[ing] the statute's protections." Moreover, the narrowing interpretation favored by the court would also, by definition, narrow the scope of the stalking statute from the scope the statute otherwise would apparently have had. For example, as previously noted, the narrowing interpretation preferred by the court would interpret the stalking statute as inapplicable to malicious speech on matters of private concern, even though the court does not hold that the First Amendment precludes imposition of criminal liability based on such speech.

The sole specific example of "weaken[ing]" the opinion for the court provides is Beachum v. United States , 197 A.3d 508 (D.C. 2018). Supra at 1166. I do not view that example as well-chosen. The facts in Beachum were as follows: (1) beginning in the spring or summer of 2016, Mr. Beachum regularly tried to engage the complaining witness in conversation; (2) although the complaining witness rebuffed him each time, he persisted, making her feel uncomfortable; (3) in July 2016, Mr. Beachum tried to get the complaining witness to come over and talk to him while he slid his hand down his pants; (4) in the following months, Mr. Beachum continued to try to talk to the complaining witness, at one point offering her $50; (5) in January 2017, Mr. Beachum approached the complaining witness as she was entering her home and said he wanted to speak with her; (6) the complaining witness tried to ignore him, but he persisted, being "extremely more aggressive" than in previous encounters, causing the complaining witness to go into her home and call the police; (7) Mr. Beachum went to his van, wrote something on a piece of paper, returned to the complaining witness's house, and started ringing the doorbell and banging on the door for about a minute; (8) when the police arrived, the complaining witness found a note in her mailbox asking the complaining witness to be Mr. Beachum's best friend; and (9) the note and Mr. Beachum's escalating conduct made the complaining witness afraid. Beachum , 197 A.3d at 509. In my view, a reasonable factfinder could have found Mr. Beachum guilty of stalking on those facts under the narrowing interpretation I prefer. A reasonable factfinder could conclude that Mr. Beachum's conduct, on two separate occasions, was a severe intrusion on the complaining witness's personal privacy and autonomy and that Mr. Beachum intended, knew, or should have known that his conduct on each occasion would have put a reasonable person in the complaining witness's circumstances to fear for her personal safety. Specifically, when considered in light of the conduct that had preceded them, the incident described in (3) above and the incident described in (5)-(9) would each in my view reasonably support such findings. Although the court expresses the view that Mr. Beachum's conduct was "no more objectively upsetting than [Dr.] Mashaud's," supra at 1166, I very much disagree. That is particularly true given that, for reasons I have explained, I focus in this case solely on Dr. Mashaud's initial email.

Finally, the court repeatedly criticizes this opinion for not reaching out to decide other issues not necessary in my view to decide this case, while at the same time itself declining to resolve numerous issues that the court views as unnecessary to resolve. E.g. , supra at 1166–67, 1168, 1168–69. It is certainly true that both opinions decline to decide a number of issues. My point is not that all issues that can be raised must be decided. Rather, my point is that an accurate comparison of the two approaches requires a full understanding of the complexities and issues that each approach would create. My further point is that one of the issues left unresolved by the court—whether the phrase "constitutionally protected activity," understood as a whole and in the context of § 22-3133(b), is reasonably susceptible to the interpretation favored by the court—must properly be decided now.

For the foregoing reasons, I respectfully dissent in part and concur only in the judgment.