United States v. Christopher Arthur

Opinion

PUBLISHED

UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

No. 24-4306

UNITED STATES OF AMERICA, Plaintiff – Appellee,

v. CHRISTOPHER CLARK ARTHUR, Defendant – Appellant.

Appeal from the United States District Court for the Eastern District of North Carolina, at Wilmington. James C. Dever, III, District Judge. (7:22-cr-00005-D-1)

Argued: September 10, 2025 Decided: December 3, 2025

Before GREGORY and AGEE, Circuit Judges and Roderick C. Young, United States Judge for the Eastern District of Virginia, sitting by designation.

Affirmed by published opinion. Judge Agee wrote the opinion in which Judge Young joined. Judge Gregory wrote a dissenting opinion.

ARGUED: Andrew DeSimone, OFFICE OF THE FEDERAL PUBLIC DEFENDER, Raleigh, North Carolina, for Appellant. Joseph Patrick Minta, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellee. ON BRIEF: G. Alan DuBois, Federal Public Defender, OFFICE OF THE FEDERAL PUBLIC DEFENDER, Raleigh, North Carolina, for Appellant. Sue J. Bai, Gavan W. Duffy Gideon, Appellate Section, National Security Division, UNITED STATES DEPARTMENT OF JUSTICE,

2

Washington, D.C.; Daniel P. Bubar, Acting United States Attorney, David A. Bragdon, OFFICE OF THE UNITED STATES ATTORNEY, Raleigh, North Carolina, for Appellee.

3

AGEE, Circuit Judge:

18 U.S.C. § 842(p)(2)(B) makes it unlawful to teach “any person” how to make or use “explosive[s],” “destructive device[s],” or “weapon[s] of mass destruction,” while “knowing that such person intends to use [that] teaching, demonstration, or information for, or in furtherance of, an activity that constitutes a Federal crime of violence.” (emphasis added).

Christopher Arthur was convicted under this statute and sentenced to 300 months in prison. He now brings a facial overbreadth challenge to § 842(p)(2)(B)’s constitutional validity—a question of first impression in this Circuit. In his view, the statute runs afoul of the First Amendment because it “regulates a substantial amount of constitutionally- protected speech.” Opening Br. 3. Arthur also contends that the district court erroneously applied U.S.S.G. § 3A1.4’s “terrorism enhancement” when issuing his sentence, leading it to calculate a significantly higher Sentencing Guidelines range than would otherwise have been applicable.

For the reasons discussed below, we find that § 842(p)(2)(B) is not unconstitutionally overbroad. Likewise, we find no reversible error in the district court’s decision to apply the terrorism enhancement. We therefore affirm the district court’s decision in full.

4

I.

A.

While Arthur’s main appellate challenge is a facial one, we provide a brief overview of the facts 1 as relevant to his sentencing challenge and to exemplify the type of fact pattern that may lead to a § 842(p)(2)(B) prosecution.

Arthur founded Tackleberry Solutions in October 2017 and initially planned for the company to sell military gear. Later, he shifted the company’s mission to “training,” with the goal of “help[ing] the average person to be able to defend themselves” against “a tyrannical government of our own or an invading tyrannical government.” J.A. 546–47. He published videos and sold manuals online with titles including “Fatal Funnels, Wartime Tactics, Repelling the Assault,” J.A. 385, and “Quick Reaction Force, Modern Day Minutemen, Improvised Explosives,” J.A. 324–25. The latter publication warned that “[t]hese are wartime tactics, not peacetime tactics. Some of these methods may be highly illegal. We do not advocate the use of this knowledge to hurt the innocent. Be sure to know the laws in your area.” J.A. 325.

The FBI began investigating Arthur following a fatal incident in June 2020 involving one of his customers, Joshua Blessed. While searching Blessed’s home in

1

Because the Government prevailed below, “we present the facts in the light most favorable to [it], drawing all reasonable inferences in its favor.” United States v. Bolden, 325 F.3d 471, 479 n.5 (4th Cir. 2003). With respect to the sentencing issues, “we review the facts in the light most favorable to the district court’s determinations.” Id.

5

Richmond, Virginia, the FBI found fourteen live pipe bombs that were identical to those described in Arthur’s manuals, as well as six manuals that he wrote.

Shortly thereafter, the FBI had a confidential informant—“Buckshot”—contact Arthur for training. Buckshot initiated contact with Arthur by watching the Tackleberry Solutions video, “How to Repel a Trained Military Force,” and requesting a free PDF from Arthur through a link in that video. J.A. 741. Arthur provided the PDF while noting that he “had to keep parts of this information off the internet. Especially since explosives are such a touchy topic.” Id. He also told Buckshot that there was “no digital copy of some of the things” he wanted to “share with [him].” Id.

Eventually, Arthur invited Buckshot to join him for in-person training, for which Buckshot would be charged a fee. Buckshot accepted his invitation and, upon his arrival, explained to Arthur that “[the] ATF’s been to my house. . . . [T]hey’re probably coming back. . . . [W]hen they do, I want to be ready.” J.A. 38. Arthur responded that he could “guaran-damn-tee” that the agents would return and that Buckshot had two choices: “Stand and fight or be, ah, not exactly where you’re supposed to be.” Id. Once Buckshot relayed that moving was not an option, Arthur spent the next three hours teaching him how to fortify his residence against the returning federal agents.

Arthur first recommended that Buckshot train attack dogs and build an electrified fence around his property. He explained that such a setup would create a “fatal funnel” that would delay law enforcement from approaching the house, at which point Buckshot could “start lobbing . . . grenades on them with [his] freaking [shotgun].” J.A. 63. Arthur also

6

suggested mounting cans of Tannerite 2 around the property that could be detonated with a rifle shot. In addition to this “perimeter defense,” Arthur suggested that it “wouldn’t be a bad idea” for Buckshot to “put[] some [improvised explosive devices (IEDs)] right up around the doors [of the house].” J.A. 56, 71. He noted that he kept such an IED on his front porch.

Arthur also suggested “a setup called the Spiderweb,” which he described as “a freaking death box.” J.A. 42. The “Spiderweb” involved blocking most entrances to Buckshot’s house and then placing remotely operated explosives near the remaining entrances, along with a “sentry gun” that could be remotely fired. J.A. 78–80. Arthur even went so far as to offer to “help [Buckshot] design [and] build it,” and later showed Buckshot how to use a lightbulb to make a detonator, to “[g]ive [him] something to start with.” J.A. 84, 91–92. The detonator would allow Buckshot to wire a switch to an explosive, as Arthur had done with the IED on his own front porch. They also discussed modifying a shotgun into a “thumper” that could launch homemade grenades. See J.A. 94–95. Buckshot paid Arthur for the training, and the two agreed to stay in contact.

Arthur was eventually arrested in January 2022 at a gun show where he had agreed to again meet Buckshot. Around the same time, FBI agents executed a search warrant on Arthur’s house and found numerous IEDs and firearms.

2

Tannerite is a commercially available explosive, commonly used to make exploding targets for marksmanship purposes. See J.A. 304.

7

B.

A grand jury indicted Arthur for violating 18 U.S.C. § 842(p)(2)(B) based on the training he provided to Buckshot (Count 1). Broadly speaking, that statute prohibits “any person” from “teach[ing] or demonstrat[ing] to any person” how to make or use explosives, while “knowing that such person intends to use” the information they learned “for, or in furtherance of,” a federal crime of violence. § 842(p)(2)(B). Arthur was also indicted on numerous other violations of 26 U.S.C. § 5861(d) and 18 U.S.C. § 922(k) in connection with the various explosives and illegal firearms found during the search of his house (Counts 2–9). 3 Arthur filed a pretrial motion to dismiss Count 1 on the ground that § 842(p)(2)(B)

is facially overbroad in violation of the First Amendment. The district court ultimately

3

Those statutes prohibit possession of unregistered firearms and receipt of firearms with obliterated serial numbers, respectively. See 26 U.S.C. § 5861(d); 18 U.S.C. § 922(k). Arthur was charged with the following crimes in Counts 2–9: knowing receipt and possession of an unregistered Palmetto State Armory PA-15 rifle having a barrel of less than 16 inches, in violation of 26 U.S.C. §§ 5841, 5845, 5861(d), 5871 (Count 2); knowing receipt and possession of an unregistered Gemtech Halo firearm silencer, in violation of 18 U.S.C. §§ 921(a)(3)(C) and (a)(24), and 26 U.S.C. §§ 5841, 5845(a)(7), 5861(d), 5871 (Count 3); knowing receipt and possession of an unregistered improvised explosive device rigged and wired to detonate with a 9-volt battery and switch, in violation of 26 U.S.C. §§ 5841, 5845(f), 5861(d), 5871 (Count 4); three counts of knowing receipt and possession of an unregistered improvised explosive device designed to detonate upon striking matchheads on a striking plate, in violation of 26 U.S.C. §§ 5841, 5845(f), 5861(d), 5871 (Counts 5, 6, and 7); knowing receipt and possession of an unregistered improved explosive device consisting of a green military-grade ammunition can and a 1-lb container of exploding target materials, in violation of 26 U.S.C. §§ 5841, 5845(f), 5861(d), 5871 (Count 8); and knowing possession of a Gemtech Halo firearm silencer that had been shipped and transported in interstate commerce, from which the manufacturer’s (importer’s) serial number had been removed, altered, and obliterated, in violation of 18 U.S.C. §§ 922(k), 924(a)(1)(B).

8

rejected this argument, holding that the statute is “not overbroad and does not restrict protected speech.” J.A. 142. That conclusion was based, in large part, on the court’s finding that § 842(p)(2)(B)’s “knowledge requirement effectively limits the scope of the statute,” leading it to “more closely resemble[] a criminal aiding and abetting statute.” J.A. 142; see J.A. 142–43 (rejecting Arthur’s reliance on Brandenburg v. Ohio, 395 U.S. 444 (1969) (per curiam), and noting that “the Fourth Circuit limits Brandenburg’s protection to mere or ‘abstract’ advocacy. . . . Section 842(p)(2)(B) and its knowledge requirement ensure the prohibited conduct goes well beyond ‘abstract advocacy’”).

At the ensuing three-day trial, the Government’s witnesses included five FBI special agents, an FBI chemist, an FBI explosives expert, and an ATF agent. These witnesses testified primarily about the search of Arthur’s residence and the items found there. The jury also heard a recording of Arthur’s meeting with Buckshot. Arthur testified in his own defense, admitting that he possessed all the items charged in Counts 2 through 9, and that he had produced Tackleberry Solution’s training materials. The case was then sent to the jury, which delivered a guilty verdict on all counts after two hours of deliberations.

At sentencing, the parties disputed whether Arthur’s conduct “involved, or was intended to promote, a federal crime of terrorism,” thereby triggering the enhancement set forth in U.S.S.G. § 3A1.4. The district court ultimately concluded that the enhancement applied because Arthur’s conduct “involve[d] distributing information knowing that it would be used to murder [or] attempt to murder an officer or employee of the United States in violation of 18 U.S.C. § 1114.” J.A. 775. Alternatively, the court concluded that the enhancement applied because Arthur’s conduct was intended to promote a violation of that

9

same section. See J.A. 775–76 (“The defendant was encouraging and contributing to the murder or attempted murder of the federal law enforcement officers.”). The court also explicitly found that one of Arthur’s purposes “was to [a]ffect the conduct of Government by coercion and to retaliate against Government conduct.” J.A. 776.

After rejecting Arthur’s other objections to the Presentence Investigation Report, the district court calculated his offense level to be 43 and his criminal history category to be VI, resulting in an advisory guidelines sentence of life imprisonment. Under U.S.S.G. § 5G1.2(d), this guidelines sentence was reduced to 1,140 months’ imprisonment—240 months for Count 1, 120 months for each of Counts 2 through 8, and 60 months for Count 9.

The district court then weighed the sentencing factors under 18 U.S.C. § 3553(a).

Based on all the evidence before it, the court determined that a downward variance from the guidelines sentence was warranted, but also emphasized that Arthur had admitted to significant criminal activity without “expressing remorse.” J.A. 797 (opining that “teaching someone to kill an ATF agent is seriously wrong and worthy of just punishment”). In the end, the court sentenced Arthur to 240 months for Count 1, 120 months for each of Counts 2 through 8 (to be served concurrently), and 60 months for Count 9 (to be served consecutively), for a total sentence of 300 months’ imprisonment, to be followed by a three-year term of supervised release. The court included, as part of its sentencing pronouncement, that it would have imposed the same sentence even if it had miscalculated the guidelines range.

10

Arthur timely noticed this appeal, and this Court has jurisdiction under 18 U.S.C.

§ 3742 and 28 U.S.C. § 1291.

II.

Arthur raises two issues on appeal, the first relating to whether 18 U.S.C.

§ 842(p)(2)(B) is facially overbroad in violation of the First Amendment, and the second relating to the district court’s application of a sentencing guideline enhancement.

The former issue is one that we review de novo. See United States v. Miselis, 972 F.3d 518, 525 (4th Cir. 2020) (utilizing de novo review for a First Amendment facial overbreadth challenge). The latter, by contrast, calls for a mixed standard of review: “In reviewing a Sentencing Guidelines application, we review factual findings for clear error and legal conclusions de novo.” United States v. Adepoju, 756 F.3d 250, 256 (4th Cir. 2014) (“Where a Guidelines application involves a mixed question of law and fact, the applicable standard turns on the nature of the circumstances at issue.”).

III.

Arthur’s primary argument is that 18 U.S.C. § 842(p)(2)(B) is facially overbroad in violation of the First Amendment. The Government disagrees, contending that Arthur failed to meet his burden of showing that § 842(p)(2)(B) “prohibits a substantial amount of protected speech relative to its plainly legitimate sweep.” Resp. Br. 14 (quoting United States v. Hansen, 599 U.S. 762, 770 (2023)). In our view, the Government is correct. To

11

explain why, we first provide a brief outline of First Amendment overbreadth principles, before applying those principles to the case at bar.

A.

The First Amendment provides that “Congress shall make no law . . . abridging the freedom of speech.” U.S. Const. amend. I. Here, Arthur does not argue that the First Amendment protects the communications for which he was prosecuted. Instead, he raises an overbreadth challenge: that § 842(p)(2)(B) punishes “so much protected speech that it cannot be applied to anyone, including him.” Hansen, 599 U.S. at 769 (emphasis in original); see Opening Br. 1.

Overbreadth challenges are unusual. To begin, “litigants typically lack standing to assert the constitutional rights of third parties.” Hansen, 599 U.S. at 769. What’s more, “litigants mounting a facial challenge to a statute normally ‘must establish that no set of circumstances exists under which the [statute] would be valid.’” Id. (emphasis in original) (quoting United States v. Salerno, 481 U.S. 739, 745 (1987)). Departing from both of these rules, the overbreadth doctrine instructs courts “to hold a statute facially unconstitutional even though it has lawful applications, and even at the behest of someone to whom the statute can be lawfully applied.” Id. (emphasis added).

Courts have typically justified the overbreadth doctrine on the ground that “it provides breathing room for free expression.” Id. The idea being, overbroad laws “‘may deter or “chill” constitutionally protected speech,’ and if would-be speakers remain silent, society will lose their contributions to the ‘marketplace of ideas.’” Id. at 769–70 (quoting Virginia v. Hicks, 539 U.S. 113, 119 (2003)). To shield against these harms, “the

12

overbreadth doctrine allows a litigant (even an undeserving one) to vindicate the rights of the silenced, as well as society’s broader interest in hearing them speak.” Id. at 770.

Because invalidation for overbreadth inevitably “destroys some good along with the bad,” it is a “strong medicine” that should not be “casually employed.” Id. (cleaned up). “To justify facial invalidation, a law’s unconstitutional applications must be realistic, not fanciful, and their number must be substantially disproportionate to the statute’s lawful sweep.” Id.; see N.Y. State Club Ass’n v. City of New York, 487 U.S. 1, 14 (1988); see Miselis, 972 F.3d at 531. In the absence of such a lopsided ratio, courts must handle unconstitutional applications in the typical manner—case-by-case. Hansen, 599 U.S. at 769.

B.

Overbreadth challenges usually proceed in several steps. First, a court must determine the scope of the challenged provision. Miselis, 972 F.3d at 532. In doing so, it “must seek to avoid any ‘constitutional problems’ by asking whether the statute is ‘subject to [] a limiting construction.’” Id. at 531 (quoting New York v. Ferber, 458 U.S. 747, 769 n.24 (1982)).

Next, “[it] must . . . determine whether, so construed, the statute ‘criminalizes a substantial amount of protected expressive activity.’” Id. (quoting United States v. Williams, 523 U.S. 285, 297 (2008)). The overbreadth challenger bears the “burden of demonstrating, from the text of the law and from actual fact, that substantial overbreadth exists.” Hicks, 539 U.S. at 122 (cleaned up). A statute’s overbreadth must not only be substantial “in an absolute sense, but also relative to [its] plainly legitimate sweep.”

13

Miselis, 972 F.3d at 531 (quoting Williams, 553 U.S. at 292). Additionally, at this step, the court must consider whether the proscribed speech falls within one of the few categories of unprotected speech. Id. at 532; see also Stevens, 559 U.S. at 468–69 (recognizing “obscenity, defamation, incitement, and speech integral to criminal conduct” as among the “well-defined and narrowly limited classes of speech, the prevention and punishment of which have never been thought to raise any Constitutional problem” (cleaned up)).

Finally, if the statute “proves ‘impermissibly overbroad,’ [the court] must assess whether ‘the unconstitutional portion’ is ‘severable’ from the remainder[.]” Miselis, 972 F.3d at 531 (quoting Ferber, 458 U.S. at 769 n.24). If so, only that portion of the law should be invalidated. Ferber, 458 U.S. at 769 n.24.

Altogether, “these efforts to preserve a statute from facial invalidation reflect the notion” that the overbreadth doctrine should only be applied as a last resort, “in cases where it is ‘truly warranted.’” Miselis, 972 F.3d at 531 (quoting Ferber, 458 U.S. at 769); see also id. at 530 (describing facial overbreadth challenges as “disfavored”).

C.

With the above principles as our guide, we turn to Arthur’s challenge. The statute he seeks to invalidate—18 U.S.C. § 842(p)(2)(B)—provides:

It shall be unlawful for any person . . . to teach or demonstrate to any person the making or use of an explosive, a destructive device, or a weapon of mass destruction, or to distribute to any person, by any means, information pertaining to, in whole or in part, the manufacture or use of an explosive, destructive device, or weapon of mass destruction, knowing that such person intends to use the teaching, demonstration, or information for, or in furtherance of, an activity that constitutes a Federal crime of violence.

14

This language is straightforward and substantially resolves step one of the overbreadth inquiry (i.e., determining what the statute covers): § 842(p)(2)(B) prohibits individuals from teaching or otherwise disseminating information relating to making and using explosives, destructive devices, and weapons of mass destruction. 4 But that prohibition exists only when the teacher/distributor of that information knows that its recipient (“such person”) intends to use it for, or in furtherance of, a federal crime of violence. In other words, Congress narrowed § 842(p)(2)(B)’s potentially broad scope by including a mens rea requirement of knowledge within it.

Having established the scope of conduct covered by § 842(p)(2)(B), we now turn to the second step of the overbreadth analysis: “determin[ing] whether, so construed, the statute ‘criminalizes a substantial amount of protected expressive activity.’” Miselis, 972 F.3d at 531 (quoting Williams, 523 U.S. at 297). Because we conclude that the answer to this question is no, Arthur’s overbreadth challenge fails.

To begin, it is not at all clear that § 842(p)(2)(B) prohibits any “protected expressive activity,” let alone a substantial amount. Id. That’s because the speech and “expressive” activity it proscribes fall largely within one of the “well-defined and narrowly limited

4

“[E]xplosive,” “destructive device,” and “weapon of mass destruction” are defined elsewhere in the United States Code. See 18 U.S.C. § 842(p)(1)(A)–(C) (defining these terms by reference to other statutory provisions). The dissent makes much of these definitions, arguing that they broaden § 842(p)(2)(B)’s scope. See Diss. Op. 34–35. But even if that’s true, § 842(p)(2)(B)’s reach is still substantially limited by its knowledge requirement, as discussed below.

15

classes of [unprotected] speech”: speech integral to criminal conduct. Stevens, 559 U.S. at 468–69 (quoting Chaplinsky v. New Hampshire, 315 U.S. 568, 571–72 (1942)).

The “speech integral to criminal conduct” exception arose in Giboney v. Empire Storage & Ice Co., 336 U.S. 490 (1949). There, unions planned to picket to force a distributor into an agreement that violated state anti-trade restraints law, but a state court prevented them from doing so. Id. at 491–92. The Supreme Court upheld the state court’s ruling, concluding that “the constitutional freedom for speech and press” does not “extend[] its immunity to speech or writing used as an integral part of conduct in violation of a valid criminal statute.” Id. at 498. On this point, the Court observed that the First Amendment does not prohibit legislatures from “mak[ing] a course of conduct illegal merely because the conduct was in part initiated, evidenced, or carried out by means of language, either spoken, written, or printed.” Id. at 502. To hold otherwise would make it “practically impossible” to enforce laws against speech-adjacent conduct “deemed injurious to society.” Id.

Since Giboney, the speech integral to criminal conduct exception has been held to encompass a wide variety of crimes that are often committed through speech. Those crimes include—but are not limited to—conspiracy, solicitation, perjury, extortion, and aiding and abetting. See Rice v. Paladin Enters., 128 F.3d 233, 244 (4th Cir. 1997) (collecting examples and noting that “[w]ere the First Amendment to bar or [] limit government regulation of such speech brigaded with action, the government would be powerless to protect the public from countless of even the most pernicious criminal acts and civil

16

wrongs” (cleaned up)); see also Hansen, 599 U.S. at 783 (collecting examples including promotion of contraband, solicitation of unlawful employment, and illegal picketing).

Here, we have little trouble concluding that § 842(p)(2)(B) falls within the speech integral to criminal conduct exception. Simply put, one who teaches another how to make “explosive[s],” “destructive device[s],” or “weapon[s] of mass destruction,” while knowing that the recipient of that information intends to use it to commit a federal crime of violence, has effectively facilitated the commission of the other’s crime. 18 U.S.C. § 842(p)(2)(B). That is, but for the proscribed communications, the other person would lack the means to commit their intended crime. See id. (prohibiting “teach[ing],” “demonstra[ting],” and “distribut[ing]” certain information when it is known that the recipient “intends to use the teaching, demonstration, or information for, or in furtherance of, . . . a federal crime of violence”). Those communications are therefore necessary—or “integral”—to the other person’s intended crime. See Integral, Cambridge English Dictionary (last accessed Sept. 16, 2025) https://dictionary.cambridge.org/us/dictionary/english/integral [https://perma.cc/A7WM-KQ7Q] (defining “integral” as “necessary and important as a part of a whole”); Integral, Merriam-Webster’s Dictionary (last accessed Sept. 16, 2025) https://www.merriam-webster.com/dictionary/integral [https://perma.cc/5HHE-8T5F] (defining integral as “essential to completeness”). The facts in this case illustrate that general point: Buckshot told Arthur he wanted to kill ATF agents who were bothering him, but he needed Arthur’s guidance to make that a reality. Arthur, in turn, provided Buckshot with all the information he needed to accomplish that objective. Arthur’s actions were thus

17

integral to Buckshot’s intended crime. 5 In effect, but for Arthur’s instruction, Buckshot could not commit his intended crime.

An additional hypothetical helps to crystallize this point. Suppose that, rather than seek out guidance on how to make explosives, Buckshot went to Arthur in search of the explosives themselves. Suppose also that Buckshot relayed his nefarious intent to murder federal agents to Arthur, and that Arthur provided Buckshot with the explosives he needed to carry out his plot. Under those circumstances, Arthur’s provision of the explosives would certainly be deemed integral to Buckshot’s intended crime. After all, he needed explosives to commit the crime, and Arthur provided them. The facts here—and under § 842(p)(2)(B) more generally—are functionally no different. Arthur taught Buckshot everything he needed to know to create the explosives he desired, and he did so knowing that Buckshot intended to use those explosives to commit a federal crime of violence. Arthur’s teachings were thus integral to Buckshot’s intended crime. See Rice, 128 F.3d at 243 (“[I]t is . . . well established that speech which, in its effect, is tantamount to legitimately proscribable nonexpressive conduct may itself be legitimately proscribed, punished, or regulated incidentally to the constitutional enforcement of generally applicable statutes.”).

Further bolstering this conclusion is this Court’s decision in United States v. Miselis.

There, the appellant brought a First Amendment facial overbreadth challenge to the Anti- Riot Act. Miselis, 972 F.3d at 525. While our analysis largely hinged on Brandenburg, 395

5

To be clear, Arthur raises only a facial challenge. The facts of this case merely illustrate how and why the speech integral to criminal conduct exception applies in the context of a § 842(p)(2)(B) prosecution.

18

U.S. 444, and its incitement-related progeny, we made an additional observation that is instructive here. Namely, we noted that “[s]peech taking some form ‘other than abstract advocacy,’ . . . such as that which ‘constitutes . . . aiding and abetting of criminal conduct,’ doesn’t implicate the First Amendment.” Id. at 533 (quoting Rice, 128 F.3d at 239, 242– 43). In doing so, we “effectively recognize[d] a second category of unprotected speech”— aiding and abetting, and other crimes of its ilk—“which may be proscribed without regard to whether it’s directed [at] and likely to produce imminent lawlessness.” Id.

The speech and activities prohibited by § 842(p)(2)(B) are simply a far cry from the sort of protected “abstract advocacy” contemplated in Miselis and Brandenburg. See id.; Williams, 553 U.S. at 299–300 (suggesting that Brandenburg only protects “abstract advocacy”). Instead, the statute’s prohibitions are much more akin to the type of “aiding and abetting of criminal conduct” that this Court has held may be limited without running afoul of the First Amendment. Miselis, 972 F.3d at 533; see Rice, 128 F.3d at 244–45. In fact, the only real difference between aiding and abetting, on the one hand, and the conduct proscribed by § 842(p)(2)(B), on the other, is that the former requires the abettor to share the same criminal intent as the principal perpetrator. See, e.g., Rosemond v. United States, 572 U.S. 65, 71 (2014) (“As at common law, a person is liable . . . for aiding and abetting a crime if . . . he (1) takes an affirmative act in furtherance of that offense, (2) with the intent of facilitating the offense’s commission.”). Section 842(p)(2)(B), by contrast, requires knowledge that the recipient of the bombmaking information intends to commit a crime. See 18 U.S.C. § 842(p)(2)(A). But for purposes of the First Amendment, this is a distinction without a difference. That’s because, practically speaking, someone violating

19

§ 842(p)(2)(B) is aiding—i.e., facilitating—the underlying crime by intentionally sharing the specified information with someone that they know intends to use it to commit a proscribed crime. And because that sort of facilitation is undoubtedly “integral” to the underlying crime, it is unprotected speech. See Hansen, 599 U.S. at 783. 6 The only other circuit to have considered a First Amendment challenge similar to that raised here rejected it. In National Mobilization Committee to End War in Viet Nam v. Foran, the Seventh Circuit considered a facial challenge to 18 U.S.C. § 231(a)(1)—a statute with language akin to that in § 842(p)(2)(B). 7 411 F.2d 934 (7th Cir. 1969). In rejecting the facial challenge raised, the court held that the challengers improperly “ignore[d] the ‘knowing, or having reason to know or intending’ language of the statute.” Id. at 937. The court went on to state that “[t]he requirement of intent of course narrows the scope of the enactment by exempting innocent or inadvertent conduct from its proscription.” Id. So too here. Section 842(p)(2)(B)’s knowledge requirement restricts its reach and ensures that one who disseminates bombmaking information to another without knowledge of the recipient’s illegal motives remains protected by the First Amendment. See Model Penal Code § 2.02(7) (“A person acts knowingly with respect to a material

6

In any event, it seems plausible that § 842(p)(2)(B), when implicated, will apply in aiding and abetting situations because of its knowledge requirement. And as noted above, aiding and abetting crimes have explicitly been recognized as falling within the speech integral to criminal conduct exception. See, e.g., Rice, 128 F.3d at 244.

7

Section 231(a)(1) prohibits “teach[ing] or demonstrat[ing] to any other person the use, application, or making of any firearm or explosive or incendiary device, . . . knowing or having reason to know or intending that the same will be unlawfully employed for use in, or in furtherance of, a civil disorder.”

20

element of an offense when: if the element involves . . . the attendant circumstances, he is aware that . . . such circumstances exist.” (emphasis added)). 8 Finally, we reiterate that “[t]o justify facial invalidation, a law’s unconstitutional applications must be realistic, not fanciful, and . . . substantially disproportionate to the statute’s lawful sweep.” Hansen, 599 U.S. at 770. For the reasons discussed above (and below), we hold that many, if not all, of § 842(p)(2)(B)’s applications fall within the “speech integral to criminal conduct” exception and thus outside the First Amendment’s ambit. 9 But even if some of its applications theoretically do not fall within that exception, Arthur has still fallen well short of carrying his burden of showing that the number of purportedly unconstitutional applications of this statute is “substantially disproportionate to [its] lawful sweep.” Id. (cleaned up).

To recap, there are several longstanding exceptions to the protections afforded by the First Amendment. One such exception relates to “speech integral to criminal conduct.” Because the communications proscribed by § 842(p)(2)(B) are primarily that sort of unprotected speech, Arthur’s facial overbreadth challenge must fail.

8

The Fifth Circuit faced a somewhat similar First Amendment challenge in United States v. Featherston, 461 F.2d 1119 (5th Cir. 1972) (relying, in part, on Foran to reject a vagueness challenge to 18 U.S.C. § 231(a)). However, because Featherston primarily concerned a vagueness challenge rather than an overbreadth one, it is not on all fours with this case.

9

Some applications of § 842(p)(2)(B) may even fall within the incitement exception discussed in Brandenburg and Miselis, particularly in factual scenarios like Arthur’s. The potential application of that exception further broadens the scope (and relative share) of unprotected speech prohibited by the statute.

21

D.

Arthur raises various arguments in an attempt to support his facial overbreadth challenge, but none have merit. 10 First, he argues that the “speech integral to criminal conduct” exception does not save § 842(p)(2)(B) from invalidation because the “examples of [such speech] . . . [all] require the defendant’s specific intent to commit a crime, which is absent from this statute.” Reply Br. 4; see Reply Br. 4–9. It is correct that many of the examples of speech integral to criminal conduct include some sort of specific intent to commit an underlying crime. See, e.g., Hansen, 599 U.S. at 771 (noting that “aiding and abetting . . . is the provision of assistance to a wrongdoer with the intent to further an offense’s commission”). But Arthur overstates the import of this observation, as neither the Supreme Court nor this Court has ever limited this exception to only apply where the defendant possesses a specific intent to commit an underlying crime. See, e.g., Rice, 128 F.3d at 248 (recognizing that “the First Amendment may in some contexts stand as a bar to the imposition of liability on the basis of mere foreseeability or knowledge that the information one imparts could be misused,” but expressly declining to “engage in an extended discussion of the existence or scope” of any such “intent-based limitation” (emphasis added)). And with good reason—it operates as a relatively flexible exception. See Giboney, 336 U.S. at 688–89; Rice, 128 F.3d at 243– 44 (acknowledging the various scenarios in which the speech integral to criminal conduct

10

The dissent raises many of the same points as Arthur. See, e.g., Diss. Op. 35–43.

Suffice it to say, those arguments fall flat for the same reasons that Arthur’s do, as explained herein.

22

exception has been extended); United States v. Osinger, 753 F.3d 939, 950 (9th Cir. 2014) (Watford, J., concurring) (noting that the boundaries of this exception “have not been clearly defined,” leaving its “precise scope . . . unsettled”). The main limiting principle for this exception is in its substance—that is, whether the speech was truly integral to the criminal conduct in question. And for the reasons already discussed, that requirement is plainly met with § 842(p)(2)(B).

Arthur also contends that the Government “regularly proves guilty knowledge” with ease. Reply Br. 17; See Reply Br. 17–23. It is therefore problematic, his argument goes, that § 842(p)(2)(B) requires only “knowledge” of another’s intent to commit a federal crime of violence. There are at least two defects with this line of argument. First, it’s not at all clear that the sort of knowledge proscribed by § 842(p)(2)(B) is as easy to prove as Arthur suggests. See Flores-Figueroa v. United States, 556 U.S. 647, 655 (2009) (acknowledging the “difficulty in many circumstances of proving beyond a reasonable doubt that a defendant has the [requisite] knowledge” to be convicted). After all, this case presents an issue of first impression. If guilty knowledge was so easy to prove in this context, it seems likely this issue would have arisen previously in the subsections 26-year history. But as counsel acknowledged at argument, this appears to be one of the first ever prosecutions under § 842(p)(2)(B). See Oral Arg. 2:20:52. And second, nearly all of the cases marshalled by Arthur in support of this point arose after a guilty verdict. See Reply Br. 17–23. The appellate court in each instance therefore needed to draw all inferences in favor of the guilty verdict. In other words, the deck was stacked against reversal. But at a trial in the first instance, it won’t be axiomatic that a relatively weak prosecution will lead

23

to a guilty verdict. By relying on cases in this posture, Arthur overstates any potential chilling effect of § 842(p)(2)(B).

Next, Arthur argues that two of the main cases cited by the Government—

Featherston and Foran—are distinguishable because they (1) did not deal with the precise issue raised here and (2) involved a limited analysis of the legal issues. On this point, he’s somewhat correct. Featherston held only that a similar statute (18 U.S.C. § 231(a)) was “sufficiently definite to apprise men of common intelligence of its meaning and application.” 461 F.2d at 1122. That is, it rejected a vagueness challenge, not a facial overbreadth one. See id. Foran, by contrast, did involve a facial overbreadth challenge to § 231(a). And there, the court held that the statute passed constitutional muster: “[Appellants] ignore[] the knowing, or having reason to know or intending language of the statute. The requirement of intent of course narrows the scope of the enactment by exempting innocent or inadvertent conduct from its proscription.” Foran, 411 F.2d at 937 (internal quotations and citation omitted). This language suggests that the “intent” requirement of § 231(a) was relevant to the court’s holding in Foran, but as Arthur points out, § 842(p)(2)(B) does not have a separate intent requirement.

Nonetheless, the main reason Arthur’s challenge fails—the speech integral to criminal conduct exception—was not at issue, much less addressed, in either decision. So those decisions have little bearing here. Moreover, Foran’s language about an intent requirement “narrow[ing] the scope of the enactment” does support our conclusion. Id. In that respect, § 842(p)(2)(B)’s “knowledge” requirement narrows its scope, just as § 231(a)(1)’s intent requirement narrowed its scope.

24

Finally, Arthur unleashes a barrage of hypotheticals to argue that § 842(p)(2)(B) is facially overbroad in violation of the First Amendment. See, e.g., Opening Br. 16–17; Reply Br. 16–17; Oral Arg. 2:24:30–2:30:50. His hypotheticals can be broken down into two categories: (1) those raised in his briefing, and (2) those raised at oral argument. Neither group moves the needle in his favor.

Beginning with the briefing hypotheticals, Arthur rightly conceded at argument that they were unpersuasive. Recall that for facial invalidation to be warranted, “a law’s unconstitutional applications must be realistic, not fanciful, and . . . substantially disproportionate to the statute’s lawful sweep.” Hansen, 599 U.S. at 770. Recall also that Arthur bears the burden of making this showing. Id. He failed to do so, offering only far- fetched hypothetical scenarios in his briefing. See id.; see, e.g., Opening Br. 16–17 (suggesting that § 842(p)(2)(B) will “chill military instructors’ willingness to teach necessary information out of fear that, if any class member is later revealed to be a member of a gang or extremist group, the instructor could be prosecuted”); id. at 17 (arguing that physics teachers “may well refrain from teaching their students basic concepts related to explosions” out of concern that their students may “use that information to commit violence”). And in any event, none of these hypotheticals implicate the statutory “knowledge” element.

The hypotheticals Arthur raised for the first time at argument are no more persuasive. In each instance, he posited a factual scenario in which the speaker taught or otherwise disseminated bombmaking information to a broad audience. See Oral Arg. 22:24:42–2:30:50. He further posited that, at some point after dissemination, the original

25

speaker became aware of at least one individual who intended to use that information to commit a crime. Arthur contends that this state of affairs would leave the speaker with a Hobson’s choice: face criminal liability under § 842(p)(2)(B), or recall the information in question and, in doing so, have their speech chilled.

While these hypotheticals are marginally less fanciful than those raised in Arthur’s briefing, they remain unconvincing. For one, in each instance, the speaker did not disseminate the relevant information with knowledge that another person intended to use it to commit a crime of violence. Rather, that knowledge only came about afterward. So, the speaker’s original “teach[ings] or “demonstrat[ions]” would not be covered by the statute. See § 842(p)(2)(B).

Moreover, it’s not clear that failing to take the information down after the fact would lead to criminal liability. Again, § 842(p)(2)(B) contemplates that the speaker knows—at the time they “teach or demonstrate . . . the making or use of an explosive, [] destructive device, or [] weapon of mass destruction,” or “distribute . . . information pertaining to” the same—that a specific person intends to use that specific information to commit a crime of violence. 18 U.S.C. § 842(p)(2)(B). But in the later hypotheticals Arthur provided, the speaker did not know of the other person’s intent until after the fact. In other words, the speaker’s “teach[ing]” and “distribut[ion]” had effectively concluded before the other individual obtained the requisite knowledge. And because § 842(p)(2)(B) requires that

26

knowledge to exist at the time of the proscribed teaching or distribution, a prosecution would not be successful under such circumstances. 11, 12 In the end, Arthur is unable to carry his burden of proving that § 842(p)(2)(B) is facially overbroad. We therefore reject his First Amendment facial overbreadth challenge.

11

An example helps to illustrate this point. Imagine Person A sends a book on bombmaking to a group of ten people. Afterward, one of those people—Person B— contacts Person A to let him know that he plans to use the information in the book to commit a crime of violence. In this situation, Person A has not violated § 842(p)(2)(B). At the time he “taught” the group, he did not know that Person B intended to use the information to commit a crime of violence. And once he did know, his “teaching” had concluded (i.e., he’d already sent the book). Because the teaching and knowledge did not happen or exist simultaneously, Person A did not violate § 842(p)(2)(B). The same would be true under the hypotheticals raised by Arthur.

12

Briefly, the Government argues that § 824(p)(2)(B)’s statutory counterpart—

§ 842(p)(2)(A)—is the more applicable provision for the broad dissemination contemplated by Arthur’s hypotheticals at oral argument. While we need not reach the issue, the Government raises a good point.

Section 842(p)(2)(A) broadly prohibits “any person” from “teach[ing] or demonstrat[ing] the making or use of an explosive, a destructive device, or a weapon of mass destruction, or to distribute by any means information pertaining to . . . the manufacture or use of an explosive, destructive device, or weapon of mass destruction, with the intent that the teaching, demonstration, or information be used for, or in furtherance of,” a federal crime of violence.

By contrast, § 842(p)(2)(B) prohibits “any person” from “teach[ing] or demonstrating to any person” the making or use of those same weapons, and “distribut[ing] to any person, by any means, information pertaining to . . . the manufacture or use” of those weapons, while “knowing that such person intends to use the teaching, demonstration, or information for, or in furtherance of,” a federal crime of violence. (emphasis added)

Even setting aside the obvious distinction between the two statutes—the differing mens rea requirements—there is another crucial difference: § 842(p)(2)(B) requires that the teaching, demonstration, or dissemination of information be “to any person,” with the knowledge that “such person” will use it improperly. This suggests that the conduct proscribed by § 842(p)(2)(B) must be directed toward some specified person(s), rather than just sent out into the ether. Section 842(p)(2)(A) contains no such limiter, and thus more naturally reads as being aimed at the sort of broad dissemination Arthur posits in his hypotheticals.

27

IV.

Arthur’s second appellate challenge relates to the district court’s imposition of a terrorism enhancement under U.S.S.G. § 3A1.4. The district court concluded that this enhancement applied because Arthur’s conduct “involve[d] distributing information knowing that it would be used to murder [or] attempt to murder an officer or employee of the United States in violation of 18 U.S.C. § 1114.” J.A. 775. It alternatively concluded that the enhancement applied because Arthur’s conduct was intended to promote a violation of that same section. We discern no reversible error in these conclusions.

U.S.S.G. § 3A1.4 instructs district courts to increase a defendant’s offense level and criminal-history category “[i]f the offense is a felony that involved, or was intended to promote, a federal crime of terrorism.” A “federal crime of terrorism,” in turn, “consist[s] of two elements: (1) the commission of one of a list of specified felonies . . . and (2) a specific intent requirement, namely, that the underlying felony was ‘calculated to influence or affect the conduct of government by intimidation or coercion, or to retaliate against government conduct.’” United States v. Chandia, 514 F.3d 365, 376-77 (4th Cir. 2008) (quoting 18 U.S.C. § 2332b(g)(5)); see also U.S.S.G. § 3A1.4 cmt. n.1 (instructing that “federal crime of terrorism” in the Guidelines has the meaning given the term in 18 U.S.C. § 2332b(g)(5)).

Here, all agree that the federal crime of violence that Arthur knew Buckshot intended to commit—the murder or attempted murder of federal law enforcement officers—qualifies as a federal crime of terrorism (via 18 U.S.C. §§ 1114 and 2332b(g)(5)). Opening Br. 24; Resp. Br. 35. The only question remaining is whether Arthur’s offense,

28

18 U.S.C. § 842(p)(2)(B), “involved, or was intended to promote” Buckshot’s federal crime of terrorism.

To reiterate, there are two ways in which the § 3A1.4 enhancement can be triggered:

(1) if the defendant’s offense is a felony that “involved” a “federal crime of terrorism,” or (2) if the defendant’s offense is a felony that “intended to promote[] a federal crime of terrorism.” U.S.S.G. § 3A1.4; see United States v. Kobito, 994 F.3d 696, 702 (4th Cir. 2021). As this Court recognized in Kobito, the “‘involve’ prong . . . suggests that the predicate offense and related conduct must include a federal crime of terrorism.” 994 F.3d at 702. The “intended to promote” prong has a slightly different sweep, applying “whenever the defendant commits a felony with ‘a goal or purpose . . . to bring or help bring into being a [federal crime of terrorism]’ . . . even if the defendant ‘has not necessarily completed, attempted, or conspired to commit [that] crime.’” Id. (first quoting United States v. Mandhai, 375 F.3d 1243, 1248 (11th Cir. 2004); and then quoting United States v. Graham, 275 F.3d 490, 516 (6th Cir. 2001)). The district court correctly recognized that either prong applies on the facts of this case.

First, the “involve” prong. Arthur’s crime necessarily “involved” a federal crime of terrorism because he provided Buckshot with information integral to Buckshot’s intended federal crime of terrorism. Not only that, but Arthur did so with the knowledge that Buckshot would use that information to commit that crime. Indeed, the government had to—and did—prove a close link between the training that Arthur gave Buckshot and Buckshot’s purported intent to murder federal officials. Arthur’s crime therefore

29

“involved” or “ha[d] within or as part of itself” Buckshot’s intended federal crime of terrorism. Id.

Arthur’s offense also “intended to promote” a federal crime of terrorism for essentially the same reasons. That is, he violated § 842(p)(2)(B) by teaching Buckshot how to create certain bombs and explosives. And the record reflects that he did so “with ‘a goal or purpose . . . to help bring into being’” Buckshot’s intended federal crime of terrorism (again, 18 U.S.C. § 1114). Id. The facts bear this out: Buckshot told Arthur what his intentions were, and Arthur knowingly taught him how to bring his plans to fruition. See J.A. 776 (finding that Arthur’s offense intended to promote a violation of 18 U.S.C. § 1114 by “encouraging and contributing to the murder of . . . federal law enforcement officers,” insofar as he taught Buckshot how to build explosives that he described variously as “stupid lethal” and a “‘freaking death box’ for ATF agents”); see also id. (finding that Arthur “reasonably foresaw the use of [those explosives and] tactics to kill ATF agents”); Graham, 275 F.3d at 518–19 (holding that the defendant’s 18 U.S.C. § 371 conspiracy was “intended to promote” a federal crime of terrorism where the targets of the latter crime “were reasonably foreseeable to the defendant”). 13 Arthur resists this conclusion, but his arguments are meritless. For instance, he argues that his crime “did not involve a crime of terrorism because it did not include a crime of terrorism.” Opening Br. 26. But that’s not true. His crime did such a crime:

13

And under Kobito, it is irrelevant that Arthur himself did “not . . . complete[], attempt[], or conspire[] to commit [a federal] crime [of terrorism].” Kobito, 994 F.3d at 702 (cleaned up).

30

Buckshot’s intention to kill federal law enforcement agents, in violation of 18 U.S.C. § 1114. That offense, and Arthur’s knowledge thereof, is necessarily included—i.e., “involved”—in the offense Arthur committed. See 18 U.S.C. § 842(p)(2)(B).

Arthur also argues that the district court failed to make certain findings relevant to its conclusion on the “intended to promote” prong of U.S.S.G. § 3A1.4. See Opening Br. 27 (“[T]he district court found [only] that Mr. Arthur’s ‘purpose, among others, in teaching the confidential human source about how to make IEDs was to [a]ffect the conduct of Government by coercion and to retaliate against Government conduct.’ But crucially, the district court never made the requisite finding that Mr. Arthur intended for his teaching to be used to kill ATF agents.” (emphasis in original)). Yet this argument relies on a misreading of the record, as the district court did make the requisite findings. First, it found that Arthur intended “to [a]ffect the conduct of Government by coercion and to retaliate against Government conduct[,]” satisfying 18 U.S.C. § 2332b(g)(5)’s specific intent requirement. J.A. 776. And the district court further found that Arthur “help[ed] bring into being” a violation of 18 U.S.C. § 1114 by “encouraging and contributing to the murder or attempted murder” of federal law enforcement officials. J.A. 776 (emphasis added); see Kobito, 994 F.3d at 702.

Finally, even assuming Arthur is correct that the district court erred in applying § 3A1.4’s terrorism enhancement, any error was harmless. A guidelines error is harmless if the record shows that “(1) ‘the district court would have reached the same result even if it had decided the Guidelines issue the other way,’ and (2) ‘the sentence would be

31

reasonable even if the Guidelines issue had been decided in the defendant’s favor.’” United States v. Mills, 917 F.3d 324, 330 (4th Cir. 2019).

Here, there’s no dispute that the first requirement is satisfied, as the district court explicitly stated it would have “impose[d] the same sentence as an alternative variant sentence if [it] ha[d] in any way miscalculated the advisory guidelines range.” J.A. 799– 800. The second requirement calls for slightly more analysis, as we must “examine[] the totality of the circumstances”—including the extent of any variance—“to see whether the sentencing court abused its discretion in concluding that the sentence it chose satisfied the standards set forth in [18 U.S.C.] § 3553(a).” Mills, 917 F.3d at 331 (quoting United States v. Gomez-Jimenez, 750 F.3d 370, 383 (4th Cir. 2014)). And in doing so, we must “give due deference to the district court’s decision that the § 3553(a) factors, on a whole, justify the extent of the variance.” Id. (quoting Gall v. United States, 552 U.S. 38, 51 (2007)).

Applying these principles to the facts of this case, we find that the district court’s individualized application of the § 3553(a) factors supports its conclusion that, even if the applicable advisory range was instead 168–210 months, 14 a variance to a 300-month sentence was “necessary to satisfy the basic aims of sentencing.” Id. To that end, the record reflects that the district court walked through all the relevant factors and properly exercised its discretion in fashioning Arthur’s sentence. See J.A. 793–800 (considering, among other things, the nature and circumstances of the offense, Arthur’s history and characteristics,

The parties agree that this would be the proper guidelines range if the terrorism

14

enhancement did not apply. See Resp. Br. 41; Reply Br. 29

32

and the need to promote respect for the law and provide just punishment). Under these circumstances, we have little issue concluding that Arthur’s sentence “would be reasonable even if the Guidelines issue had been decided in [his] favor.” Mills, 917 F.3d at 330. Any error attending the Guidelines issue raised by Arthur is therefore harmless.

In sum, Arthur advances a cramped reading of both the record and U.S.S.G.

§ 3A1.4’s “involved” or “intended to promote” requirement. The district court correctly rejected his arguments on its way to applying § 3A1.4’s enhancement. And even if it had erred, any error was harmless for the reasons just stated. We therefore affirm the sentence imposed by the district court.

V.

For the foregoing reasons, the district court’s judgment is

AFFIRMED.

33

GREGORY, Circuit Judge, dissenting:

Christopher Clark Arthur brings this challenge to 18 U.S.C. § 842(p)(2)(B), arguing that the statute is facially overbroad in violation of the First Amendment. Because I have concerns about the wide range of speech encompassed by the statutory language as well as the novelty of criminalizing speech when the speaker lacked specific intent to commit a crime, I cannot agree with my colleagues that this statute should survive overbreadth analysis.

I.

Given the concern that overbroad laws “may deter or chill constitutionally protected speech—especially when the overbroad statute imposes criminal sanctions,” courts allow First Amendment facial challenges “based merely on hypothetical applications of the law to nonparties . . . even if the law is constitutional as applied to the plaintiff.” Virginia v. Hicks, 539 U.S. 113, 119 (2003); Preston v. Leake, 660 F.3d 726, 738-39 (4th Cir. 2011).

The first step in a facial overbreadth analysis “is to construe the challenged statute.”

United States v. Williams, 553 U.S. 285, 293 (2008). Section 842(p)(2)(B) prohibits “distribut[ing] to any person, by any means, information pertaining to, in whole or in part, the manufacture or use of an explosive, destructive device, or weapon of mass destruction” so long as the distributor “know[s]” that a recipient intends to use the information in furtherance of a federal crime of violence. 18 U.S.C. § 842(p)(2)(B). The majority asserts that Arthur’s conduct—which involved instructing a confidential informant how to set up deadly explosive defenses on his property—“exemplif[ies] the type of fact pattern that may

34

lead to a § 842(p)(2)(B) prosecution.” Majority Op. at 4. But as the majority recognizes, Arthur does not bring a challenge to the statute as applied to his own conduct; our task is to determine whether the law on the whole “criminalizes a substantial amount of protected activity.” Williams, 553 U.S. at 293.

To appreciate the full scope of the statute, we must consider how it defines the terms “explosive” and “destructive device.” The term “explosive” means

gunpowders, powders used for blasting, all forms of high explosives, blasting materials, fuzes (other than electric circuit breakers), detonators, and other detonating agents, smokeless powders, other explosive or incendiary devices within the meaning of paragraph (5) of section 232 of this title, and any chemical compounds, mechanical mixture, or device that contains any oxidizing and combustible units, or other ingredients, in such proportions, quantities, or packing that ignition by fire, by friction, by concussion, by percussion, or by detonation of the compound, mixture, or device or any part thereof may cause an explosion.

18 U.S.C. §§ 842(p)(1), 844(j). We have held that gasoline, even outside a pressurized container, qualifies as an explosive for the purpose of this section, United States v. Lee, 726 F.2d 128, 130–31 (4th Cir. 1984), and several of our sister circuits have issued even more expansive holdings. See, e.g., United States v. Hepp, 656 F.2d 350, 352 (8th Cir. 1981) (holding that releasing methane into the air creates an “explosive” under this statute). The section defining “destructive device” lists a number of particular items which that category comprises, but it also includes a catchall provision that incorporates into the definition “any combination of parts either designed or intended for use in converting any device into any destructive device.” 18 U.S.C. §§ 842(p)(1), 921(a)(4). While the facts of Arthur’s case might imply that the statute is solely concerned with bombs, the reality of the statutory language is that it applies to far more common and less threatening items.

35

The breadth of these definitions is exacerbated by the breadth of the rest of Section 842(p)(2)(B). The statute does not restrict criminal punishment only to those who publish assembly instructions for explosives and destructive devices. Section 842(p)(2)(B) prohibits distributing information “pertaining to, in whole or in part, the manufacture or use” of an explosive or destructive device. 18 U.S.C. § 842(p)(2)(B) (emphasis added). The law is hardly cabined to those who communicate bombmaking instructions; the statute prohibits discussion of anything pertaining, in part to an explosive or destructive device. Consider a discussion about the production of methane gas, which our sister circuits have held need only be released into the air to create an “explosive.” By the plain language of the statute, that discussion qualifies as “pertaining . . . in part, to the manufacture . . . of an explosive.” 18 U.S.C. § 842(p)(2)(B). Or consider a university professor giving a lecture on simple physics topics; the language of the statute is likely to cover this speech as well, since the concepts of potential and kinetic energy easily “pertain[] to” incendiary applications. A statute that couples vague definitions with strict prohibitions leaves enormous discretion in the hands of the government to penalize speech.

II.

Since the statute prohibits a wide range of speech, we must consider whether the mens rea requirement sufficiently narrows its scope to ensure it does not penalize too much protected speech.

A conviction for violating Section 842(p)(2)(B), requires the perpetrator “know[]”

that a recipient of the information “intends” to use it “for, or in furtherance of, an activity

36

that constitutes a Federal crime of violence.” 18 U.S.C. § 842(p)(2)(B). Whether mere knowledge—without intent—establishes the speaker’s sufficient culpability so as to justify withholding First Amendment protection has not been squarely decided by this Court. The only case on this point cited by the majority is National Mobilization Committee to End War in Viet Nam v. Foran, 411 F.2d 934, 937 (7th Cir. 1969), in which the Seventh Circuit devoted one paragraph to a similar question. But that opinion was issued before the Supreme Court’s seminal decision in Brandenburg v. Ohio clarified that the First Amendment protects advocacy of lawlessness. 395 U.S. 444 (1969). Until today, this Court has refrained from deciding whether a speaker’s mere knowledge of a listener’s criminal intent is sufficient to permit civil liability. See Rice v. Paladin Enters., 128 F.3d 233, 247 (4th Cir. 1997) (“[T]he First Amendment may in some contexts stand as a bar to the imposition of liability on the basis of mere foreseeability or knowledge that the information one imparts could be used for an impermissible purpose.”). When the law imposes criminal sanctions, our First Amendment scrutiny is even more exacting. Hicks, 539 U.S. at 119.

Knowledge is too readily proven during prosecution to sufficiently winnow this broad statute. To show knowledge, the prosecution generally must present evidence that the speaker is “aware” that a recipient intends to use the information in furtherance of a federal crime of violence. Barnes v. United States, 412 U.S. 837, 845 (1973). But adjudicators may “impute the element of knowledge” when “the evidence supports an inference of deliberate ignorance.” United States v. Ruhe, 191 F.3d 376, 384 (1999). Criminal defendants can satisfy a knowledge requirement for failure to investigate

37

suspicious circumstances, as with a pharmacist purchasing medication at below-market rate, or a recipient of funds from illegal activities who failed to inquire about the source of suspicious deposits. United States v. George, 233 F. App’x 402, 404 (5th Cir. 2007); United States v. Olla, 754 F. App’x 168, 171 (4th Cir. 2018).

This low bar for criminal knowledge, combined with the broad language of the statute, imperils First Amendment protection for much valuable speech. Consider, for instance, the university professor discussed above, who is scheduled to give a lecture on the physics of combustion, or even simply on the topic of potential energy, which surely constitutes a “part” of information about explosives. If the professor had reason to believe a listener would weaponize this information—perhaps a potential attendee sent a letter outlining malicious intentions, or an audience member wore a t-shirt suggesting an affinity for violence—then the professor could conceivably be prosecuted under Section 842(p)(2)(B) for providing restricted information knowing that an audience member intended to use that information for nefarious purposes. The same could be said for a publisher of an instructional manual for safe use of explosives in construction and demolition. If the publisher received prior notice of a potential reader’s inclination to weaponize the manual’s information, the publisher would be at risk of prosecution under Section 842(p)(2)(B). In both examples, protected and socially valuable speech is stifled because of the possibility that a rogue audience member would misuse the information provided, even if the speaker did not intend such misuse.

My colleagues in the majority may quibble with whether these examples, or numerous others that fit into a similar pattern, would satisfy the knowledge requirement to

38

permit liability under Section 842(p)(2)(B). While I believe that they would, the more pertinent question is whether they constitute speech that might be chilled under this statute, especially “given the ordinary citizen’s predictable tendency to steer wide of the unlawful zone.” Counterman v. Colorado, 600 U.S. 66, 77–78 (2023) (internal citation omitted). If ordinary citizens become aware that they can be punished for providing information they anticipate will aid a crime, they will surely hesitate to provide information when they believe there is some small chance of criminal activity. As a result, even if the majority is correct about the difficulty of proving knowledge to a jury, an ordinary citizen aware of the risks will refrain from protected speech. See Counterman, 600 U.S. at 75 (“A speaker may be unsure about the side of a line on which his speech falls. Or he may worry that the legal system will err, and count speech that is permissible as instead not. Or he may simply be concerned about the expense of becoming entangled in the legal system. The result is self-censorship of speech that could not be proscribed . . . .” (internal citations omitted)).

At its core, this statute brings the First Amendment into new territory, where speech is prohibited not because of the intent of the speaker, but because of the intent of the listener. When a criminal statute requires the speaker to have a criminal intent, First Amendment concerns are lessened. United States v. Hansen, 599 U.S. 762, 783 (2023). The chilling effect of a criminal statute premised upon the intent of the speaker has minimal chilling effect, because the speaker will always know whether they have the requisite intent. When the speech is criminalized as a result of the listener’s intent, however, the speaker is in a far more precarious position. As a result, the prudent speaker who is unsure about the motivations of an audience will refrain from speech. An “important tool to

39

prevent that outcome—to stop people from steering wide of the unlawful zone—is to condition liability on the [government]’s showing of a culpable mental state.” Counterman, 600 U.S. at 75. For this statute, the knowledge requirement insufficiently establishes culpability.

I do not doubt that this statute restricts some legitimately proscribable speech; the circumstances of Arthur’s case demonstrate that this statute has at least some constitutional applications. But the reduced mens rea requirement, in combination with the broad language of the statute, creates an unacceptably high risk of chilling protected expression. Conditioning liability on a more culpable mental state—as Section 842(p)(2)(A) does— “will shield some otherwise proscribable . . . speech . . . . But the added element reduces the prospect of chilling fully protected expression.” Counterman, 600 U.S. 66, 75 (2023).

The government assures us that cases like Arthur’s are emblematic of the types of Section 842(p)(2)(B) cases they intend to prosecute, and speech outside this narrow band of conduct bears no additional risk. But “we would not uphold an unconstitutional statute merely because the Government promised to use it responsibly.” United States v. Stevens, 559 U.S. 460, 480 (2010). “[T]he First Amendment protects against the Government; it does not leave us at the mercy of noblesse oblige.” Id.

III.

The statute restricts information based on its content, and as such is “presumptively invalid” unless the government meets its burden to rebut that presumption. United States v. Playboy Entertainment Grp., Inc., 529 U.S. 803, 817 (2000). The majority concludes

40

that the government has met its burden by demonstrating that the entire law only proscribes speech “integral to criminal conduct,” a band of speech which receives no First Amendment protection. Maj. Op. at 14.

However, the majority misconstrues this “historic and traditional categor[y]” of unprotected speech. Stevens, 559 U.S. at 468. As we have described it, speech which is “tantamount to legitimately proscribable nonexpressive conduct may itself be legitimately proscribed, punished, or regulated incidentally to the constitutional enforcement of generally applicable statutes.” Rice, 128 F.3d at 243. For instance, the crimes of extortion, perjury, and solicitation all contain speech components, but an extortionist does not earn the protection of the First Amendment merely because the extortion is effectuated by speaking. See id. at 244. In other words, “speech is not protected by the First Amendment when it is the very vehicle of the crime itself.” United States v. Varani, 435 F.2d 758, 762 (6th Cir. 1970).

To criminalize speech as “speech integral to criminal conduct,” there must be an underlying criminal offense that the statute penalizes. Naturally, whenever a statute creates criminal penalties for speech, a train of circular logic could allow us to conclude that the speech is integral to criminal conduct: i.e., the criminal conduct of speaking in violation of the challenged statute. But this logic would dispose entirely of the protections of the First Amendment. The category of “speech integral to criminal conduct” refers to that speech that is itself merely a means of committing the underlying crime; in those instances, “the justifications for free speech . . . do not reach communications that are simply means

41

to get a crime successfully committed.” Rice, 128 F.3d at 243 (quoting Kent Greenawalt, Speech, Crime, and the Uses of Language 85 (1989)).

As the majority recognizes, the underlying crime to which this speech is purportedly “integral” is the “Federal crime of violence.” Majority Op. at 18–19. While the restricted speech may facilitate the Federal crime of violence, it is stretching the category beyond its historical limits to claim that the speech is “tantamount to” a federal crime of violence, or “simply a means” of committing a federal crime of violence. Rice, 12 F.3d at 243. The principle that the First Amendment permits “restrictions upon the content of speech in a few limited areas” is not an invitation for courts to expand the scope of these “First Amendment Free Zone[s]” to filter out speech we deem undesirable. R.A.V. v. St. Paul, 505 U.S. 377, 382 (1992) ; Bd. of Airport Comm’rs of L.A. v. Jews for Jesus, Inc., 482 U.S. 569, 574 (1987).

Two further differences highlight the gulf between Section 842(p)(2)(B) and every other example we have recognized of speech integral to criminal conduct. The first and clearest is that all other speech integral to criminal conduct is penalized only when the speaker has the specific intent to commit the crime. Criminal solicitation, conspiracy, extortion, and perjury each require the speaker intend to carry out the underlying criminal act. See Model Penal Code §§ 5.02 (solicitation); 5.03 (conspiracy); 223.4 (extortion); 241.1 (perjury). The Supreme Court has defended the “speech integral to unlawful conduct” category by noting that “[s]peech intended to bring about a particular unlawful act has no social value; therefore, it is unprotected.” Hansen, 599 U.S. at 783 (emphasis added). As I discussed above, a speaker prosecuted under Section 842(p)(2)(B) need not

42

have any intent to commit a crime, much less an intent to carry out the underlying “Federal crime of violence” with which the statute is purportedly concerned.

This leads to the second important distinction between Section 842(p)(2)(B) and the category of speech integral to criminal conduct: unlike any other speech that falls into that category, Section 842(p)(2)(B) restricts sharing publicly available and socially valuable information. As the U.S. Department of Justice acknowledged in its 1997 report on bomb manufacturing, “anyone interested in manufacturing a bomb, dangerous weapon, or a weapon of mass destruction can easily obtain detailed instructions from readily accessible sources such as legitimate reference books, the so-called underground press, and the Internet.” U.S. Dep’t of Justice, Report on the Availability of Bombmaking Information, the Extent to Which its Dissemination Is Controlled by Federal Law, and the Extent to Which Such Dissemination May Be Subject to Regulation Consistent with the First Amendment to the United States Constitution 1-2 (Apr. 29, 1997). The report further recognized that any statute attempting to quell the dissemination of this information would need to leave untouched “the wholly legitimate publication and teaching of such information, or otherwise violate the First Amendment.” Id. at 2.

The First Amendment provides strong protection for the provision of publicly available facts, such as those restricted by Section 842(p)(2)(B). The Supreme Court has long hesitated to punish speakers whose only misdeed was providing truthful information through a new avenue. See The Florida Star v. B.J.F., 491 U.S. 524, 533 (1989); see also Greenawalt, Speech, Crime & the Uses of Language at 86–87 (“A speaker may conceivably think a communication has significant value for the listener beyond the listener’s

43

immediate purpose, but, even if the speaker does not think that, perhaps a recognition of the speaker’s autonomy should include allowing him ordinarily to say what he believes to be true to his acquaintances, regardless of the use he thinks they plan to make of it.”). When the state criminalizes sharing information that exists in the public domain, it places a burden on the speaker to refrain from speech but does not prevent the listener from accessing that information through other, legal channels. Absent an intent to facilitate the commission of a crime, “it is a limited set of cases indeed where, despite the accessibility of the public to certain information, a meaningful public interest is served by restricting its further release by other entities.” Florida Star, 491 U.S. at 535.

The categories of speech excepted from First Amendment protection are historically “limited,” and the First Amendment does not permit government “to disregard these traditional limitations.” Stevens, 559 U.S. at 468. Sharing factual knowledge with no intent to beget a crime does not fit within the category of “speech integral to criminal conduct,” which is primarily concerned with speech that takes the place of a criminal act. Because Section 842(p)(2)(B) does not include a specific intent requirement, and it penalizes the distribution of publicly available information, I cannot agree that the speech it restricts is “speech integral to criminal conduct” that receives no First Amendment protection.

IV.

As Justice Hugo Black wrote, “My view is, without deviation, without exception, without any ifs, buts, or whereases, that freedom of speech means that you shall not do something to people either for the views they have, or the views they express, or the words

44

they speak or write.” One Man’s Stand for Freedom: Mr. Justice Black and the Bill of Rights 478 (Irving Dillard ed. 1963). While there are no doubt benefits to restricting certain categories of speech, “[t]he First Amendment itself reflects a judgment by the American people that the benefits of its restrictions on the Government outweigh the costs. Our Constitution forecloses any attempt to revise that judgment simply on the basis that some speech is not worth it.” Stevens, 559 U.S. at 470. I believe the shocking breadth of the statutory language, combined with the ease of proving knowledge and the substantial danger of chilling protected speech, justify imposing the “strong medicine” of facial invalidation for overbreadth. Hansen, 599 U.S. at 784. I would reverse, and therefore I dissent.