We must decide whether the First Amendment permits the government to punish a threat without proving that it was made with the intent to threaten the victim.

* In early 1998, Paul Kent Cassel and his girlfriend, Anastasia Kafteranis, were living on property owned by Kafteranis and located near Randsburg, California, in the Mojave Desert. The federal government owned several lots near Kafteranis's property and it sought to sell them, acting through the Bureau of Land Management (BLM). Cassel apparently liked his privacy, though, and was not about to let neighbors move in without doing what he could to stop them.

Arthur and Alice Rinard, a married couple, were interested in buying one of the government lots, and in January 1998 they visited the property to look around. As they were walking around, Cassel approached them. He was accompanied by two of his dogs. One of the dogs—a certain "Mr. Mooch Face"—was extremely ugly and at least somewhat aggressive, probably because it had once been run over by a car. Cassel began a conversation with the Rinards that would continue over the following two days. Cassel's participation in the conversation consisted mostly of providing the Rinards with a series of dramatic reasons why the property that the Rinards were considering was quite undesirable. Cassel claimed, among other things, that the government's maps misidentified the boundaries between the various lots; that bidding on one of the lots—lot 107—was pointless because Cassel and Kafteranis were going to purchase it no matter what the cost; that it would cost at least twenty thousand dollars to get the permits needed to build a residence on the property; that the surrounding area was inhabited by child molesters, murderers, producers of illegal drugs, devil-worshipers, and witches; that the ground was a toxic waste dump contaminated with cyanide; that local law enforcement officials were corrupt; that mining explosions had damaged Kafteranis's own house; and that a neighbor had developed a disease known as "silica lung."

Cassel invited the Rinards to join him and Kafteranis for dinner, and despite his generally unneighborly demeanor, they agreed. Cassel kept up his invective during the meal, and his dogs continued to appear aggressive. He ultimately succeeded in dissuading the Rinards from purchasing the lot they had been considering—not, according to Mr. Rinard's testimony, because they believed his stories about nearby witches, but because they did not want Cassel as a neighbor. The Rinards told the BLM about their conversations with Cassel, and Mr. Rinard informed the BLM that while he did not feel personally intimidated by Cassel, he believed that others might.

About a month later, another couple, Mickey and Terry Goodin, came to visit two of the lots. Again Cassel greeted the couple, this time accompanied by only one of his dogs. Cassel's conversation with the Goodins proceeded along the same lines as his earlier conversation with the Rinards. He referred to the BLM as "crooks" and threw in at least one less polite term to boot; he described the toxic waste dump, the child molesters, the drug labs, the devil worshipers, and the cyanide in the ground. Mr. Goodin found Cassel obnoxious and could tell he did not want neighbors. He testified that Cassel told him "that if I [Goodin] tried to build anything on Lot 107, that it would definitely burn. He would see to that. That if I left anything there, it would be stolen, vandalized. He would see to that." Cassel denied making that statement. The Goodins ended up buying a lot about a quarter of a mile from Kafteranis's property, but Mr. Goodin testified that he did not bid on lot 107 because of Cassel's threat to burn any house the Goodins might build. Kafteranis eventually purchased lot 107 at auction.

In November 2000, Cassel was charged in the Eastern District of California with two counts of interfering with a federal land sale under 18 U.S.C. § 1860 and two counts of witness tampering under 18 U.S.C. § 1512(c). In April 2001, the government filed a superseding information dropping one of the witness tampering counts. Cassel was tried before a magistrate judge by his consent, and a jury convicted Cassel on all remaining charges.

In September 2001, before his sentencing hearing, Cassel wrote a letter to the magistrate judge requesting new counsel. The court denied that request after a hearing and sentenced Cassel to five months' imprisonment and 150 days of home confinement. Cassel appealed to the district court, which affirmed his conviction and sentence. This appeal timely followed.

Cassel argues that the statute under which he was convicted, 18 U.S.C. § 1860, is facially unconstitutional, both because it punishes constitutionally protected speech1 and because it is unconstitutionally vague. He further contends that the trial court erroneously instructed the jury and improperly denied his request for new counsel. We consider these claims in turn.2

* 18 U.S.C. § 1860 punishes, in relevant part, "[w]hoever, by intimidation ... hinders, prevents, or attempts to hinder or prevent, any person from bidding upon or purchasing any tract of" federal land at public sale. Cassel argues that this language criminalizes a wide range of constitutionally protected speech merely because such speech has the effect—intended or not—of "intimidati[ng]" a potential land buyer and thus making it more difficult for the government to sell its land.

* As an initial matter, we must consider what level of scrutiny we are to apply to the statute. Its language is not addressed specifically to speech: rather, it criminalizes the act of interfering with a federal land sale "by intimidation," whether through speech or conduct. The government appears to contend, on that basis, that the statute should not be analyzed as a restriction on speech at all, but rather as a perfectly legitimate restriction on conduct that only incidentally restricts some speech. Cf. United States v. Brice, 926 F.2d 925, 931 (9th Cir. 1991) ("If conduct contains both speech and non-speech elements, and if Congress has the authority to regulate the non-speech conduct, incidental restrictions on freedom of speech are not constitutionally invalid."). But when the definition of a crime or tort embraces any conduct that causes or might cause a certain harm, and the law is applied to speech whose communicative impact causes the relevant harm, we treat the law as content-based. See, e.g., Hustler Magazine v. Falwell, 485 U.S. 46, 108 S. Ct. 876, 99 L. Ed. 2d 41 (1988) (holding that "public figures and public officials" must show actual malice to prevail on a state-law tort claim for intentional infliction of emotional distress based on the defendant's speech). 18 U.S.C. § 1860 is just such a law: its prohibition applies to any conduct that causes the harm of interfering by intimidation with a public land sale. Since it was applied in this case to Cassel's speech (as well as his allegedly intimidating conduct, such as approaching the couples while accompanied by aggressive dogs), we treat it as content-based.

In arguing to the contrary, the government directs our attention to the Fifth Circuit's decision in United States v. Hicks, 980 F.2d 963 (5th Cir. 1992). In Hicks, the Fifth Circuit upheld a conviction under 49 U.S.C.App. § 1472(j), which prohibits "intimidat[ing] or threaten[ing]" a member of the crew onboard an airplane. The court held that "the statute reasonably regulates the time, place, and manner of speech, irrespective of its particular content. The content of passengers' speech is thus regulated only in an incidental fashion." Hicks, 980 F.2d at 971. Accordingly, it analyzed the statute under the deferential standard applied to content-neutral restrictions on speech.3 See id.

We cannot agree with this analysis, however. "Time, place, and manner" restrictions regulate matters such as how loud speech can be, see, e.g., Ward v. Rock Against Racism, 491 U.S. 781, 109 S. Ct. 2746, 105 L. Ed. 2d 661 (1989), or whether residents may place signs in their lawns, see, e.g., City of Ladue v. Gilleo, 512 U.S. 43, 114 S. Ct. 2038, 129 L. Ed. 2d 36 (1994). These are issues that have nothing to do with the content of the speech. Excessively loud speech is objectionable whether it consists of rock music or nursery rhymes, and signs on lawns may be unsightly whether they advertise political candidates or a garage sale. In contrast, 18 U.S.C. § 1860 punishes speech precisely because of the "intimidat[ing]" message it contains. It must, therefore, be analyzed as a content-based restriction on speech.

Not all content-based restrictions on speech are unconstitutional, of course. The Supreme Court has recognized that certain categories of speech are of such low value and inflict such serious harm that they are outside the protection of the First Amendment. One such category is composed of "true threats." See, e.g., United States v. Hanna, 293 F.3d 1080, 1084 (9th Cir. 2002) ("[T]he Court [has] left no doubt that true threats could be criminalized because they are not protected speech."). The government argues that § 1860 comports with the First Amendment because it only punishes speech that falls within this unprotected category. Cassel recognizes that true threats are unprotected, but he contends that only intentional threats are unprotected by the First Amendment and that the crime defined in § 1860 lacks the constitutionally necessary mens rea element of intent. We are thus faced with the question whether intent to threaten the victim is required in order for speech to fall within the First Amendment exception for threats.

As a preliminary matter, we should be clear about precisely what is at issue. On one hand, it is not in dispute that, to be convicted under § 1860, the defendant must intend the speech or conduct that has the effect of intimidating the potential land buyer. In other words—though it is almost too obvious to mention—a defendant whose involuntary epileptic fit caused him to appear menacing during a land auction, frightening off a potential buyer, would doubtless not have violated the statute. (We intimate no view as to whether this minimal element of intent is required by the Constitution; the issue is not before us.)

On the other hand, no one now contends that the Constitution requires proof that the defendant intends to carry out the threat—that is, actually to inflict the harm that he has threatened. One of the chief evils wrought by a threat is its deleterious and coercive effect on the victim (and, here, in the consequent effect on the government's ability to sell its land), and that effect is not diminished merely because the defendant is bluffing. See Roy v. United States, 416 F.2d 874, 877 (9th Cir. 1969); Virginia v. Black, 538 U.S. 343, 359-60, 123 S. Ct. 1536, 155 L. Ed. 2d 535 (2003).

Rather, the disputed question is whether the government must prove that the defendant intended his words or conduct to be understood by the victim as a threat. Cassel argues that it must. The government's position is that mere negligence with regard to the victim's understanding is enough: in other words, speech is punishable if a reasonable person would understand it as a threat, whether or not the speaker meant for it to be so understood.

The Supreme Court has decided very few cases directly addressing the threat exception. For many years, its only significant pronouncement on the subject was its opinion in Watts v. United States, 394 U.S. 705, 89 S. Ct. 1399, 22 L. Ed. 2d 664 (1969). The defendant in Watts had announced at a public rally against the Vietnam War in 1966 that "[i]f they ever make me carry a rifle the first man I want to get in my sights is L.B.J." Id. at 706, 89 S. Ct. 1399. He was convicted of violating 18 U.S.C. § 871(a), which prohibits "knowingly and willfully mak[ing] any threat to take the life of or to inflict bodily harm upon the President of the United States." Id. at 705 & n. *, 89 S. Ct. 1399. The Court overturned his conviction, holding that First Amendment principles required it to construe the term "threat" in the statute to exclude the sort of political hyperbole the defendant had used. Id. at 707-08, 89 S. Ct. 1399. The Court declined to address the question of what mens rea requirement the statute imposed,4 see id., and so it did not explain whether intent to threaten is a necessary part of a constitutionally punishable threat.

Our own cases, it must be said, have not been entirely clear or consistent on that question. We have often held or implied that negligence is enough. In Roy v. United States, 416 F.2d 874 (9th Cir. 1969), we were called upon to answer the specific statutory question left unresolved by the Supreme Court in Watts: what degree of intent is required by 18 U.S.C. § 871(a)'s prohibition of "knowingly and willfully" communicating a threat against the President. Roy did not raise constitutional issues in his appeal, and we did not address them except to say, in a footnote, that "there does not appear to be a free speech issue in this case." Roy, 416 F.2d at 879 n. 17. Nevertheless, as a matter of statutory construction, we held that the statute required

only that the defendant intentionally make a statement ... under such circumstances wherein a reasonable person would foresee that the statement would be interpreted by those to whom the maker communicates the statement as a serious expression of an intention to inflict bodily harm upon ... the President, and that the statement not be the result of mistake, duress, or coercion.

Id. at 877-78. The reference to a "reasonable person" makes clear that we were describing a negligence standard, with no requirement of actual intent to threaten. We subsequently reaffirmed Roy's holding in United States v. Hanna, 293 F.3d 1080 (9th Cir. 2002), and stated clearly that "Roy's `reasonable speaker' standard does not violate the First Amendment." Id. at 1084.

In United States v. Orozco-Santillan, 903 F.2d 1262 (9th Cir. 1990), we upheld the defendant's conviction for threatening a federal law enforcement officer under 18 U.S.C. § 115(a)(1)(B). We were not primarily concerned with constitutional questions, but we observed that "[a] `true' threat, where a reasonable person would foresee that the listener will believe he will be subjected to physical violence upon his person, is unprotected by the first amendment." Orozco-Santillan, 903 F.2d at 1266. Again, our reference to a "reasonable person" seems to suggest that the First Amendment permits punishing a threat made with only negligence as to the statement's threatening character.5 See also Wurtz v. Risley, 719 F.2d 1438, 1441 (9th Cir. 1983) (stating that "implicit in the nature of ... punishable threats is a reasonable tendency to produce in the victim a fear that the threat will be carried out" and holding that a Montana statute was overbroad because it contained no such requirement); cf. United States v. Meeker, 527 F.2d 12 (9th Cir. 1975) (holding, without mentioning constitutional issues, that 49 U.S.C.App. § 1472(j), which prohibits "intimidat[ing] or threaten[ing]" a crew member onboard an airplane "so as to interfere with the performance" of the crew member's duties, does not require specific intent to interfere with the crew member's performance).

Yet at other times the language of our cases has suggested that intent is required. In United States v. Gilbert, 813 F.2d 1523 (9th Cir. 1987), we upheld, against an overbreadth challenge, a provision of the Fair Housing Act that criminalizes "willfully ... intimidat[ing]" anyone for various reasons related to racial discrimination in housing. See 42 U.S.C. § 3631. A conviction under the statute requires proof that the defendant specifically intended "to injure, intimidate or interfere with the victim." See, e.g., United States v. McInnis, 976 F.2d 1226, 1230 (9th Cir. 1992). Addressing Gilbert's First Amendment challenge to § 3631, we observed that "the element of intent specified in section 3631 [is] the determinative factor separating protected expression from unprotected criminal behavior." Gilbert, 813 F.2d at 1529; see also id. ("[T]he statute's requirement of intent to intimidate serves to insulate the statute from unconstitutional application to protected speech."). That language seems to imply that a subjective "intent to intimidate" (or intent to threaten) is necessary before a threat can constitutionally be punished. Cf. United States v. Twine, 853 F.2d 676 (9th Cir. 1988) (holding, without discussing constitutional issues, that subjective intent to threaten is necessary to a conviction under 18 U.S.C. §§ 875(c) and 876, which punish threats to kidnap or injure another person).

On one occasion, we seemed to advocate both positions within the confines of a single opinion. In Planned Parenthood of the Columbia/Willamette, Inc. v. American Coalition of Life Activists, 290 F.3d 1058 (9th Cir. 2002) (en banc), we rejected a First Amendment challenge to the Freedom of Access to Clinics Entrances Act (FACE), 18 U.S.C. § 248, which provides a right of action against whoever by "threat of force ... intentionally intimidates any person because that person is or has been... providing reproductive health services." 18 U.S.C. §§ 248(a)(1), (c)(1)(A). In the course of holding that FACE does not require intent or ability to carry out the threat, we repeated our statement in Orozco-Santillan that "[a] true threat, that is one where a reasonable person would foresee that the listener will believe he will be subjected to physical violence upon his person, is unprotected by the first amendment." Planned Parenthood, 290 F.3d at 1075 (quoting Orozco-Santillan, 903 F.2d at 1265). We also wrote that "the only intent requirement for a true threat is that the defendant intentionally or knowingly communicate the threat."6 Id. Thus far, someone reading our decision in that case would think our holding was clear: no subjective intent to threaten is necessary. Indeed, one of the dissents described the majority's holding in just that way. Id. at 1108 n. 9 (Berzon, J., dissenting) ("The majority opinion does not appear to embrace any such subjective intent standard as a constitutional requirement.. . .").

Yet our opinion in Planned Parenthood also noted that the statute in question contained an explicit requirement of intent to threaten (or "intent to intimidate"), and we cited Gilbert for the proposition that "the requirement of intent to intimidate serves to insulate the statute from unconstitutional application to protected speech." Id. at 1076 (quoting Gilbert, 813 F.2d at 1529). Moreover, we observed that "the requirement of intent to intimidate cures whatever risk there might be of overbreadth." Id. at 1079. And we wrote that the reasonable-person definition of a true threat, "coupled with the statute's irequirement of intent to intimidate, comports with the First Amendment." Id. at 1088 (emphasis added). In these passages we seemed to hold—or at least keep open the possibility—that subjective intent to threaten is necessary. But see Hanna, 293 F.3d at 1084 (citing Planned Parenthood for the proposition that subjective intent is unnecessary).

We need not delve deeper into the vagaries of our own case law. For in an opinion more recent than any of our relevant cases, the Supreme Court revisited the topic of threats and the First Amendment. In Virginia v. Black, 538 U.S. 343, 123 S. Ct. 1536, 155 L. Ed. 2d 535 (2003), the Court held that it does not violate the Constitution for a state to prohibit burning a cross with the intent to intimidate. The state of Virginia had enacted a statute providing as follows:

It shall be unlawful for any person or persons, with the intent of intimidating any person or group of persons, to burn, or cause to be burned, a cross on the property of another, a highway or other public place. ... Any such burning of a cross shall be prima facie evidence of an intent to intimidate a person or group of persons.

Va.Code Ann. § 18.2-423 (1996). The Supreme Court of Virginia consolidated the appeals of three people convicted under this statute and held the statute unconstitutional on its face on the grounds that, although it punishes speech within the generally unprotected category of true threats, it "selectively chooses only cross burning because of its distinctive message." Black v. Virginia, 262 Va. 764, 553 S.E.2d 738, 744 (2001). The U.S. Supreme Court disagreed, holding that a state may single out "cross burnings done with the intent to intimidate" for punishment because "burning a cross is a particularly virulent form" of threat. Black, 538 U.S. at 363, 123 S. Ct. 1536. "Thus, just as a State may regulate only that obscenity which is the most obscene due to its prurient content, so too may a State choose to prohibit only those forms of intimidation that are most likely to inspire fear of bodily harm." Id.

The Court laid great weight on the intent requirement. It offered this definition of unprotected "true threats" and "intimidation":

"True threats" encompass those statements where the speaker means to communicate a serious expression of an intent to commit an act of unlawful violence to a particular individual or group of individuals. The speaker need not actually intend to carry out the threat.... Intimidation in the constitutionally proscribable sense of the word is a type of true threat, where a speaker directs a threat to a person or group of persons with the intent of placing the victim in fear of bodily harm or death.

Id. at 359-60, 123 S. Ct. 1536 (citations omitted). The clear import of this definition is that only intentional threats are criminally punishable consistently with the First Amendment. First, the definition requires that "the speaker means to communicate... an intent to commit an act of unlawful violence." A natural reading of this language embraces not only the requirement that the communication itself be intentional, but also the requirement that the speaker intend for his language to threaten the victim.

The Court's insistence on intent to threaten as the sine qua non of a constitutionally punishable threat is especially clear from its ultimate holding that the Virginia statute was unconstitutional precisely because the element of intent was effectively eliminated by the statute's provision rendering any burning of a cross on the property of another "prima facie evidence of an intent to intimidate." At the trial of one of the petitioners, the jury had been instructed that "[t]he burning of a cross, by itself, is sufficient evidence from which you may infer the required intent." Id. at 364, 123 S. Ct. 1536. This language mirrored Virginia's Model Jury Instruction for the statute, see Va. Model Jury Instructions, Criminal, Instruction No. 10.250 (1998 & Supp. 2001), and the Supreme Court of Virginia had not disavowed the instruction; accordingly, Justice O'Connor's plurality opinion held the Court bound by the instruction's interpretation of the provision. Id. at 364-65, 123 S. Ct. 1536.

Because the prima facie evidence provision makes it unnecessary for the government actually to prove the defendant's intent, the plurality held that the statute violated the First Amendment. See id. at 365, 123 S. Ct. 1536 ("[T]he prima facie provision strips away the very reason why a State may ban cross burning with the intent to intimidate."). Its holding thus affirms our own dictum—not always adhered to in our cases—that "the element of intent [is] the determinative factor separating protected expression from unprotected criminal behavior." United States v. Gilbert, 813 F.2d 1523, 1529 (9th Cir. 1987). The Court emphasized that the very "same act" "may mean that a person is engaging in constitutionally proscribable intimidation [or] only that the person is engaged in core political speech." Black, 538 U.S. at 365, 123 S. Ct. 1536. "[A] burning cross is not always intended to intimidate," the Court noted, and when it is not so intended it may not be punished as a threat. Id.; see also id. at 366, 123 S. Ct. 1536 (noting that the prima facie provision "does not distinguish between a cross burning done with the purpose of creating anger or resentment and a cross burning done with the purpose of threatening or intimidating a victim" (emphasis added)); id. at 367, 123 S. Ct. 1536 ("The ... provision... ignores all of the contextual factors that are necessary to decide whether a particular cross burning is intended to intimidate. The First Amendment does not permit such a shortcut.").

Although Justice O'Connor's opinion was only for a four-Justice plurality of the Court, each of the other opinions—with the possible exception of Justice Thomas's dissent—takes the same view of the necessity of an intent element. Justice Scalia agreed that the Virginia statute was unconstitutional insofar as it failed to require the state to prove the defendant's intent. Id. at 368, 123 S. Ct. 1536 (Scalia, J., concurring in part, concurring in the judgment in part, and dissenting in part). He disagreed only with the Court's facial invalidation of the statute; instead, he would have permitted case-by-case challenges to convictions in which the prima facie provision operated to remove the state's burden to prove intent. Id. at 372-73, 123 S. Ct. 1536.

Justice Souter, joined by Justices Kennedy and Ginsburg, agreed that the prima facie evidence provision rendered the statute facially unconstitutional because it effectively eliminated the intent requirement. Id. at 385, 123 S. Ct. 1536 (Souter, J., concurring in the judgment in part and dissenting in part) (noting that "the symbolic act of burning a cross ... is consistent with both intent to intimidate and intent to make an ideological statement"). He disagreed only with the plurality's holding that the Virginia Supreme Court could, on remand, apply a narrowing construction to the prima facie evidence provision and thus save the statute as a whole from facial unconstitutionality. Id. at 387, 123 S. Ct. 1536.

Justice Thomas dissented outright; he would have held that the statute reaches only conduct, not expression, and thus raises no First Amendment concerns at all. Id. at 394-95, 123 S. Ct. 1536 (Thomas, J., dissenting). Alternatively, even under First Amendment analysis, he would have held that "the fact that the statute permits a jury to draw an inference of intent to intimidate from the cross burning itself presents no constitutional problems." Id. at 395, 123 S. Ct. 1536.

Thus, eight Justices agreed that intent to intimidate is necessary and that the government must prove it in order to secure a conviction.7 We cannot but conclude that the same principle governs in the case before us.8 See Frederick Schauer, Intentions, Conventions, and the First Amendment, 55 Sup.Ct. Rev. 197, 217 (2003) ("[I]t is plain that ... the Black majority ... believed that the First Amendment imposed upon Virginia a requirement that the threatener have specifically intended to intimidate."); Roger C. Hartley, Cross Burning—Hate Speech as Free Speech, 54 Cath. U.L.Rev. 1, 33 ("Black now confirms that proof of specific intent (aim) must be proved also in threat cases."); Lauren Gilbert, Mocking George: Political Satire as "True Threat" in the Age of Global Terrorism, 58 U. Miami L.Rev. 843, 883-84 (2004); cf. Jennifer E. Rothman, Freedom of Speech and True Threats, 25 J.L. & Pub. Pol'y 283, 317-18 (arguing, before Black, for a subjective intent requirement, and observing that "First Amendment law often requires proof of a specific state of mind before finding a speaker liable or allowing a criminal conviction of the speaker").

The Court's definition of a constitutionally proscribable threat is, of course, binding on us even though it is in tension with some of the holdings and language in prior cases of this circuit.9 See Miller v. Gammie, 335 F.3d 889, 899 (9th Cir. 2003) (en banc) (holding that when a three-judge panel is faced with intervening precedent from a higher court that is "clearly irreconcilable" with a prior holding of this court, the panel is bound by the intervening authority). We are therefore bound to conclude that speech may be deemed unprotected by the First Amendment as a "true threat" only upon proof that the speaker subjectively intended the speech as a threat.10

We turn now to Cassel's claim that 18 U.S.C. § 1860 is facially unconstitutional because it fails to require the requisite subjective intent. His argument relies on a literal reading of the statute:11 taken literally, the phrase "[w]hoever, by intimidation... hinders [or] prevents [a bid on public land]" would apply to anyone whose conduct happened to intimidate a potential bidder, whether or not the intimidation was intentional.

Yet this literal reading ignores a cardinal principle of statutory construction: "The existence of a mens rea is the rule of, rather than the exception to, the principles of Anglo-American criminal jurisprudence." United States v. Pasillas-Gaytan, 192 F.3d 864, 868 (9th Cir. 1999) (quoting Dennis v. United States, 341 U.S. 494, 500, 71 S. Ct. 857, 95 L. Ed. 1137 (1951)). Thus, except in unusual circumstances, we construe a criminal statute to include a mens rea element even when none appears on the face of the statute.12 See, e.g., id. (construing 18 U.S.C. § 1425, which criminalizes the unlawful procurement of naturalization, to include a requirement that the defendant either know he is not eligible for naturalization or knowingly lie in his application).

Cassel's argument is all the more untenable in light of the principle that "every reasonable construction must be resorted to, in order to save a statute from unconstitutionality." United States v. Buckland, 289 F.3d 558, 564 (9th Cir. 2002) (quoting Hooper v. California, 155 U.S. 648, 657, 15 S. Ct. 207, 39 L. Ed. 297 (1895)); INS v. St. Cyr, 533 U.S. 289, 299-300 & n. 12, 121 S. Ct. 2271, 150 L. Ed. 2d 347 (2001); see also Miller v. French, 530 U.S. 327, 336, 120 S. Ct. 2246, 147 L. Ed. 2d 326 (2000) (instructing courts to avoid "constitutionally doubtful constructions"). Having held that intent to threaten is a constitutionally necessary element of a statute punishing threats, we do not hesitate to construe 18 U.S.C. § 1860 to require such intent. Cassel's argument that the statute is facially unconstitutional therefore fails.

Cassel also argues that the statute is unconstitutional because it is unacceptably vague. This contention is inconsistent with our decision in United States v. Tabacca, 924 F.2d 906 (9th Cir. 1991). In Tabacca we rejected a vagueness challenge to 49 U.S.C.App. § 1472(j), which criminalizes "intimidat[ing], or threaten[ing]" a member of the crew onboard an airplane. We held that the term "intimidate" was not unconstitutionally vague because the statute

clearly includes modifying language which provides parameters of conduct. The proscribed acts must occur while aboard an aircraft and must "interfere with the performance by ... [an] attendant of his duties...."

Id. at 913. Those limitations on the statute's scope, we held, gave the defendant reasonable notice of the sort of "intimidation" that was proscribed. Similar limitations exist in this case: the "intimidation" punishable under 18 U.S.C. § 1860 must occur in connection with a public sale of land, and it must hinder or prevent someone from bidding in such a sale. That suffices to give "fair notice to those who might violate" the statute, and so Cassel's vagueness challenge fails. See Gilbert, 813 F.2d at 1530.

We now consider Cassel's argument that the jury instructions given by the trial court were erroneous and require reversal of his conviction. The trial judge instructed the jury as follows:

Intimidation is to make a person timid or fearful through the use of words and conduct that would put an ordinary, reasonable person in fear or apprehension for the purpose of compelling or deterring legal conduct of that person.

Cassel argues that the instruction failed to include a sufficient mens rea element. He also argues that the trial court should have instructed the jury that intimidation must involve the belief of the victim that there is reasonable fear of bodily harm or violence.

In view of our holding that 18 U.S.C. § 1860 includes a mens rea element of subjective intent, Cassel's first contention is correct, for the instruction included no such requirement. It did require that the intimidation be "for the purpose of compelling or deterring legal conduct," but that is by no means the same thing.13 It is not enough that the defendant intend to influence the potential bidder; rather, he must intend to do so by means of a threat. Were it otherwise, a defendant could be convicted under § 1860 who, for example, merely attempted to dissuade the potential bidder by informing him of dangerous conditions in the area surrounding the land to be sold. Such a result comports neither with the First Amendment nor with the statute's purpose, which is to punish interference by intimidation, not to prevent potential land buyers from hearing negative viewpoints on the land for sale.

A jury instruction that omits a necessary element of the offense is fatally flawed and requires reversal unless the error is harmless beyond a reasonable doubt. See United States v. Pasillas-Gaytan, 192 F.3d 864 (9th Cir. 1999); Neder v. United States, 527 U.S. 1, 8-15, 119 S. Ct. 1827, 144 L. Ed. 2d 35 (1999). The record in this case does not suffice to establish Cassel's intent with the degree of certainty necessary to take a contested factual issue out of the hands of the jury. There was conflicting testimony at trial as to precisely what Cassel said and did to the Rinards and the Goodins, and the fact that the jury convicted Cassel based on the instruction quoted above does not demonstrate beyond a reasonable doubt that it would have convicted him under an instruction that accords with our decision today.

Cassel also contends that "intimidation" under § 1860 requires a fear of bodily harm on the victim's part. Several aspects of this contention require consideration. First, insofar as Cassel is suggesting that the victim must actually be intimidated, we cannot agree. The proper focus is on the defendant's conduct and its effect on the federal land sale, not the subjective emotions which the victim experiences. Cf. United States v. Meeker, 527 F.2d 12 (9th Cir. 1975) ("Nor is proof that the victim was in fact frightened for his own physical safety required in order to find that a defendant performed the criminal act of intimidation."); United States v. Alsop, 479 F.2d 65, 67 n. 4 (9th Cir. 1973) (suggesting a jury instruction that "[t]o take ... `by intimidation'" under 18 U.S.C. § 2113(a), which prohibits bank robbery, "means wilfully to take ... in such a way that would put an ordinary, reasonable person in fear of bodily harm"). As we have noted, "[w]ere it otherwise, a fearless banker could never be robbed by intimidation. We refuse to reach such an absurd result." Id. at 67. It would be equally absurd to hold that a fearless bidder can never be deterred by intimidation.

Second, we do not agree that the scope of the statute is necessarily limited to threats of bodily harm. "Intimidation," both in common and legal usage, can refer to the act of placing someone in fear of injury other than harm to the body. See, e.g., 720 Ill. Comp. Stat. 5/12-6(a)(1) ("A person commits intimidation when, with intent to cause another to perform or to omit the performance of any act, he communicates... a threat to ... [i]nflict physical harm on the person threatened or any other person or on property" (emphasis added)). In some cases, arising under other statutes, we have approved jury instructions defining "intimidation" as involving a threat of bodily harm. See, e.g., United States v. Alsop, 479 F.2d 65, 67 n. 4 (9th Cir. 1973). We are aware of no case, however, squarely presenting the issue whether threatened injury to property falls within the ambit of a statute prohibiting intimidation. The purpose of 18 U.S.C. § 1860 is to prevent interference with federal land sales, and we see no reason why Congress would have intended to exclude interference accomplished through threats to property from the statute's scope. A potential bidder who fears that his house will be burned to the ground may be just as "intimidat[ed]" as one who fears an assault on his person. We conclude that "intimidation" under 18 U.S.C. § 1860 requires the threat of harm inflicted by the defendant upon the victim's person or property. Cf. United States v. Viefhaus, 168 F.3d 392, 394 (10th Cir. 1999) (defining a "true threat" to include "a declaration of intention ... to injure another or his property by the commission of some unlawful act" (emphasis added)).

Nevertheless, Cassel's objection contains a kernel of merit. For while the jury instruction correctly stated that "intimidation" involves "words and conduct that would put an ordinary, reasonable person in fear or apprehension," it failed to specify that the statute requires "fear or apprehension" of injury inflicted by the defendant. Whether the threat is of injury to person or property, there is no doubt that it must be a threat of injury brought about—rather than merely predicted—by the defendant.14 Indeed, the First Amendment requires as much. See Planned Parenthood, 290 F.3d at 1076.

Again, this error in the jury instruction cannot be said to be harmless. Most of the statements Cassel made to potential buyers did not suggest any threat of force on his part. For example, his claim that the area was inhabited by devil worshipers and producers of illegal drugs could fall within the court's instruction that intimidation consists of words and conduct that "would put an ordinary, reasonable person in fear or apprehension." Such claims, however, cannot constitute intimidation under § 1860, because they do not involve a threat of harm to person or property inflicted by the defendant. Accordingly, Cassel is entitled to a new trial on this ground as well.

Finally, we address Cassel's argument that the trial court improperly denied his request for new counsel. We review the denial of a motion for substitution of counsel for abuse of discretion. United States v. Gonzalez, 113 F.3d 1026, 1028 (9th Cir. 1997). The trial court's discretion "must be exercised ... within the limitations of the Sixth Amendment, which grants criminal defendants a qualified constitutional right to ... counsel of their choice." Id. (quoting United States v. D'Amore, 56 F.3d 1202, 1204 (9th Cir. 1995)) (second omission in original). We consider three factors when evaluating the trial court's decision: (1) the adequacy of the court's inquiry into the defendant's complaint, (2) the extent of conflict between the defendant and counsel, and (3) the timeliness of the motion and the extent of resulting inconvenience or delay. Id.

* The first factor we consider is whether the trial court has "conduct[ed] an inquiry adequate to create a sufficient basis for reaching an informed decision." Gonzalez, 113 F.3d at 1028. The trial judge in this case had an extended conversation with Cassel about the distrust Cassel felt for his attorney; Cassel had several opportunities to present his concerns and the court gave every indication that it was willing to take Cassel seriously. When Cassel described his disagreement and frustration with his attorney's decision to make certain objections and pursue certain lines of questioning, the judge carefully explained to Cassel why his attorney had objected and why the objection made perfect sense from a legal perspective. The judge even gave Cassel an example of the sort of behavior on a lawyer's part that would justify removing him from the case and gave Cassel an opportunity to compare his attorney's conduct to the judge's example. Our review of the record satisfies us that the magistrate judge conducted an adequate inquiry.

ii

The second factor concerns the degree to which the conflict between the defendant and his attorney prevented the attorney from providing effective assistance. See Gonzalez, 113 F.3d at 1029. The record supplies no indication that Cassel's attorney provided inadequate assistance. As the district court noted, despite Cassel's statement that his attorney considered him paranoid, "Mr. Cassel never suggested that at any time he and [his attorney] Mr. Rainwater were unable to communicate about his defense." The magistrate judge, in fact, complimented the attorney's performance, noting that he was at all times "a very strong advocate within the legal bounds, and within the professional ethical bounds" on Cassel's behalf and that he was "a very focused lawyer." Cassel did report a great deal of frustration and lack of trust in his attorney, which weighs in favor of his request. Nevertheless, nothing he said indicated either that a conflict of interest existed or that the attorney-client relationship had broken down in a way that would diminish the attorney's ability to represent Cassel. Cf. Jackson v. Ylst, 921 F.2d 882, 888 (9th Cir. 1990).

iii

The third factor is the timeliness of the defendant's request. Cassel made his request for new counsel on the day of sentencing. If granted, it would certainly have resulted in some delay, possibly substantial, as a new attorney became familiar enough with the case to represent Cassel at sentencing.

iv

Two of the three relevant factors—the adequacy of the magistrate judge's inquiry and the timeliness of the defendant's request—weigh heavily in favor of the magistrate judge's decision not to grant Cassel's request for new counsel. The third factor—the degree of conflict between client and attorney—is slightly more equivocal but still supports the judge's decision: Cassel's mistrust of his lawyer was real, but nothing suggests that it threatened to affect the attorney's ability to represent him. We are therefore satisfied that the magistrate judge did not abuse his discretion in denying Cassel's request.

Although Cassel's facial challenge to 18 U.S.C. § 1860 fails, his conviction was based on jury instructions that inadequately described the elements of the crime. Accordingly, the district court's judgment of conviction is VACATED and the case is REMANDED for a new trial.