dissenting:
“This is the kind of case that could persuade the cynical reader who believes he has seen it all that something new and different may still be lurking around the corner.” Homan v. Goyal, 711 A.2d 812, 814 (D.C.1998). At issue is whether the trial judge correctly stated the elements of the crime of threats to do bodily harm. D.C.Code § 22-407 (2001). According to the judge, the government was required to prove beyond a reasonable doubt that the defendant, Lee Carrell, (1) spoke words or otherwise communicated to the complaining witness words (2) that would cause a person to reasonably believe that he or she would be seriously harmed if the event occurred; and (3) that he intended to utter the words that constituted the threat. (Emphasis and numerals added.) Carrell has no quarrel with the judge’s statement of the first and second elements, but he contends that the third was incorrect, and that the government should instead have been required to prove that Carrell “intended to utter these words as a threat.”
The unusual feature of this case is that over a thirty-year period, this court has used the language advocated by Carrell on five occasions and the competing articulation urged on us by the government and adopted by the trial judge in only three. Because the earliest of these eight cases— Campbell v. United States, 450 A.2d 428, 431 n. 5 (D.C.1982) — used the phrase “intended to utter the words that constituted the threat” but did not specify that the defendant must have “intended to utter these words as a threat,” the government contends, and the majority apparently agrees, that the court is required by M.A.P. v. Ryan, 285 A.2d 310, 312 (D.C.1971), to adhere to the Campbell articulation and to reject the alternative language used in five of our seven post-Campbell decisions, beginning with United States v. Baish, 460 A.2d 38, 42 (D.C.1983), decided one year after Campbell. For the reasons set forth in some detail below, I do not agree with this contention, and I conclude, to the contrary, that the issue is an open one under District of Columbia law and that we should now adopt the Baish articulation.
My colleagues also maintain that the Campbell language is correct because threats to do bodily harm should be treat*172ed as a general intent offense and that proof of specific intent should not be required. Although I find this issue difficult in light of some of the federal decisions cited by the majority, I nevertheless conclude that affirmance is not warranted.
I.
I begin by addressing a preliminary issue. The court holds, and I agree, that whether one applies the “intent to utter the words” standard advocated by the government or the “intent to utter these words as a threat” standard favored by the defendant, the evidence, as described in the majority opinion, was sufficient to establish Carrell’s guilt beyond a reasonable doubt. In addressing a claim of evidentia-ry insufficiency, we must of course view the record in the light most favorable to the prosecution. In re S.W., 45 A.3d 151, 154 (D.C.2012). Although the words used by Carrell: “I could kill you. I could kill you. I could fucking kill you right now” do not explicitly describe what Carrell proposed to do in the future, as a threat ordinarily would, I do not doubt that under the circumstances, including Carrell’s simultaneous assault on Ms. Ringenburg (and the unambiguously threatening remarks he made the previous evening to the effect that he would kill her friends or family if she called them to defend her) could reasonably be viewed by an impartial trier of fact as an implicit threat. This would be true even if the government is required to prove, as I believe that it should be, that Carrell intended his words as a threat.
This does not mean, however, that it is of no consequence whether the judge applied the “intent to utter the words” standard or the more exacting “intent to utter the words as a threat” test. It would not be accurate to say that the government necessarily wins either way. On the contrary, the result of the appeal could well turn on which of the two articulations was applied in the court’s analysis. If the government had to prove only that Carrell intended to utter the words, then the judge’s affirmative credibility finding as to Ms. Ringenburg’s testimony effectively disposed of the case. If, on the other hand, the prosecution was required to prove that Carrell intended to utter the words as a threat, it is not at all obvious that a reasonable trier of fact, whether judge or jury, would inevitably convict him. Guilt or innocence is a question for the trier of fact, and at this stage of the proceeding, the judge need not — indeed, she may not — view the record in the light most favorable to the prosecution. Rather, she must presume innocence and acquit unless the government proves the defendant’s guilt beyond a reasonable doubt. “The reasonable doubt standard of proof requires the factfinder to reach a subjective state of near certitude of the guilt of the accused.” Rivas v. United States, 783 A.2d 125, 133 (D.C.2001) (en banc). In this case, the judge having believed that the relevant events had occurred as Ms. Ringenburg described them, there was no plausible way to acquit Carrell of the assault charge against him. Whether, however, “I could fucking kill you right now,” despite its phrasing as to what Carrell would be able to do, was intended to threaten Ms. Ringenburg regarding what he would do, is far less obvious and strikes me as a classic question of fact for the jury (or, in this case for the judge). We therefore cannot decide the appeal, in my judgment, without first determining whether the trial judge’s statement of the elements of the offense was consistent with the law of this jurisdiction.
II.
As noted above, an astonishing feature of this case is that in addressing the ques*173tion whether the prosecution was required to prove “intent to utter as a threat” rather than only “intent to utter the words,” both parties have been able to cite apparently impressive supportive District of Columbia precedent. Over the past three decades, there have been at least eight decisions by our court purporting to recite the elements of the offense. Only three of these decisions, including Campbell, the earliest of the eight, use the government’s articulation in setting forth the intent element of threats to do bodily harm. See also Evans v. United States, 779 A.2d 891, 894 (D.C.2001) (citing Campbell); Joiner-Die v. United States, 899 A.2d 762, 764 (D.C.2006) (citing Evans and Campbell ).1 In none of the three, however, was there any issue raised as to whether the prosecution must prove intent to threaten or merely intent to utter. Campbell dealt primarily with the admission of “other crimes” evidence; Evans concerned the question whether “attempted threats” is a crime; and Joiner-Die addressed the sufficiency of the evidence and the allegedly improper curtailment of cross-examination. Because the issue presented in the present appeal did not arise in any of these cases, it could not have been authoritatively decided.
In five of the eight cases, including the three most recent ones, the court stated that “the government must prove ... that the defendant intended to utter these words as a threat.” Baish, 460 A.2d at 42 (emphasis added) (citing Campbell and Gurley); Clark v. United States, 755 A.2d 1026, 1030 (D.C.2000) (citing Baish); Jenkins v. United States, 902 A.2d 79, 86 (D.C.2006) (citing Clark and Baish); Hunter v. United States, 980 A.2d 1158, 1163 (D.C.2009) (citing Jenkins); In re S.W., 45 A.3d at 155 (citing Clark and Baish). In none of these five cases, however, just as in the three which adopt the Campbell language, was the court presented with the issue now before us. Baish primarily involved territorial jurisdiction; Clark raised the question whether evidence of a threat made following an unlawful arrest should be suppressed as fruit of the poisonous tree; the issue in Jenkins was whether ostensibly harmless words constituted a threat because of the context in which they were spoken; Hunter raised an issue of merger; and In re £>.W. dealt largely with First Amendment questions that may arise in construing the threats statute. Use of the language in Baish rather than that in Campbell would not have affected the result in any of the five.
Finally, although Campbell was decided in 1982 and Baish in 1983, no opinion in either line of cases contains any expression of disagreement with the language of the other group. Indeed, Baish cited Campbell as precedent. Evidently, although the court in S.W. observed that the intent element “has been subject to controversy” due to the two “competing formulation[s]” described above, 45 A.3d at 155 n. 10, no judge of this court had suggested in any published opinion that we confront the apparent difference between the two articulations until this case came along.2 The two *174lines of cases have thus coexisted in our jurisprudence for three decades. There has been no “adversarial crossing of swords,” Murray v. District of Columbia, 870 A.2d 25, 27 (D.C.2005), in any of our cases with respect to which of the two competing statements of the intent element correctly applies the law of this jurisdiction. Under these circumstances, I am unable to agree with the government or with my colleagues in the majority that we are required by M.A.P. v. Ryan, or by Thomas v. United States, 731 A.2d 415, 420 & n. 6 (D.C.1999), to adhere to the language of Campbell and to reject that of Baish and of the four decisions that have followed Baish.
Nothing in M.A.P. v. Ryan requires a division of this court to follow language in a prior case which is unnecessary for the decision of that case. Albertie v. Louis & Alexander Corp., 646 A.2d 1001, 1005 (D.C.1994) (citations omitted). In District of Columbia v. Sierra Club, 670 A.2d 354, 360 (D.C.1996), we stated that “[t]he rule of stare decisis is never properly invoked unless in the decision put forward as precedent the judicial mind has been applied to and passed upon the precise question” (quoting Fletcher v. Scott, 201 Minn. 609, 277 N.W. 270, 272 (1938) (citations omitted)). “Questions which merely lurk in the record, neither brought to the attention of the court nor ruled upon, are not to be considered as having been so decided as to constitute precedents.” Id. (quoting Webster v. Fall, 266 U.S. 507, 511, 45 S.Ct. 148, 69 L.Ed. 411 (1925)); see also Thompson v. United States, 546 A.2d 414, 423 n. 14 (D.C.1988) (quoting Webster). There is, moreover, no consensus among the judges of our court as to which of the two articulations — Campbell or Baish — would have been more appropriate here (although the Baish language has been used more often).3 If we were nevertheless required to follow Campbell and not Baish, this would have to be because M.A.P. v. Ryan and Thomas should be deemed so inflexible that no distinction may be made between issues contested in the earlier case and those not contested. We would also be required to ignore the apparent rejection of the language of Campbell in a majority of our decisions since that case was decided. Albertie and Sierra Club establish that the doctrine of M.A.P. v. Ryan was not intended to be so expansive and rigid.
It is true that in Thomas, 731 A.2d at 420, we stated that “the rule is fundamental in our jurisprudence that no division of this court will overrule a prior decision of this court.” Washington v. Guest Servs., 718 A.2d 1071, 1075 (D.C.1998) (quoting M.A.P. v. Ryan, 285 A.2d at 312) (internal quotation marks omitted). “This court will not lightly deem one of its decisions to *175have been implicitly overruled and thus stripped of its precedential authority.” Lee v. United States, 668 A.2d 822, 828 (D.C.1995). Where a division of this court fails to adhere to earlier controlling authority, we are required to follow the earlier decision rather than the later one. See Taylor v. First Am. Title Co., 477 A.2d 227, 230 (D.C.1984); Thomas, 731 A.2d at 420 n. 6. In most circumstances — arguably in all circumstances, my colleagues perhaps assume — that would be that. Nevertheless, I know of no case in which the rule of M.A.P. v. Ryan has been applied to a situation comparable to the one before us here, in which not a single one of the decisions relied on arose in a case presenting the question as to which of the two articulations of the intent element is correct, so that this issue has never previously been litigated in this jurisdiction. Further, over the years, there appears to have been more support in the court for the 1983 articulation than for the one in 1982. Accordingly, I disagree with the government and the majority opinion as to the applicability here of M.A.P. v. Ryan.
III.
If, as I believe, neither Campbell nor Baish resolves the issue before us, then it surely remains an open question in this jurisdiction. As I see it, the notion that one can threaten another without intending that the words be uttered as a threat is, at least, counter-intuitive. As far as I am aware, folks threaten others intentionally, not inadvertently (except perhaps in situations in which nobody was meant to hear the threat). As this court observed in Baish, 460 A.2d at 42, and the majority reiterates here, ante at 167, Black’s Law Dictionary defines “threat” as “a communicated intent to inflict harm on any person or on property.” Intent to threaten is thus the essence of the concept. In Postell v. United States, 282 A.2d 551, 553 (D.C.1971), which the majority also quotes, ante at 166, we concluded that “the gist of the crime is that the words used are of such a nature as to convey a menace or fear of bodily harm to the ordinary hearer.” To “convey a menace” surely brims with intent to threaten. As a practical matter, to require proof only of intent to utter, and not of intent to threaten, would take any intent requirement out of the statute, for it would be a rare case indeed in which threatening words were uttered inadvertently or by accident, rather than intentionally. Yet although the statute does not state what intent is required, all of our cases, whether they follow Campbell or Baish, recognize that intent is the third element of the offense.
My colleagues in the majority insist that our en banc court has resolved the issue now before us in Holt v. United States, 565 A.2d 970, 971-72 (D.C.1989) (en banc). I disagree. The issue in Holt was whether intent to extort was an element of our felony threats statute. A majority of the court4 held that it was not. In the present case, Holt would support a holding that the government was not required to prove that Carrell intended to kill Ms. Ringen-burg, or perhaps to do her physical harm. The decision does not mean, however, that the prosecution need not prove that the threat was made intentionally. It is worth noting that the government does not rely on Holt, or even cite that case in its brief.5
*176Despite what I regard as the common sense of the issue — with few if any exceptions, threats are, by definition, intended to threaten — this case is difficult for me in light of the federal appellate decisions cited by the majority, ante at 169-70. See, e.g., United States v. DeAndino, 958 F.2d 146, 148-50 (6th Cir.1992). These decisions construe 18 U.S.C. § 875(c) (2012), which is similar (though not identical) to our statute in relevant respects, and a majority of them support the government’s position in this case, largely on the theory that threats are a “general intent” crime.6 There is, however, contrary authority as well. See, e.g. United States v. Sutcliffe, 505 F.3d 944, 953 (9th Cir.2007) (§ 875(c) requires proof of intent to threaten); United States v. Sovie, 122 F.3d 122, 125 (2d Cir.1997) (it is not necessary to prove that the defendant had the specific intent to carry out the threat, but only that he intended to communicate a threat); Robinson v. United States, 506 A.2d 572, 574 (D.C.1986) (in analogous prosecution for “intent to frighten” assault, ie., threats, prosecution must prove that the defendant intended either to cause injury or to create apprehension in the victim by engaging in some threatening conduct); Com. v. Spencer, 40 Mass.App.Ct. 919, 663 N.E.2d 268, 271 (1986) (in case of a threatened assault, “there must be an actual intention to cause apprehension.”) Further, although the phrasing of statutes prohibiting threats of bodily harm (as a form of assault, or as an independent offense) varies widely, “the gravamen of the offense is the intentional placing of a person in fear, [and] the intent to cause fear is an essential element of the crime.” 31A Am.Jur.2d Extortion, Blackmail and Threats § 48 (2013).
Because the intent of the threats statute with respect to the precise question before us is not at all clear, I believe that this is an appropriate occasion to include in our calculus the venerable but fundamentally just “rule of lenity.” As the Supreme Court explained in United States v. Bass, 404 U.S. 336, 347, 92 S.Ct. 515, 30 L.Ed.2d 488 (1971) (quoting United States v. Universal C.I.T. Credit Corp., 344 U.S. 218, 221-22, 73 S.Ct. 227, 97 L.Ed. 260 (1952)):
In various ways over the years, we have stated that ‘when choice has to be made between two readings of what conduct Congress has made a crime, it is appropriate, before we choose the harsher alternative, to require that Congress should have spoken in language that is clear and definite.’
Accord, Ruffin v. United States, 76 A.3d 845, 858 (D.C.2013) (discussing the District’s felony threats statute) (“criminal statutes are to be strictly construed and should not be interpreted to extend criminal liability beyond that which [the legislature] has plainly and unmistakably proscribed”). Accordingly, “[t]he rule of lenity requires ambiguous criminal statutes to be interpreted in favor of the defendants subjected to them.” United States v. Santos, 553 U.S. 507, 514, 128 *177S.Ct. 2020, 170 L.Ed.2d 912 (2008); accord, McBoyle v. United States, 283 U.S. 25, 27, 51 S.Ct. 340, 75 L.Ed. 816 (1931) (per Holmes, J.); United States v. Wiltberger, 5 Wheat. 76, 18 U.S. 76, 95, 5 L.Ed. 37 (1820) (per Marshall, C.J.). In this case, as in Lemon v. United States, 564 A.2d 1368, 1381 (D.C.1989) we are obliged to “consider the rule of lenity, which provides that criminal statutes should be strictly construed and genuine ambiguities resolved in favor of the defendant.” To be sure, the rule of lenity is a “secondary canon of construction.” Washington v. District of Columbia Dep’t of Pub. Works, 954 A.2d 945, 948 (D.C.2008). It “serves as an aid for resolving an ambiguity [and] is not to be used to beget one.” Callanan v. United States, 364 U.S. 587, 596, 81 S.Ct. 321, 5 L.Ed.2d 312 (1961). Simply put, it is the last resort, not the first. When, however, the intent of the statute cannot be determined from its language, purpose, or legislative history, any remaining ambiguity is to be resolved in favor of the accused, in conformity with the rule. Washington, 954 A.2d at 948-49.
In this case, the language of the statute does not answer the issue before us, no pertinent legislative history has been cited to us, and the coexistence for thirty years of the two competing lines of authority itself suggests that with respect to what is required to prove intent, resort to the rule of lenity is warranted and, in my view, should carry the day. Surely, the legislature has not “unmistakably proscribed,” Ruffin, supra, alleged threats where no intent to threaten has been shown.
In sum, there being no binding authority either way, I would hold, for the reasons stated above, that the prosecution must prove, in a threats case such as this, that the defendant intended to utter the words as a threat. I would affirm Carrell’s conviction for assault, vacate his conviction for attempted threats, and remand the case to the trial court for further proceedings. I would not necessarily require the judge to hold a new trial, for she could, in the exercise of her discretion, make what I regard as the requisite finding as to Carrell’s intent on the existing record.
I respectfully dissent.7
. The government also cites Gurley v. United States, 308 A.2d 785, 787 (D.C.1973), but as I read that opinion, it neither explicitly nor implicitly addresses the precise issue here presented. Further, in my view, Tolentino v. United States, 636 A.2d 433, 435 (D.C.1994); Beard v. United States, 535 A.2d 1373, 1378 (D.C.1988); and United States v. Smith, 337 A.2d 499, 503 (D.C.1975), cited by the majority, likewise did not address the question in dispute here, but concerned the second element of the offense of threats, rather than the third, which deals with the required intent.
. In Jenkins, 902 A.2d at 87 n. 11, the government made the same argument based on M.A.P. v. Ryan as it makes in this case, but we found it unnecessary to reach it.
. Perhaps this is not a case in which “statistics ... tell much, and courts listen,” cf. Harris v. District of Columbia Comm'n on Human Rights, 562 A.2d 625, 632 (D.C.1989) (quoting State of Alabama v. United States, 304 F.2d 583, 586 (5th Cir.1962)), but it is interesting to note that while six judges have written or joined opinions adhering to the Campbell phrasing (a seventh concurred only in the result in Campbell), thirteen have written or joined opinions adopting the Baish articulation (a fourteenth judge concurred only in part in Hunter). Three of the six judges who have written or joined Campbell language opinions, including one of the two judges in the majority in Campbell itself, have also written or joined opinions approving the language from Baish. This suggests that several judges — perhaps most — may believe that there is not a great difference between the analyses in Campbell and Baish, especially as the court in the latter case cited the former as authority. In any event, if the votes of the three judges who have agreed to both alternative phrasings, are excluded from my somewhat unorthodox "poll,” then we have, as it were, ten votes for the Baish language and only three for Campbell. I do not suggest that these numbers are dispositive, but they do add a little potentially instructive context.
. Whatever goes around comes around, but not always as might be anticipated. In Holt, decided almost a quarter of a century ago, I voted with the en banc majority to affirm the defendant’s conviction, while Judge Newman joined Judge Julia Cooper Mack’s powerful dissent. Now it is I who am outvoted! Sic transit gloria!
. Aside from its failure to cite the principal authorities relied on by my colleagues in sup*176port of their decision, the government has not treated the issue of intent in this context as one of major importance. In its brief, it has devoted only a single paragraph to the issue, primarily to argue that M.A.P. v. Ryan controls and nullifies the Baish line of cases. Moreover, in In re S.W., 45 A.3d at 155 n. 10, which was decided in 2012, and which is the last of the relevant cases to reach this court, the government did not contest, in the trial court or on appeal, the proposition that proof of intent to threaten was required.
. In its brief, the government alludes to what it calls the "well-accepted understanding that threats are a general-intent crime.” The only authority it cites for that proposition is Jones v. United States, 477 A.2d 231, 239 (D.C.1984), a first-degree murder case that did not involve the threats statute at all.
. At the conclusion of its opinion, the court states that by failing to request certain specific findings, Carrell has waived the claim that the judge erred by failing to do so. See Super. Ct.Crim. R. 23(c). I agree. I believe, however, that although Carrell's brief could be clearer on the point, his basic contention is that the judge applied the wrong legal standard.