dissenting.
The issue in this case is the meaning of the phrase, in 18 U. S. C. §2119, “with the intent to cause death or serious bodily harm.” (For convenience’ sake, I shall refer to it in this opinion as simply intent to kill.) As recounted by the Court, petitioner’s accomplice, Vernon Lennon, “testified that the plan was to steal the cars without harming the victims, but that he would have used his gun if any of the drivers had given him a‘hard time.’” Ante, at 4. The District Court instructed the jury that the intent element would be satisfied if petitioner possessed this “conditional” *13intent. Today’s judgment holds that instruction to have been correct.
I dissent from that holding because I disagree with the following, utterly central, passage of the opinion:
“[A] carjacker’s intent to harm his victim may be either 'conditional’ or ‘unconditional.’ The statutory phrase at issue theoretically might describe (1) the former, (2) the latter, or (3) both species of intent.” Ante, at 7 (footnote omitted).
I think, to the contrary, that in customary English usage the unqualified word “intent” does not usually connote a purpose that is subject to any conditions precedent except those so remote in the speaker’s estimation as to be effectively nonexistent — and it never connotes a purpose that is subject to a condition which the speaker hopes will not occur. (It is this last sort of “conditional intent” that is at issue in this case, and that I refer to in my subsequent use of the term.) “Intent” is “[a] state of mind in which a person seeks to accomplish a given result through a course of action.” Black’s Law Dictionary 810 (6th ed. 1990). One can hardly “seek to accomplish” a result he hopes will not ensue.
The Court’s division of intent into two categories, conditional and unconditional, makes the unreasonable seem logical. But Aristotelian classification says nothing about linguistic usage. Instead of identifying two categories, the Court might just as readily have identified three: unconditional intent, conditional intent, and feigned intent. But the second category, like the third, is simply not conveyed by the word “intent” alone. There is intent, conditional intent, and feigned intent, just as there is agreement, conditional agreement, and feigned agreement — but to say that in either case the noun alone, without qualification, “theoretically might describe” all three phenomena is simply false. Conditional intent is no more embraced by the unmodified word “intent” than a sea lion is embraced by the unmodified word “lion.”
*14If I have made a categorical determination to go to Louisiana for the Christmas holidays, it is accurate for me to say that I “intend” to go to Louisiana. And that is so even though I realize that there are some remote and unlikely contingencies — “acts of God,” for example — that might prevent me. (The fact that these remote contingencies are always implicit in the expression of intent accounts for the humorousness of spelling them out in such expressions as “if I should live so long,” or “the Good Lord willing and the creek don’t rise.”) It is less precise, though tolerable usage, to say that I “intend” to go if my purpose is conditional upon an event which, though not virtually certain to happen (such as my continuing to live), is reasonably likely to happen, and which I hope will happen. I might, for example, say that I “intend” to go even if my plans depend upon receipt of my usual and hoped-for end-of-year bonus.
But it is not common usage — indeed, it is an unheard-of usage — to speak of my having an “intent” to do something, when my plans are contingent upon an event that is not virtually certain, and that I hope will not occur. When a Mend is seriously ill, for example, I would not say that “I intend to go to his funeral next week.” I would have to make it clear that the intent is a conditional one: “I intend to go to his funeral next week if he dies.” The earjacker who intends to kill if he is met with resistance is in the same position: He has an “intent to kill if resisted”; he does not have an “intent to kill.” No amount of rationalization can change the reality of this normal (and as far as I know exclusive) English usage. The word in the statute simply will not bear the meaning that the Court assigns.
The Government makes two contextual arguments to which I should respond. First, it points out that the statute criminalizes not only carjackings accomplished by “force and violence” but also those accomplished by mere “intimidation.” Requiring an unconditional intent, it asserts, would make the number of covered carjackings accomplished by in*15timidation “implausibly small.” Brief for United States 22. That seems to me not so. It is surely not an unusual carjacking in which the criminal jumps into the passenger seat and forces the person behind the wheel to drive off at gunpoint. A carjacker who intends to kill may well use this modus operandi, planning to kill the driver in a more secluded location. Second, the Government asserts that it would be hard to imagine an unconditional-intent-to-kill case in which the first penalty provision of §2119 would apply, i. e., the provision governing cases in which no death or bodily harm has occurred. Id., at 23. That is rather like saying that the crime of attempted murder should not exist, because someone who intends to kill always succeeds.
Notwithstanding the clear ordinary meaning of the word “intent,” it would be possible, though of course quite unusual, for the word to have acquired a different meaning in the criminal law. The Court does not claim — and falls far short of-establishing — such “term-of-art” status. It cites five state cases (representing the majority view among the minority of jurisdictions that have addressed the question) saying that conditional intent satisfies an intent requirement; but it acknowledges that there are cases in other jurisdictions to the contrary. See ante, at 10, n. 9 (citing State v. Irwin, 55 N. C. App. 305, 285 S. E. 2d 345 (1982); State v. Kinnemore, 34 Ohio App. 2d 39, 295 N. E. 2d 680 (1972)); see also Craddock v. State, 204 Miss. 606, 37 So. 2d 778 (1948); McArdle v. State, 372 So. 2d 897 (Ala. Crim. App.), writ denied, 372 So. 2d 902 (Ala. 1979). As I understand the Court’s position, it is not that the former eases are right and the latter wrong, so that “intent” in criminal statutes, a term of art in that context, includes conditional intent; but rather that “intent” in criminal statutes may include conditional intent, depending upon the statute in question. That seems to me not an available option. It is so utterly clear in normal usage that “intent” does not include conditional intent, that only an accepted convention in the criminal law could *16give the word a different meaning. And an accepted convention is not established by the fact that some courts have thought so some times. One must decide, I think, which line of cases is correct, and in my judgment it is that which rejects the conditional-intent rule.
There are of course innumerable federal criminal statutes containing an intent requirement, ranging from intent to steal, see 18 U. S. C. §2113, to intent to defeat the provisions of the Bankruptcy Code, see § 152(5), to intent that a vessel be used in hostilities against a friendly nation, see §962, to intent to obstruct the lawful exercise of parental rights, see § 1204. Consider, for example, 21 U. S. C. § 841, which makes it a crime to possess certain drugs with intent to distribute them. Possession alone is also a crime, but a lesser one, see §844. Suppose that a person acquires and possesses a small quantity of cocaine for his own use, and that he in fact consumes it entirely himself. But assume further that, at the time he acquired the drug, he told his wife not to worry about the expense because, if they had an emergency need for money, he could always resell it. If conditional intent suffices, this person, who has never sold drugs and has never “intended” to sell drugs in any normal sense, has been guilty of possession with intent to distribute. Or consider 18 U. S. C. §2390, which makes it a crime to enlist within the United States “with intent to serve in armed hostility against the United States.” Suppose a Canadian enlists in the Canadian army in the United States, intending, of course, to fight all of Canada’s wars, including (though he neither expects nor hopes for it) a war against the United States. He would be criminally liable. These examples make it clear, I think, that the doctrine of conditional intent cannot reasonably be applied across-the-board to the criminal code. I am unaware that any equivalent absurdities result from reading “intent” to mean what it says— a conclusion strongly supported by the fact that the Government has cited only a single case involving another federal *17statute, from over two centuries of federal criminal jurisprudence, applying the conditional-intent doctrine (and that in circumstances where it would not at all have been absurd to require real intent).1 The course selected by the Court, of course — “intent” is sometimes conditional and sometimes not — would require us to sift through these many statutes *18one-by-one, making our decision on the basis of such ephemeral indications of “congressional purpose” as the Court has used in this case, to which I now turn.
Ultimately, the Court rests its decision upon the fact that the purpose of the statute — which it says is deterring carjacking — “is better served by construing the statute to cover both the conditional and the unconditional species of wrongful intent.” Ante, at 9. It supports this statement, both premise and conclusion, by two unusually uninformative statements from the legislative history (to stand out in that respect in that realm is quite an accomplishment) that speak generally about strengthening and broadening the carjacking statute and punishing earjackers severely. Ante, at 9, n. 7. But every statute intends not only to achieve certain policy objectives, but to achieve them by the means specified. Limitations upon the means employed to achieve the policy goal are no less a “purpose” of the statute than the policy goal itself. See Director, Office of Workers’ Compensation Programs v. Newport News Shipbuilding & Dry Dock Co., 514 U. S. 122, 185-136 (1995). Under the Court’s analysis, any interpretation of the statute that would broaden its reach would further the purpose the Court has found. Such reasoning is limitless and illogical.
The Court confidently asserts that “petitioner’s interpretation would exclude from the coverage of the statute most of the conduct that Congress obviously intended to prohibit.” Ante, at 9. It seems to me that one can best judge what Congress “obviously intended” not by intuition, but by the words that Congress enacted, which in this case require intent (not conditional intent) to kill. Is it implausible that Congress intended to define such a narrow federal crime? Not at .all. The era when this statute was passed contained well publicized instances of not only carjackings, and not only carjackings involving violence or the threat of violence (as, of course, most of them do); but also of carjackings in which the perpetrators senselessly harmed the ear owners *19when that was entirely unnecessary to the crime. I have a Mend whose father was killed, and whose mother was nearly killed, in just such an incident — after the car had already been handed over. It is not at all implausible that Congress should direct its attention to this particularly savage sort of carjacking — where killing the driver is part of the intended crime.2
Indeed, it seems to me much more implausible that Congress would have focused upon the ineffable “conditional intent” that the Court reads into the statute, sending courts and juries off to wander through “would-a, could-a, should-a” land. It is difficult enough to determine a defendant’s actual intent; it is infinitely more difficult to determine what the defendant planned to do upon the happening of an event that the defendant hoped would not happen, and that he himself may not have come to focus upon. There will not often be the accomplice’s convenient confirmation of conditional intent that exists in the present case. Presumably it will be up to each jury whether to take the carjacker (“Your car or *20your life”) at his word. Such a system of justice seems to me so arbitrary that it is difficult to believe Congress intended it. Had Congress meant to cast its carjacking net so broadly, it could have achieved that result — and eliminated the arbitrariness — by defining the crime as “carjacking under threat of death or serious bodily injury.” Given the language here, I find it much more plausible that Congress meant to reach — as it said — the caijacker who intended to kill.
In sum, I find the statute entirely unambiguous as to whether the earjacker who hopes to obtain the ear without inflicting harm is covered. Even if ambiguity existed, however, the rule of lenity would require it to be resolved in the defendant’s favor. See generally United States v. Wiltberger, 5 Wheat. 76, 95 (1820). The Government’s statement that the rule of lenity “has its primary application in eases in which there is some doubt whether the legislature intended to criminalize conduct that might otherwise appear to be innocent,” Brief for United States 81 (emphasis added), is carefully crafted to conceal the fact that we have repeatedly applied the rule to situations just like this. For example, in Ladner v. United States, 358 U. S. 169 (1958), the statute at issue made it a crime to assault a federal officer with a deadly weapon. The defendant, who fired one shotgun blast that wounded two federal officers, contended that under this statute he was guilty of only one, and not two, assaults. The Court said, in an opinion joined by all eight Justices who reached the merits of the case:
“This policy of lenity means that the Court will not interpret a federal criminal statute so as to increase the penalty that it places on an individual when such an interpretation can be based on no more than a guess as to what Congress intended. If Congress desires to create multiple offenses from a single act affecting more than one federal officer, Congress can make that meaning clear. We thus hold that the single discharge of a shot*21gun alleged by the petitioner in this case would constitute only a single violation of §254.” Id., at 178.
In Bell v. United States, 349 U. S. 81 (1955), the issue was similar: whether transporting two women, for the purpose of prostitution, in the same vehicle and on the same trip, constituted one or two violations of the Mann Act. In an opinion authored by Justice Frankfurter, the Court said:
“When Congress leaves to the Judiciary the task of imputing to Congress an undeclared will, the ambiguity should be resolved in favor of lenity. And this is not out of any sentimental consideration, or for want of sympathy with the purpose of Congress in proscribing evil or antisocial conduct. It may fairly be said to be a presupposition of our law to resolve doubts in the enforcement of a penal code against the imposition of a harsher punishment.” Id., at 83.
If that is no longer the presupposition of our law, the Court should say so, and reduce the rule of lenity to a historical curiosity. But if it remains the presupposition, the rule has undeniable application in the present case. If the statute is not, as I think, clear in the defendant’s favor, it is at the very least ambiguous and the defendant must be given the benefit of the doubt.
* * *
This seems to me not a difficult case. The issue before us is not whether the “intent” element of some common-law crime developed by the courts themselves — or even the “intent” element of a statute that replicates the common-law definition — includes, or should include, conditional intent. Rather, it is whether the English term “intent” used in a statute defining a brand new crime bears a meaning that contradicts normal usage. Since it is quite impossible to say that longstanding, agreed-upon legal usage has converted this word into a term of art, the answer has to be no. And it would be no even if the question were doubt*22ful. I think it particularly inadvisable to introduce the new possibility of “conditional-intent” prosecutions into a modern federal criminal-law system characterized by plea bargaining, where they will predictably be used for in terrorem effect. I respectfully dissent.
The one case the Government has come up with is Shaffer v. United States, 308 F. 2d 654 (CA5 1962), cert. denied, 373 U. S. 939 (1963), which upheld a conviction of assault “with intent to do bodily harm” where the defendant had said that if any persons tried to leave the building within five minutes after his departure “he would shoot their heads off,” 308 F. 2d, at 655. In my view, and in normal parlance, the defendant did not “intend” to do bodily harm, and there would have been nothing absurd about holding to that effect.
The Government cites six other federal cases, Brief for United States 14-15, n. 5, but they are so inapposite that they succeed only in demonstrating the weakness of its assertion that conditional intent is the federal rule. Two of them, United States v. Richardson, 27 F. Cas. 798 (No. 16,155) (CCDC 1837), and United States v. Myers, 27 F. Cas. 43 (No. 15,845) (CCDC 1806), involve convictions for simple assault with no specific intent, and do not even contain any dictum bearing upon the present question. A third, United States v. Arrellano, 812 F. 2d 1209, 1212, n. 2 (CA9 1987), contains nothing but dictum, since the jury found no intent of any sort. A fourth, United States v. Marks, 29 M. J. 1 (Ct. Mil. App. 1989), involved a defendant who tried to set fire to material that he assertedly believed was flame resistant. The crime he was convicted of, aggravated arson, was, as the court specifically stated, “a general intent crime,” id., at 3. And the last two cases, United States v. Dworken, 855 F. 2d 12 (CA1 1988), and United States v. Anello, 765 F. 2d 253 (CA1), cert. denied sub nom. Wendolkowski v. United States, 474 U. S. 996 (1985), both involved conspiracy to possess drugs with intent to distribute. Defendants contended that they could not be convicted because they did not intend to complete the conspired-for transaction unless the quality of the drugs (and, in the case of Dworken, the price as well) was satisfactory. Of course the intent necessary to conspire for a specific-intent crime is not the same as the intent necessary for the crime itself, particularly insofar as antecedent conditions are concerned. And in any event, since it can hardly be thought that the conspirators wanted the quality and price of the drugs to be inadequate, neither case involved the conditional intent that is the subject of the present ease.
Note that I am discussing what was a 'plausible congressional purpose in enacting this language — not what I necessarily think was the real one. I search for a plausible purpose because a text without one may represent a "scrivener’s error” that we may properly correct. See Green v. Bock Laundry Machine Co., 490 U. S. 504, 528-529 (1989) (Scalia, J., concurring in judgment); see also United States v. X-Citement Video, Inc., 513 U. S. 64, 82 (1994) (Scalia, J., dissenting). There is no need for such correction here; the text as it reads, unamended by a meaning of “intent” that contradicts normal usage, makes total sense. If I were to speculate as to the real reason the “intent” requirement was added by those who drafted it, I think I would select neither the Court’s attribution of purpose nor the one I have hypothesized. Like the District Court, see 921 F. Supp. 155, 158 (EDNY 1996), and the Court of Appeals for the Third Circuit, see United States v. Anderson, 108 F. 3d 478, 482-483 (1997), I suspect the “intent” requirement was inadvertently expanded beyond the new subsection 2119(3), which imposed the death penalty — where it was thought necessary to ensure the constitutionality of that provision. Of course the actual intent of the draftsmen is irrelevant; we are governed by what Congress enacted.