with whom
Justice Blackmun and Justice O’Connor join, dissenting.Congress sometimes uses slightly different language to convey the same message. Thus, Congress uses the terms “subsequent offense,” “second or subsequent offense,” and “second or subsequent conviction” in various sections of the Criminal Code, all to authorize enhanced sentences for repeat offenders.1 On some occasions, Congress meticulously defines the chosen term to identify those offenses committed after a prior conviction “has become final”;2 more frequently, *138it relies on settled usage and the reader’s common sense to impart the same meaning.
In certain sections of the Code, even absent a definition, the context makes perfectly clear that the word “subsequent” describes only those offenses committed after a prior conviction has become final. Title 18 U. S. C. § 1302, for instance, which prohibits mailing of lottery tickets, authorizes a 5-year prison sentence for “any subsequent offense.” A literal reading of that phrase, like the one adopted by the majority today, presumably would justify imposition of five 5-year sentences if a defendant who sold six lottery tickets through the mail were charged in a single indictment. But it is absurd to think that Congress intended to treat such a defendant as a repeat offender, subject to penalty enhancement, “simply because he managed to evade detection, prosecution, and conviction for the first five offenses and was ultimately tried for all six in a single proceeding.” Ante, at 137.
In other Code sections, where context is less illuminating, the long-established usage of the word “subsequent” to distinguish between first offenders and recidivists is sufficient to avoid misunderstanding by anyone familiar with federal criminal practice.3 Thus, in a 1955 opinion construing the undefined term “subsequent offense,” the First Circuit noted that most “subsequent offender” statutes had been construed to provide that any offense “committed subsequent to a conviction calls for the increased penalty.” Gonzalez v. United States, 224 F. 2d 431, 434 (1955). The court continued:
“In the United States courts uniformly this has been held to be the rule. In Singer v. United States, [278 F. 415 (1922)], the Court of Appeals for the Third Circuit considered a substantially similar statute to that presently before us and held that a second offense within the *139meaning of the statute could occur only after a conviction for the first offense. See, e. g., United States v. Lindquist, [285 F. 447 (WD Wash. 1921)], and Biddle v. Thiele, [11 F. 2d 235 (CA8 1926)]. The Court of Appeals for the Fifth Circuit said in Holst v. Owens, [24 F. 2d 100, 101 (1928)]: Tt cannot legally be known that an offense has been committed until there has been a conviction. A second offense, as used in the criminal statutes, is one that has been committed after conviction for a first offense. Likewise, a third or any subsequent offense implies a repetition of crime after each previous conviction.’ Similarly, in Smith v. United States, [41 F. 2d 215, 217 (CA9 1930)], the court stated: Tn order that a conviction shall affect the penalty for subsequent offenses, it must be prior to the commission of the offense.’ ” Ibid.
Congress did not define the term “subsequent conviction” when it enacted § 924(e) in 1968. It is fair to presume, however, that Congress was familiar with the usage uniformly followed in the federal courts. See NLRB v. Amax Coal Co., 453 U. S. 322, 329 (1981); Perrin v. United States, 444 U. S. 37, 42-45 (1979). Indeed, given the settled construction of repeat offender provisions, it is hardly surprising that Congressman Poff, who proposed the floor amendment that became § 924(c), felt it unnecessary to elaborate further. Cf. Morissette v. United States, 342 U. S. 246, 263 (1952) (“[W]here Congress borrows terms of art... absence of contrary direction may be taken as satisfaction with widely accepted definitions, not as a departure from them”). It is also unsurprising that there appears to have been no misunderstanding of the term “second or subsequent conviction” for almost 20 years after the enactment of § 924(c).
Section 924(e) was construed by this Court for the first time in Simpson v. United States, 435 U. S. 6 (1978), a case involving sentencing of a defendant who had committed two bank robberies, two months apart. Convicted in two sepa*140rate trials, the defendant was sentenced in each for bank robbery, and in each to 10 years under § 924(c), then the maximum authorized term for a first-time offender. Id., at 9. Apparently, nobody considered the possibility that the defendant might have been treated as a repeat offender at his second trial, and sentenced under § 924(c)’s “second or subsequent conviction” provision. ’ In any event, despite the fact that the literal language of the statute would have authorized the § 924(c) sentences, id., at 16-17 (Rehnquist, J., dissenting), the Court set them aside, applying the rule of lenity and concluding that Congress did not intend enhancement under § 924(c) when, as in Simpson’s case, a defendant is also sentenced under a substantive statute providing for an enhancement for use of a firearm. Id., at 14-15.
In Busic v. United States, 446 U. S. 398 (1980), the Court construed the first offender portion of § 924(e) even more narrowly than in Simpson, again rejecting a literal reading of the statutory text that would have supported a contrary result. In his dissenting opinion, Justice Stewart succinctly described § 924(c) as a “general enhancement provision— with its stiff sanctions for first offenders and even stiffer sanctions for recidivists.”4 This understanding that the term “second or subsequent conviction” was used to describe recidivism seemingly was shared by other judges, as several years were to elapse before the construction adopted by the Eleventh Circuit in United States v. Rawlings, 821 F. 2d *1411543, cert. denied, 484 U. S. 979 (1987), and endorsed by the Court today, appeared in any reported judicial opinion.
At oral argument, the Government was unable to tell us how the “second or subsequent conviction” language of §,924(c) was construed by Government prosecutors prior to 1987, when Rawlings was decided. Tr. of Oral Arg. 27-28. It seems to me, however, quite likely that until 1987, the Government read the “second or subsequent” section of § 924(e) as a straightforward recidivist provision, just as Justice Stewart did in 1980. That reading certainly would comport with the Government’s submissions to this Court in Simpson, supra, and Busic, supra, both of which describe the “second or subsequent conviction” provision in terms of recidivism.5 It would be consistent, too, with the reported cases involving § 924(c) sentencing, which make clear that the district courts were routinely imposing consecutive 5-year sentences when defendants were convicted of two separate offenses under § 924(e), apparently without objection from the Government that the second conviction warranted a longer sentence. See, e. g., United States v. Henry, 878 F. 2d 937, 938 (CA6 1989); United States v. Jim, 865 F. 2d 211, 212 (CA9), cert. denied, 493 U. S. 827 (1989); United States v. Fontanilla, 849 F. 2d 1257, 1258 (CA9 1988); United States v. Chalan, 812 F. 2d 1302, 1315 (CA10 1987), cert. denied, 488 U. S. 983 (1988).
In light of this history, I would find no ambiguity in the phrase “subsequent conviction” as used in § 924(e). Like its many counterparts in the Criminal Code, the phrase clearly is intended to refer to a conviction for an offense committed *142after an earlier conviction has become final; it is, in short, a recidivist provision. When that sensible construction is adopted, of course, the grammatical difficulties and the potential for prosecutorial manipulation that trouble the majority, see ante, at 181-134, are avoided entirely. See United States v. Neal, 976 F. 2d 601, 603 (CA9 1992) (Fletcher, J., dissenting) (“common-sense reading of § 924(c)” as recidivist statute).
Even assuming, however, that the meaning of § 924(c)’s repeat offender provision is not as obvious as I think, its history belies the notion that its text admits of only one reading, that adopted in Rawlings. Surely it cannot be argued that a construction surfacing for the first time 19 years after enactment is the only available construction. Indeed, even after Rawlings, there is no consensus on this point; some courts — and some Government prosecutors — continue to apply § 924(c) as a recidivist statute.6 In United States v. Nabors, 901 F. 2d 1351 (CA6), cert. denied, 498 U. S. 871 (1990), for instance, a ease decided in 1990, the Court of Appeals purported to follow Rawlings, but actually affirmed imposition of two 5-year sentences for convictions on two distinct § 924(c) violations.7 Similarly, in United States v. *143Luskin, 926 F. 2d 372 (CA4), cert. denied, 502 U. S. 815 (1991), decided a year later, the Court of Appeals upheld three 5-year sentences for three violations of § 924(c) committed on separate dates, even though the minimum mandatory penalty for a “second or subsequent conviction” was 10 years at the time of trial. Significantly, the Government did not challenge the 5-year sentences on the second and third convictions.8
At the very least, this equivocation on the part of those charged with enforcing § 924(c), combined with the understanding of repeat offender provisions current when § 924(c) was enacted, render the construction of § 924(e) sufficiently uncertain that the rule of lenity should apply. Cf. Simpson, 435 U. S., at 14-15; see United States v. Abreu, 962 F. 2d 1447, 1450-1451 (CA10 1992) (en banc). As one District Court judge said of § 924(c), in the course of a 1991 sentencing:
“The statute is not a model of clarity. Its use of the word ‘conviction’ rather than wording describing the offense suggests an intent to reach recidivists who repeat conduct after conviction in the judicial system for prior offenses. The legislative history suggests that Congress was trying to impose draconian punishment ‘if he does it a second time.’ 114 Cong. Rec. 22231, 22237 (1968). It is unclear whether this means a second time as a recidivist or a second time offender who has not faced deterrence by a prior sentence. Criminal statutes must be strictly construed. Nabors [901 F. 2d, at *1441358] said that ‘§ 924(c)(1) is, at best, hard to follow in simple English . .With Mr. Godwin in front of me, I decline to hold him to a higher test than one found difficult by appellate court judges.” United States v. Godwin, 758 F. Supp. 281, 283 (ED Pa. 1991).
In an effort to cure § 924(c) of any ambiguity, the Court undertakes an intricate grammatical analysis, with an emphasis on the word “conviction.”9 According to the Court, the “conviction” referred to in § 924(c) must be a finding of guilt, preceding the entry of final judgment, because sentence is imposed with the final judgment; if “conviction” referred to the final judgment itself, there would be no opportunity for sentence enhancement. Ante, at 132. The “absurdity]” of this situation, ibid., which, I note, has thus far eluded all of the courts to apply § 924(c) as a recidivist *145statute, see supra, evaporates if we assume that sentencing judges are gifted with enough common sense to understand that they may, upon entry of a second final judgment, enhance the sentence incorporated therein. In any event, the majority’s conclusion that a “second or subsequent conviction” is a finding of guilt leaves unanswered the question dispositive here: whether that second conviction (finding of guilt or entry of judgment) is subject to enhancement if it is not for an offense committed after a prior conviction has become final.
The Court finds additional support for its conclusion in the fact that at least some contrary readings of § 924(c) would “give a prosecutor unreviewable discretion either to impose or to waive the enhanced sentencing provisions” through the manner in which she charged a crime or crimes. Ante, at 133. I have already pointed out that the majority’s particular concern is not implicated if § 924(c) is treated as a straightforward recidivist provision, supra, at 142-143; under that construction, a defendant who commits a second § 924(c) offense before trial on the first would not be eligible. for sentence enhancement whether the two counts were tried separately or together. I would add only that the Court’s alternative reading does not solve the broader problem it identifies. As the Government concedes, see Tr. of Oral Arg. 31-32, prosecutors will continue to enjoy considerable discretion in deciding how many § 924(c) offenses to charge in relation to a criminal transaction or series of transactions. An armed defendant who robs a bank and, at the same time, assaults a guard, may be subject to one or two § 924(c) charges; the choice is the prosecutor’s, and the consequence, under today’s holding, the difference between a 5- and a 15-year enhancement. Cf. United States v. Jim, 865 F. 2d, at 212 (defendant charged with three counts under § 924(c), each arising from the same criminal episode); United States v. Fontanilla, 849 F. 2d, at 1257 (same).
*146Section 924(c) of the Criminal Code mandates an enhanced, 20-year sentence for repeat offenders. Between 1968, when the statute was enacted, and 1987, when textualism replaced common sense in its interpretation, the bench and bar seem to have understood that this provision applied to defendants who, having once been convicted under § 924(c), “failed to learn their lessons from the initial punishment” and committed a repeat offense. See United States v. Neal, 976 F. 2d, at 608 (Fletcher, J., dissenting).10 The contrary reading adopted by the Court today, driven by an elaborate exercise in sentence parsing, is responsive to neither historical context nor common sense. Because I cannot agree with this unwarranted and unnecessarily harsh construction of § 924(c), the meaning of which should, at a minimum, be informed by the rule of lenity, I respectfully dissent.
See, e. g., 18 U. S. C. § 1302 ("subsequent offense” related to mailing of lottery tickets); § 1735 (“second or subsequent offense” related to sexually oriented advertising); § 844(h) (“second or subsequent conviction” for felonious use of explosives).
See, e. g., 21 U. S. C. § 859(b) (1988 ed., Supp. III) (distribution of drugs to minors); 21 U. S. C. § 860(b) (1988 ed., Supp. Ill) (distribution of drugs near schools); 21 U. S. C. § 962(b) (importation of controlled substances).
See, e.g., 18 U. S. C. §2114 (“subsequent offense” of mail robbery), as interpreted in United States v. Cooper, 580 F. 2d 259, 261 (CA7 1978) (“obvious” that “subsequent offense” language must be read as applying only to offenses committed after conviction on a prior offense).
446 U. S., at 416. His full comment:
“I agree with the holding in Simpson that Congress did not intend to ‘pyramid' punishments for the use of a firearm in a single criminal transaction. Yet I find quite implausible the proposition that Congress, in enacting § 924(c)(1), did not intend this general enhancement provision — with its stiff sanctions for first offenders and even stiffer sanctions for recidivists — to serve as an alternative source of enhanced punishment for those who commit felonies, such as bank robbery and assaulting a federal officer, that had been previously singled out by Congress as warranting special enhancement, but for which a lesser enhancement sanction than that imposed by § 924(c) had been authorized.”
See Brief for United States in Busic v. United States, O. T. 1979, No. 78-6020, p. 19 (“Section 924(c) establishes mandatory minimum sentences, requires increasingly severe sentences for recidivists (without possibility of suspension or probation), and prohibits concurrent sentencing”); Brief for United States in Simpson v. United States, O. T. 1977, No. 76-5761, pp. 13-14 (discussing application of sentencing provisions “[i]f the gun-wielding bank robber were a recidivist”).
Dismissing these eases, as well as those decided pre-Rawlings, as a long line of “erroneous lower-court decisions,” ante, at 135, cannot explain why 19 years passed before the correct interpretation of a statute of “utterly no ambiguity,” ibid., made its first reported appearance.
There is some tension between the notion that the text of the'statute is clear and unambiguous and the Court of Appeals’ explanation for its holding:
“While § 924(c)(1) is, at best, hard to follow in simple English, we concur with the reasoning in Rawlings that two distinct violations of the statute trigger the subsequent sentence enhancement provisions of § 924(c)(1). Thus, the commission of two violations of § 924(c)(1) would result in a five-year consecutive sentence for the first conviction and a ten-year consecutive sentence for the second § 924(c)(1) conviction. However, because of the complexity of this issue, we find the district court’s failure to sentence Nabors to a ten-year consecutive sentence for his second § 924(c)(1) conviction not clearly erroneous.” United States v. Nabors, 901 F. 2d, at 1358-1359.
“The 1988 amendment raised the penalty for repeat violators of the statute to twenty years. In the version that was in effect at the time of the present crimes, the penalty for repeat violators was ten years. Arguably, the district judge should have sentenced appellant to one five-year and two ten-year consecutive terms of imprisonment for his convictions' under Counts V through VII. However, since the United States has not counter-appealed on this point, we will not address it.” United States v. Luskin, 926 F. 2d, at 374, n. 2.
The Court also suggests that use of the word “conviction,” rather than “offense,” distinguishes this statute from the repeat offender provisions discussed in Gonzalez v. United States, 224 F. 2d 431 (CA1 1965), supra, at 138-139. Of course, the majority’s textualist approach would lead to the same result if §924(c)'s enhancement were reserved for “second or subsequent offenses”: At the time of sentencing for two violations committed on separate dates, one violation is “second or subsequent” to the other, and the conviction itself always will establish that two “offenses” have indeed been committed. See ante, at 135.
It is true, as the Court points out in passionate defense of its reading, that the words “offense” and “conviction” are not identical. What is at issue here, however, is not whether the terms mean the same thing in all usages, but whether they mean the same thing when they are used by Congress to identify the class of repeat offenders subject to enhanced sentences. Cf. ante, at 131-132 (context gives meaning to word “conviction”). If there is any difference between the terms as so used, it only lends further support to the conclusion that § 924(e) is a recidivist provision. As discussed above, repeat offender statutes couched in terms of “offense” were understood at the time of § 924(c)’s enactment to identify offenses committed after a prior conviction. See supra, at 138-139. A fortiori, “use of the word ‘conviction’ rather than wording describing the offense suggests an intent to reach recidivists who repeat conduct after conviction in the judicial system for prior offenses.” United States v. Godwin, 758 F. Supp. 281, 283 (ED Pa. 1991).
“However, punishing first offenders with twenty-five-year sentences does not deter crime as much as it ruins lives. If, after arrest and conviction, a first offender is warned that he will face a mandatory twenty-year sentence if he commits the same crime again, then the offender will know of the penalty. Having already served at least five years in prison, he will have a strong incentive to stay out of trouble. Discouraging recidivism by people who have already been in prison and been released serves a far more valuable purpose than deterring offenders who have yet to be arrested and have no knowledge of the law’s penalties.” United States v. Jones, 965 F. 2d 1507, 1521 (CA8 1992) (internal citation omitted).