with whom Chief Justice Rehnquist and Justice Kennedy join, dissenting.
The State of New York seeks to prosecute respondent a second time for the actions that he took at 6:35 p.m. on October 3, 1987. If the Double Jeopardy Clause guaranteed the right not to be twice put in jeopardy for the same conduct, it would bar this second prosecution. But that Clause guarantees only the right not to be twice put in jeopardy for the same offense, and has been interpreted since its inception, as was its common-law antecedent, to permit a prosecution based upon the same acts but for a different crime. The Court today holds otherwise, departing from clear text and clear precedent with no justification except the citation of dictum in a recent case (dictum that was similarly unsupported, and inconclusive to boot). The effects of this innova*527tion upon our criminal justice system are likely to be substantial. In practice, it will require prosecutors to observe a rule we have explicitly rejected in principle: that all charges arising out of a single occurrence must be joined in a single indictment. Because respondent is not being prosecuted for the same offense for which he was previously prosecuted, I would reverse the judgment.
I
The Double Jeopardy Clause, made applicable to the States by the Fourteenth Amendment, Benton v. Maryland, 395 U. S. 784, 794 (1969), provides: “[N]or . shall any person be subject for the same offence to be twice put in jeopardy of life or limb.” U. S. Const., Arndt. 5. It “‘protects] an individual from being subjected to the hazards of trial and possible conviction more than once for an alleged offense.’” Burks v. United States, 437 U. S. 1, 11 (1978), quoting Green v. United States, 355 U. S. 184, 187 (1957). In Blockburger v. United States, 284 U. S. 299, 304 (1932), we summarized the test for determining whether conduct violating two distinct statutory provisions constitutes the “same offence” for double jeopardy purposes:
“The applicable rule is that where the same act or transaction constitutes a violation of two distinct statutory provisions, the test to be applied to determine whether there are two offenses or only one, is whether each provision requires proof of a fact which the other does not. Gavieres v. United States, 220 U. S. 338, 342 [(1911]), and authorities cited. In that case this court quoted from and adopted the language of the Supreme Court of Massachusetts in Morey v. Commonwealth, 108 Mass. 433 [(1871)]: ‘A single act may be an offense against two statutes; and if each statute requires proof of an additional fact which the other does not, an acquittal or conviction under either statute does not exempt the defend*528ant from prosecution and punishment under the other.’” Ibid.
Blockburger furnishes, we have observed, the “established test” for determining whether successive prosecutions arising out of the same events are for the “same offence.” Brown v. Ohio, 432 U. S. 161, 166 (1977). This test focuses on the statutory elements of the two crimes with which a defendant has been charged, not on the proof that is offered or relied upon to secure a conviction. “If each [statute] requires proof of a fact that the other does not, the Blockburger test is satisfied, notwithstanding a substantial overlap in the proof offered to establish the crimes.” Iannelli v. United States, 420 U. S. 770, 785, n. 17 (1975); see also Gore v. United States, 357 U. S. 386 (1958); American Tobacco Co. v. United States, 328 U. S. 781, 788-789 (1946).
We have departed from Blockburger*s exclusive focus on the statutory elements of crimes in only two situations. One occurs where a statutory offense expressly incorporates another statutory offense without specifying the latter’s elements. For example, in Harris v. Oklahoma, 433 U. S. 682 (1977) (per curiam), we held that a conviction for felony murder based on a killing in the course of an armed robbery barred subsequent prosecution for the underlying robbery. Although the second prosecution would not have been barred under the Blockburger test (because on its face the Oklahoma felony-murder statute did not require proof of robbery, but only of some felony), the second prosecution was impermissible because it would again force the defendant to defend against the charge of robbery. The other situation in which we have relaxed the Blockburger “elements” test occurs where a second prosecution would require relitigation of factual issues that were necessarily resolved in the defendant’s favor in the first prosecution. See Ashe v. Swenson, 397 U. S. 436 (1970).
Subject to the Harris and Ashe exceptions, I would adhere to the Blockburger rule that successive prosecutions under *529two different statutes do not constitute double jeopardy if each statutory crime contains an element that the other does not, regardless of the overlap between the proof required for each prosecution in the particular case. That rule best gives effect to the language of the Clause, which protects individuals from being twice put in jeopardy “for the same offence,” not for the same conduct or actions. “Offence” was commonly understood in 1791 to mean “transgression,” that is, “the Violation or Breaking of a Law.” Dictionarium Britannicum (Bailey ed. 1730); see also J. Kersey, A New English Dictionary (1702); 2 T. Sheridan, A General Dictionary of the English Language (1780); J. Walker, A Critical Pronouncing Dictionary (1791); 2 N. Webster, An American Dictionary of the English Language (1828). If the same conduct violates two (or more) laws, then each offense may be separately prosecuted. Of course, this is not to say that two criminal provisions create “distinct” offenses simply by appearing under separate statutory headings; but if each contains an element the other does not, i. e., if it is possible to violate each one without violating the other, then they cannot constitute the “same offence.”
Another textual element also supports the Blockburger test. Since the Double Jeopardy Clause protects the defendant from being “twice put in jeopardy,” i. e., made to stand trial (see, e. g., Respublica v. Shaffer, 1 Dall. 236, 237 (Pa. 1788)), for the “same offence,” it presupposes that sameness can be determined before the second trial. Otherwise, the Clause would have prohibited a second “conviction” or “sentence” for the same offense. A court can always determine, before trial, whether the second prosecution involves the “same offence” in the Blockburger sense, since the Constitution entitles the defendant “to be informed of the nature and cause of the accusation.” Arndt. 6. But since the Constitution does not entitle the defendant to be informed of the evidence against him, the Court’s “proof-of-same-conduct” test will be implementable before trial only if the indictment hap*530pens to show that the same evidence is at issue, or only if the jurisdiction’s rules of criminal procedure happen to require the prosecution to submit a bill of particulars that cannot be exceeded. More often than not, in other words, the Court’s test will not succeed in preventing the defendant from being tried twice.
Relying on text alone, therefore, one would conclude that the Double Jeopardy Clause meant what Blockburger said. But there is in addition a wealth of historical evidence to the same effect. The Clause was based on the English common-law pleas of auterfoits acquit and auterfoits convict, which pleas were valid only “upon a prosecution for the same identical act and crime.” 4 W. Blackstone, Commentaries 330 (1769) (emphasis added). In that respect they differed from the plea of auterfoits attaint, which could be invoked by any person under a sentence of death “whether it be for the same or any other felony.” Ibid.
The English practice, as understood in 1791, did not recognize auterfoits acquit and auterfoits convict as good pleas against successive prosecutions for crimes whose elements were distinct, even though based on the same act. An acquittal or conviction for larceny, for example, did not bar a trial for trespass based on “the same taking, because Trespass and Larceny are Offences of a different Nature, and the Judgment for the one entirely differs from that for the other.” 2 W. Hawkins, Pleas of the Crown, ch. 36, §7, p. 376 (4th ed. 1762); see also id., ch. 35, §5, at 371. Sir Matthew Hale described the rule in similar terms:
“If A. commit a burglary in the county of B. and likewise at the same time steal goods out of the house, if he be indicted of larciny for the goods and acquitted, yet he may be indicted for the burglary notwithstanding the acquittal.
“And e converso, if indicted for the burglary and acquitted, yet he may be indicted of the larciny, for they are several offences, tho committed at the same time. *531And burglary may be where there is no larciny, and larciny may be where there is no burglary.
“Thus it hath happened, that a man acquitted for stealing the horse, hath yet been arraigned and convict for stealing the saddle, tho both were done at the same time.” 2 M. Hale, Pleas of the Crown, ch. 31, pp. 245-246 (1736 ed.).
Treatises of a slightly later vintage are in accord. Thomas Starkie (frequently cited in early American cases) says:
“The plea [of auterfoits acquit] will be vicious if the offences charged in the two indictments be perfectly distinct in point of law, however nearly they may be connected in fact.
“So if the defendant be first indicted upon the more general charge, consisting of the circumstances A. and B. only, an acquittal obviously includes an acquittal from a more special charge consisting of the circumstances A. B. and C. for if he be not guilty of the former, he cannot be guilty of those with the addition of a third. But if one charge consist of the circumstances A. B. C. and another of the circumstances A. D. E. then, if the circumstance which belongs to them in common does not of itself constitute a distinct substantive offence, an acquittal from the one charge cannot include an acquittal of the other.” 1 T. Starkie, Criminal Pleading, ch. xix, pp. 322-323 (2d ed. 1822).
Likewise:
“The plea [of auterfoits acquit] cannot be sustained if the offences charged in the two indictments are in conteinplation of law dissimilar from each other, however nearly analogous in fact and in circumstances .... [I]f the former charge were such a one as the defendant could not have been convicted of the latter upon it, the acquittal *532cannot be pleaded.” 2 C. Petersdorff, Abridgment 738, n. (1825).
See also 1 J. Chitty, Criminal Law 455-457 (1816).
The cases from this period are few, but they lend support to this view. In Turner’s Case, Kelyng 30, 84 Eng. Rep.. 1068 (K. B. 1708), the defendant was acquitted on an indictment charging burglary by breaking and entering the house of Tryon and taking away great sums of money. Turner was again indicted for burglary by breaking and entering the house of Tryon and removing the money of Tryon’s servant. The court held that Turner could not “now be indicted again for the same burglary for breaking the house; but we all agreed, he might be indicted for felony, for stealing the money of [the servant]. For they are several felonies, and he was not indicted of this felony before . . . .” Even the holding of Turner’s Case — that the second indictment charged the same felony of burglary — was limited in the famous case of King v. Vandercomb, 2 Leach 708, 168 Eng. Rep. 455 (K. B. 1796). There, the defendants were first charged with burglary by breaking and entering a house and stealing goods. The Crown abandoned the prosecution because it developed at trial that the defendants had not removed any property. In a second prosecution for burglary by breaking and entering with intent to steal, the plea of auterfoits acquit was held bad:
“The circumstance of breaking and entering the house is common and essential to both the species of this offence; but it does not of itself constitute the crime in either of them; for it is necessary, to the completion of burglary, that there should not only be a breaking and entering, but the breaking and entering must be accompanied with a felony actually committed or intended to be committed; and these two offences are so distinct in their nature, that evidence of one of them will not support an indictment for the other.” Vandercomb, supra, at 717, 168 Eng. Rep., at 460 (citations omitted).
*533The court’s statement in Vandercomb that the “evidence of one of them will not support an indictment for the other,” see also id., at 720, 168 Eng. Rep., at 461, is the precise equivalent of our statement in Blockburger that “each provision requires proof of a fact which the other does not.” 284 U. S., at 304.
The early American cases adhere to the same rule. In State v. Sonnerkalb, 2 Nott & McCord 280 (S. C. 1820), the defendant was first convicted of retailing liquor without a license. He was then tried a second time for “dealing, trading or trafficking with a negro,” id., at 281, based on the same sale, and “the same evidence was given on the part of the state,” id., at 280. The court rejected the defendant’s claim that he had been convicted twice for the same offense: “[L]et it be admitted, that the defendant committed physically but one act; two offences may be committed by one act ... .” Id., at 283. Since the first offense required proof of retailing liquor (but it was “immaterial to whom he [did] retail,” id., at 282), and the second required proof of sale to a Negro (but it was immaterial what product he sold), the two offenses were different “in legal contemplation.” Ibid.
Commonwealth v. Roby, 12 Pickering 496 (Mass. 1832), after analyzing King v. Vandercomb and Chitty’s treatise, distilled the rule as follows:
“In considering the identity of the offence, it must appear by the plea, that the offence charged in both cases was the same in law and in fact. The plea will be vicious, if the offences charged in the two indictments be perfectly distinct in point of law, however nearly they may be connected in fact. . . . [I]t is sufficient if an acquittal from the offence charged in the first indictment virtually includes an acquittal from that set forth in the second, however they may differ in degree. Thus an acquittal on an indictment for murder will be a good bar to an indictment for manslaughter, and é converso, an acquittal on an indictment for manslaughter will be a bar to *534a prosecution for murder; for in the first instance, had the defendant been guilty, not of murder but of manslaughter, he would have been found guilty of the latter offence upon that indictment; and in the second instance, since the defendant is not guilty of manslaughter, he cannot be guilty of manslaughter under circumstances of aggravation which enlarge it into murder.” Id., at 504 (emphasis in original).
Unless one offense is lesser included of the other, the two are not the “same” under this test.
In State v. Standifer, 5 Porter 523 (Ala. 1837), the defendants were acquitted of murdering Levi Lowry. They were then charged with assault and battery of John Lowry, and pleaded auterfoits acquit on the grounds that the charge was based on the same affray as the previous prosecution. The court rejected the plea: “It is not of unfrequent occurrence, that the same individual, at the same time, and in the same transaction, commits two or more distinct crimes, and an acquittal of one, will not be a bar to punishment for the other.” Id., at 531. A jury could not lawfully have returned a verdict of guilty of assault on John Lowry at the first trial, and the offenses thus had “no appearance of identity.” Id., at 532.
In State v. Sias, 17 N. H. 558 (1845), the defendant was first acquitted of larceny, and then charged with obtaining property by conspiracy. The State admitted that the “facts alleged and proposed to be proved in this case are precisely the same facts, and same obtaining of the same property as the facts and taking of property which constituted the larceny in the former indictment.” Ibid. The court held that the second prosecution was not barred:
“The offence charged in this indictment is not the same as that charged in the former, and of which the defendant has been acquitted; nor is it included in the former. The defendant could not have been convicted of a conspiracy on the former indictment. He cannot be con*535victed of larceny on this. The proof in the former case may have shown [the codefendant] to be guilty of larceny, and the defendant and others of a conspiracy, but the acquittal was of the larceny charged, and not of the conspiracy, which was not charged; and of which, for that reason, the defendant could neither have been acquitted nor convicted in that case.” Id., at 559.
See also State v. Taylor, 2 Bailey 49, 50 (S. C. 1830) (conviction of “trading with a slave” does not bar prosecution for receiving goods stolen by slave “founded on the same act”; “two distinct offences were committed” because neither offense was necessarily included within the other); Hite v. State, 17 Tenn. 357, 376 (1836) (following Vandercomb)’, State v. Glasgow, Dudley 40, 43 (S. C. 1837) (following Sonnerkalb); State v. Coombs, 32 Maine 529, 530 (1851) (conviction for selling liquor does not bar prosecution for being a common seller of such liquors: “In the trial for common selling, the single acts of sale are not prosecuted. They are shown merely as evidence of the larger crime. Such proceedings do not expose to a second punishment for the same offence”); Wilson v. State, 24 Conn. 57, 63 (1855) (conviction for larceny does not bar prosecution for burglary by breaking and entering with intent to steal because each offense requires proof of facts that other does not: “A uniform doctrine on this point has prevailed, wherever it has been discussed”); State v. Warner, 14 Ind. 572 (1860) (same rule).
Thus, the Blockburger definition of “same offence” was not invented in 1932, but reflected a venerable understanding. Blockburger relied on Gavieres v. United States, 220 U. S. 338, 343 (1911), which relied on Burton v. United States, 202 U. S. 344, 380-381 (1906), which relied on Commonwealth v. Roby, swpra, one of the leading early cases. Blockburger and Gavieres also cited Morey v. Commonwealth, 108 Mass. 433, 435 (1871), which also applied Roby. We have applied the Roby-Morey-Gavieres-Blockburger formulation in virtually every case defining the “same offense” decided since *536Blockburger. See, e. g., Jones v. Thomas, 491 U. S. 376, 384-385, n. 3 (1989); United States v. Woodward, 469 U. S. 105, 108 (1985) (per curiam); Ohio v. Johnson, 467 U. S. 493, 499, n. 8 (1984); Albernaz v. United States, 450 U. S. 333, 337 (1981); Whalen v. United States, 445 U. S. 684, 691 (1980); Simpson v. United States, 435 U. S. 6, 11 (1978); Iannelli v. United States, 420 U. S., at 785, n. 17; Gore v. United States, 357 U. S., at 392.
II
The Court today abandons text and longstanding precedent to adopt the theory that double jeopardy bars “any subsequent prosecution in which the government, to establish an essential element of an offense charged in that prosecution, will prove conduct that constitutes an offense for which the defendant has already been prosecuted.” Ante, at 521 (emphasis added). The Court purports to derive that standard from our decision in Illinois v. Vitale, 447 U. S. 410 (1980), in which a motorist who caused a fatal accident was first convicted of unlawful failure to reduce speed, and later charged with involuntary manslaughter. We reversed the lower court’s determination that the second prosecution was barred by the Blockburger test, because each statute had a statutory element that the other did not: Manslaughter, but not failure to reduce speed, required proof of death; failure to reduce speed, but not manslaughter, required a failure to slow down. In remanding, however, we noted the possibility that the second prosecution might be barred on another ground:
“[I]t may be that to sustain its manslaughter case the State may find it necessary to prove a failure to slow or to rely on conduct necessarily involving such failure; it may concede as much prior to trial. In that case, because Vitale has already been convicted for conduct that is a necessary element of the more serious crime for which he has been charged, his claim of double jeopardy *537would be substantial under Brawn [v. Ohio, 432 U. S. 161 (1977),] and our later decision in Harris v. Oklahoma, 433 U. S. 682 (1977).” 447 U. S., at 420 (emphasis added).
We did not decide in Vitale that the second prosecution would constitute double jeopardy if it required proof of the conduct for which Vitale had already been convicted. We could not possibly have decided that, since the issue was not presented on the facts before us. But beyond that, we did not even say in Vitale, by way of dictum, that such a prosecution would violate the Double Jeopardy Clause. We said only that a claim to that effect would be “substantial,” ibid.; see also id. at 421, deferring to another day the question whether it would be successful. That day is today, and we should answer the question no.
To begin with, the argument that Vitale said to be “substantial” finds no support whatever in the two cases that Vitale thought gave it substance, Brown v. Ohio, 432 U. S. 161 (1977), and Harris v. Oklahoma, 433 U. S. 682 (1977). The first, Brown, involved nothing more than a straightforward application of Blockburger. There a car thief was first convicted of “joyriding,” an offense that consisted of “tak[ing], operat[ing], or keeping] any motor vehicle without the consent of its owner.” 432 U. S., at 162, n. 1. He was then charged with auto theft, which required all the elements of joyriding plus an intent permanently to deprive the owner of his car. We held that Blockburger barred the second prosecution: Because joyriding was simply a lesser included offense of auto theft, proof of the latter would “invariably” require proof of the former. 432 U. S., at 168. We did not even hint that double jeopardy would also have barred the prosecution if the two statutes had passed the Blockburger test but the second prosecution could not be successful without proving the same facts. The second case, our brief per curiam disposition in Harris, involved a prosecution for armed robbery that followed a conviction for felony murder *538based on the same armed robbery. The felony murder statute by definition incorporated all of the elements of the underlying felony charged; thus the later prosecution (rather than, as in Brown, the earlier conviction) involved a lesser included offense. “When,” we said, “conviction of a greater crime, murder, cannot be had without conviction of the lesser crime, robbery with firearms, the Double Jeopardy Clause bars prosecution for the lesser crime after conviction of the greater one.” 433 U. S., at 682. Again, we gave no indication that the second prosecution would have been barred if— not because of the statutory definition of the crimes but merely because of the circumstances of the particular case— guilt could not be established without proving the same conduct charged in the first prosecution. In short, to call the latter proposition “substantial” in Vitale took more than a little stretching of the cited cases.
I would have thought the result the Court reaches today foreclosed by our decision just a few months ago in Dowling v. United States, 493 U. S. 342 (1990). There the State, in a prosecution for robbery, introduced evidence of the defendant’s perpetration of another robbery committed in similar fashion (both involved ski masks), of which he had previously been acquitted. Proof of the prior robbery tended to establish commission of the later one. The State, in other words, “to establish an essential element of an offense charged in [the second] prosecution, [had] prove[d] conduct that constitute[d] an offense for which the defendant ha[d] already been prosecuted.” Ante, at 521. We held, however, that the Double Jeopardy Clause was not violated. The difference in our holding today cannot rationally be explained by the fact that in Dowling, unlike the present case, the two crimes were part of separate transactions; that in no way alters the central vice (according to today’s holding) that the defendant was forced a second time to defend against proof that he had committed a robbery for which he had already been prosecuted. In Dowling, as here, conduct establishing a previ*539ously prosecuted offense was relied upon, not because that offense was a statutory element of the second offense, but only because the conduct would prove the existence of a statutory element. If that did not offend the Double Jeopardy Clause in Dowling, it should not do so here.
The principle the Court adopts today is not only radically out of line with our double jeopardy jurisprudence; its practical effect, whenever it applies, will come down to a requirement that where the charges arise from a “‘single criminal act, occurrence, episode, or transaction,’” they “must be tried in a single proceeding,” Brown v. Ohio, supra, at 170 (Brennan, J., concurring) — a requirement we have hitherto “steadfastly refused” to impose, Garrett v. United States, 471 U. S. 773, 790 (1985). Suppose, for example, that the State prosecutes a group of individuals for a substantive offense, and then prosecutes them for conspiracy. Cf. Pinkerton v. United States, 328 U. S. 640, 645-646 (1946). In the conspiracy trial it will prove (if it can) that the defendants actually committed the substantive offense — even though there is evidence of other overt acts sufficient to sustain the conspiracy charge. For proof of the substantive offense, though not an element of the conspiracy charge, will assuredly be persuasive in establishing that a conspiracy existed. Or suppose an initial prosecution for burglary and a subsequent prosecution for murder that occurred in the course of the same burglary. In the second trial the State will prove (if it can) that the defendant was engaged in a burglary — not because that is itself an element of the murder charge, but because by providing a motive for intentional killing it will be persuasive that murder occurred. Under the analysis embraced by the Court today, I take it that the second prosecution in each of these cases would be barred, because the State, “to establish an essential element of an offense charged in that prosecution, will prove conduct that constitutes an offense for which the defendant has already been prosecuted.” Ante, at 521. Just as, in today’s case, proof of *540drunken driving or of crossing the median strip invalidates the second prosecution even though they are not elements of the homicide and assault offenses of which respondent is charged; so also, in the hypotheticals given, proof of the substantive offense will invalidate the conspiracy prosecution and proof of the burglary the murder prosecution.
The Court seeks to shrink the apparent application of its novel principle by saying that repetitive proof violates the Double Jeopardy Clause only if it is introduced “to establish an essential element of an offense charged in [the second] prosecution.” That is a meaningless limitation, of course. All evidence pertaining to guilt seeks “to establish an essential element of [the] offense,” and should be excluded if it does not have that tendency.
The other half of the Court’s new test does seem to import some limitation, though I am not sure precisely what it means and cannot imagine what principle justifies it. I refer to the requirement that the evidence introduced in the second prosecution must “prove conduct that constitutes an offense for which the defendant has already been prosecuted.” This means, presumably, that prosecutors who wish to use facts sufficient to prove one crime in order to establish guilt of another crime must bring both prosecutions simultaneously; but that those who wish to use only some of the facts establishing one crime — not enough facts to “prove conduct that constitutes an offense” — can bring successive prosecutions. But, one may reasonably ask, what justification is there even in reason alone (having abandoned text and precedent) for limiting the Court’s new rule in this fashion? The Court defends the rule on the ground that a successive prosecution based on the same proof exposes the defendant to the burden and embarrassment of resisting proof of the same facts in multiple proceedings, and enables the State to “rehearse its presentation of proof, thus increasing the risk of an .erroneous conviction for one or more of the offenses charged.” Ante, at 518. But that vice does not exist only when the sec*541ond prosecution seeks to prove all the facts necessary to support the first prosecution; it exists as well when the second prosecution seeks to prove some, rather than all of them— i. e., whenever two prosecutions each require proof of facts (or even a single fact) common to both. If the Court were correct that the Double Jeopardy Clause protects individuals against the necessity of twice proving (or refuting) the same evidence, as opposed to the necessity of twice defending against the same charge, then the second prosecution should be equally bad whether it contains all or merely some of the proof necessary for the first.
Apart from the lack of rational basis for this latter limitation, I am greatly perplexed (as will be the unfortunate trial court judges who must apply today’s rootless decision) as to what precisely it means. It is not at all apparent how a court is to go about deciding whether the evidence that has been introduced (or that will be introduced) at the second trial “proves conduct” that constitutes an offense for which the defendant has already been prosecuted. Is the judge in the second trial supposed to pretend that he is the judge in the first one, and to let the second trial proceed only if the evidence would not be enough to go to the jury on the earlier charge? Or (as the language of the Court’s test more readily suggests) is the judge in the second trial supposed to decide on his own whether the evidence before him really “proves” the earlier charge (perhaps beyond a reasonable doubt)? Consider application of the Court’s new rule in the unusually simple circumstances of the present case: Suppose that, in the trial upon remand, the prosecution’s evidence shows, among other things, that when the vehicles came to rest after the collision they were located on what was, for the defendant’s vehicle, the wrong side of the road. The prosecution also produces a witness who testifies that prior to the collision the defendant’s vehicle was “weaving back and forth”— without saying, however, that it was weaving back and forth over the center line. Is this enough to meet today’s require*542ment of “proving” the offense of operating a vehicle on the wrong side of the road? If not, suppose in addition that defense counsel asks the witness on cross-examination, “When you said the defendant’s vehicle was ‘weaving back and forth,’ did you mean weaving back and forth across the center line?” — to which the witness replies “yes.” Will this self-inflicted wound count for purposes of determining what the prosecution has “proved”? If so, can the prosecution then seek to impeach its own witness by showing that his recollection of the vehicle’s crossing the center line was inaccurate? Or can it at least introduce another witness to establish that fact? There are many questions here, and the answers to all of them are ridiculous. Whatever line is selected as the criterion of “proving” the prior offense — enough evidence to go to the jury, more likely than not, or beyond a reasonable doubt — the prosecutor in the second trial will presumably seek to introduce as much evidence as he can without crossing that line; and the defense attorney will presumably seek to provoke the prosecutor into (or assist him in) proving the defendant guilty of the earlier crime. This delicious role reversal, discovered to have been mandated by the Double Jeopardy Clause lo these 200 years, makes for high comedy but inferior justice. Often, the performance will even have an encore. If the judge initially decides that the previously prosecuted offense “will not be proved” (whatever that means) he will have to decide at the conclusion of the trial whether it “has been proved” (whatever that means). Indeed, he may presumably be asked to make the latter determination periodically during the course of the trial, since the Double Jeopardy Clause assuredly entitles the defendant to have the proceedings terminated as soon as its violation is evident. Even if we had no constitutional text and no prior case law to rely upon, rejection of today’s opinion is adequately supported by the modest desire to protect our criminal legal system from ridicule.
*543A limitation that is so unsupported in reason and so absurd in application is unlikely to survive. Today’s decision to extend the Double Jeopardy Clause to prosecutions that prove a previously prosecuted offense will lead predictably to extending it to prosecutions that involve the same facts as a previously prosecuted offense. We will thus have fully embraced Justice Brennan’s “same transaction” theory, which has as little support in the text and history of the Double Jeopardy Clause, but at least has the merit of being rational and easy to apply. One can readily imagine the words of our first opinion effecting this extension: “When we said in Grady that the second prosecution is impermissible if it ‘will prove conduct’ that constitutes the prior offense, we did not mean that it will establish commission of that offense with the degree of completeness that would permit a jury to convict. It suffices if the evidence in the second prosecution ‘proves’ the previously prosecuted offense in the sense of tending to establish one or more of the elements of that offense.” The Court that has done what it has today to 200 years of established double jeopardy jurisprudence should find this lesser transmogrification easy. It may, however, prove unnecessary, since prosecutors confronted with the inscrutability of today’s opinion will be well advised to proceed on the assumption that the “same transaction” theory has already been adopted. It is hard to tell what else has.
Ill
Since I do not agree with the Court’s new theory of the Double Jeopardy Clause, the question in this case for me is whether the current prosecution will place respondent in jeopardy for the “same offenses” for which he has already been convicted. The elements of the traffic offenses to which he pleaded guilty were, respectively, operating a vehicle on the wrong side of the road, N. Y. Veh. & Traf. Law § 1120(a) (McKinney 1986), and operating a vehicle while in an intoxicated condition, § 1192(3). The elements of the of*544fenses covered by the subsequent charges whose dismissal is challenged here* are, respectively, recklessly causing the death of another person, N. Y. Penal Law § 125.15 (McKinney 1987), negligently causing the death of another person, §125.10, and recklessly causing physical injury to another person, § 120.00.' Because respondent concedes, see ante, at 522, that each of these provisions contains an element, in the sense described by Blockburger, that the provisions under which he has been convicted do not, they do not constitute the “same offense” within the meaning of the Double Jeopardy Clause. I would therefore reverse the judgment.
The court below held two vehicular manslaughter counts barred under the Blockburger test, and because the State does not contest that ruling here, see ante, at 521, n. 12, I do not reach it.