Rogers v. Tennessee , 121 S. Ct. 1693 ( 2001 )

532 U.S. 451 (2001)

ROGERS
v.
TENNESSEE

No. 99-6218.

United States Supreme Court.

Argued November 1, 2000. Decided May 14, 2001. CERTIORARI TO THE SUPREME COURT OF TENNESSEE

O'Connor, J., delivered the opinion of the Court, in which Rehnquist, C. J., and Kennedy, Souter, and Ginsburg, JJ., joined. Stevens, J., filed a dissenting opinion, post, p. 467. Scalia, J., filed a dissenting opinion, in which Stevens and Thomas, JJ., joined, and in which Breyer, J., joined as to Part II, post, p. 467. Breyer, J., filed a dissenting opinion, post, p. 481.

W. Mark Ward argued the cause for petitioner. With him on the briefs were Tony Brayton and Garland Ergüden.

Michael E. Moore, Solicitor General of Tennessee, argued the cause for respondent. With him on the brief were Paul G. Summers, Attorney General, and Gordon W. Smith, Associate Solicitor General.^[*]

Justice O'Connor, delivered the opinion of the Court.

This case concerns the constitutionality of the retroactive application of a judicial decision abolishing the common law "year and a day rule." At common law, the year and a day rule provided that no defendant could be convicted of murder unless his victim had died by the defendant's act within a year and a day of the act. See, e. g., Louisville, E. & St. L. R. Co. v. Clarke, 152 U.S. 230, 239 (1894); 4 W. Blackstone, Commentaries on the Laws of England 197-198 (1769). The Supreme Court of Tennessee abolished the rule as it had existed at common law in Tennessee and applied its decision to petitioner to uphold his conviction. The question before us is whether, in doing so, the court denied petitioner due process of law in violation of the Fourteenth Amendment.

I

Petitioner Wilbert K. Rogers was convicted in Tennessee state court of second degree murder. According to the undisputed facts, petitioner stabbed his victim, James Bowdery, with a butcher knife on May 6, 1994. One of the stab wounds penetrated Bowdery's heart. During surgery to repair the wound to his heart, Bowdery went into cardiac arrest, but was resuscitated and survived the procedure. As a result, however, he had developed a condition known as "cerebral hypoxia," which results from a loss of oxygen to the brain. Bowdery's higher brain functions had ceased, and he slipped into and remained in a coma until August 7, 1995, when he died from a kidney infection (a common complication experienced by comatose patients). Approximately 15 months had passed between the stabbing and Bowdery's death which, according to the undisputed testimony of the county medical examiner, was caused by cerebral hypoxia "``secondary to a stab wound to the heart.' " 992 S.W.2d 393, 395 (Tenn. 1999).

Based on this evidence, the jury found petitioner guilty under Tennessee's criminal homicide statute. The statute, which makes no mention of the year and a day rule, defines criminal homicide simply as "the unlawful killing of another person which may be first degree murder, second degree murder, voluntary manslaughter, criminally negligent homicide or vehicular homicide." Tenn. Code Ann. § 39-13-201 (1997). Petitioner appealed his conviction to the Tennessee Court of Criminal Appeals, arguing that, despite its absence from the statute, the year and a day rule persisted as part of the common law of Tennessee and, as such, precluded his conviction. The Court of Criminal Appeals rejected that argument and affirmed the conviction. The court held that Tennessee's Criminal Sentencing Reform Act of 1989 (1989 Act), which abolished all common law defenses in criminal actions in Tennessee, had abolished the rule. See Tenn. Code Ann. § 39-11-203(e)(2) (1997). The court also rejected petitioner's further contention that the legislative abolition of the rule constituted an ex post facto violation, noting that the 1989 Act had taken effect five years before petitioner committed his crime. No. 02C01-9611—CR-00418 (Tenn. Crim. App., Oct. 17, 1997), App. 7.

The Supreme Court of Tennessee affirmed on different grounds. The court observed that it had recognized the viability of the year and a day rule in Tennessee in Percer v. State, 118 Tenn. 765, 103 S.W. 780 (1907), and that, "[d]espite the paucity of case law" on the rule in Tennessee, "both parties . . . agree that the . . . rule was a part of the common law of this State." 992 S.W.2d, at 396. Turning to the rule's present status, the court noted that the rule has been legislatively or judicially abolished by the "vast majority" of jurisdictions recently to have considered the issue. Id., at 397. The court concluded that, contrary to the conclusion of the Court of Criminal Appeals, the 1989 Act had not abolished the rule. After reviewing the justifications for the rule at common law, however, the court found that the original reasons for recognizing the rule no longer exist. Accordingly, the court abolished the rule as it had existed at common law in Tennessee. Id., at 399-401.

The court disagreed with petitioner's contention that application of its decision abolishing the rule to his case would violate the Ex Post Facto Clauses of the State and Federal Constitutions. Those constitutional provisions, the court observed, refer only to legislative Acts. The court then noted that in Bouie v. City of Columbia, 378 U.S. 347 (1964), this Court held that due process prohibits retroactive application of any "``judicial construction of a criminal statute [that] is unexpected and indefensible by reference to the law which has been expressed prior to the conduct in issue.' " 992 S.W.2d, at 402 (quoting Bouie v. City of Columbia, supra, at 354) (alteration in original). The court concluded, however, that application of its decision to petitioner would not offend this principle. 992 S.W.2d, at 402. We granted certiorari, 529 U.S. 1129 (2000), and we now affirm.

II

Although petitioner's claim is one of due process, the Constitution's Ex Post Facto Clause figures prominently in his argument. The Clause provides simply that "[n]o State shall . . . pass any . . . ex post facto Law." Art. I, § 10, cl. 1. The most well-known and oft-repeated explanation of the scope of the Clause's protection was given by Justice Chase, who long ago identified, in dictum, four types of laws to which the Clause extends:

"1st. Every law that makes an action done before the passing of the law, and which was innocent when done, criminal; and punishes such action. 2d. Every law that aggravates a crime, or makes it greater than it was, when committed. 3d. Every law that changes the punishment, and inflicts a greater punishment, than the law annexed to the crime, when committed. 4th. Every law that alters the legal rules of evidence, and receives less, or different, testimony, than the law required at the time of the commission of the offense, in order to convict the offender." Calder v. Bull, 3 Dall. 386, 390 (1798) (seriatim opinion of Chase, J.) (emphasis deleted).

Accord, Carmell v. Texas, 529 U.S. 513, 521-525 (2000); Collins v. Youngblood, 497 U.S. 37, 41-42, 46 (1990). As the text of the Clause makes clear, it "is a limitation upon the powers of the Legislature, and does not of its own force apply to the Judicial Branch of government." Marks v. United States, 430 U.S. 188, 191 (1977) (citation omitted).

We have observed, however, that limitations on ex post facto judicial decision making are inherent in the notion of due process. In Bouie v. City of Columbia, we considered the South Carolina Supreme Court's retroactive application of its construction of the State's criminal trespass statute to the petitioners in that case. The statute prohibited "entry upon the lands of another . . . after notice from the owner or tenant prohibiting such entry . . . ." 378 U.S., at 349, n. 1 (citation and internal quotation marks omitted). The South Carolina court construed the statute to extend to patrons of a drug store who had received no notice prohibiting their entry into the store, but had refused to leave the store when asked. Prior to the court's decision, South Carolina cases construing the statute had uniformly held that conviction under the statute required proof of notice before entry. None of those cases, moreover, had given the "slightest indication that that requirement could be satisfied by proof of the different act of remaining on the land after being told to leave." Id., at 357.

We held that the South Carolina court's retroactive application of its construction to the store patrons violated due process. Reviewing decisions in which we had held criminal statutes "void for vagueness" under the Due Process Clause, we noted that this Court has often recognized the "basic principle that a criminal statute must give fair warning of the conduct that it makes a crime." Id., at 350; see id., at 350-352 (discussing, inter alia, United States v. Harriss, 347 U.S. 612 (1954), Lanzetta v. New Jersey, 306 U.S. 451 (1939), and Connally v. General Constr. Co., 269 U.S. 385 (1926)). Deprivation of the right to fair warning, we continued, can result both from vague statutory language and from an unforeseeable and retroactive judicial expansion of statutory language that appears narrow and precise on its face. Bouie v. City of Columbia, 378 U. S., at 352. For that reason, we concluded that "[i]f a judicial construction of a criminal statute is ``unexpected and indefensible by reference to the law which had been expressed prior to the conduct in issue,' [the construction] must not be given retroactive effect." Id., at 354 (quoting J. Hall, General Principles of Criminal Law 61 (2d ed. 1960)). We found that the South Carolina court's construction of the statute violated this principle because it was so clearly at odds with the statute's plain language and had no support in prior South Carolina decisions. 378 U.S., at 356.

Relying largely upon Bouie, petitioner argues that the Tennessee court erred in rejecting his claim that the retroactive application of its decision to his case violates due process. Petitioner contends that the Ex Post Facto Clause would prohibit the retroactive application of a decision abolishing the year and a day rule if accomplished by the Tennessee Legislature. He claims that the purposes behind the Clause are so fundamental that due process should prevent the Supreme Court of Tennessee from accomplishing the same result by judicial decree. Brief for Petitioner 8-18. In support of this claim, petitioner takes Bouie to stand for the proposition that "[i]n evaluating whether the retroactive application of a judicial decree violates Due Process, a critical question is whether the Constitution would prohibit the same result attained by the exercise of the state's legislative power." Brief for Petitioner 12.

To the extent petitioner argues that the Due Process Clause incorporates the specific prohibitions of the Ex Post Facto Clause as identified in Calder, petitioner misreads Bouie. To be sure, our opinion in Bouie does contain some expansive language that is suggestive of the broad interpretation for which petitioner argues. Most prominent is our statement that "[i]f a state legislature is barred by the Ex Post Facto Clause from passing . . . a law, it must follow that a State Supreme Court is barred by the Due Process Clause from achieving precisely the same result by judicial construction." 378 U.S., at 353-354; see also id., at 353 ("[A]n unforeseeable judicial enlargement of a criminal statute, applied retroactively, operates precisely like an ex post facto law"); id., at 362 ("The Due Process Clause compels the same result" as would the constitutional proscription against ex post facto laws "where the State has sought to achieve precisely the same [impermissible] effect by judicial construction of the statute"). This language, however, was dicta. Our decision in Bouie was rooted firmly in well established notions of due process. See supra, at 457-458. Its rationale rested on core due process concepts of notice, foreseeability, and, in particular, the right to fair warning as those concepts bear on the constitutionality of attaching criminal penalties to what previously had been innocent conduct. See, e. g., 378 U. S., at 351, 352, 354-355. And we couched its holding squarely in terms of that established due process right, and not in terms of the ex post facto related dicta to which petitioner points. Id., at 355 (concluding that "the South Carolina Code did not give [the petitioners] fair warning, at the time of their conduct . . . , that the act for which they now stand convicted was rendered criminal by the statute"). Contrary to petitioner's suggestion, nowhere in the opinion did we go so far as to incorporate jot-for-jot the specific categories of Calder into due process limitations on the retroactive application of judicial decisions.

Nor have any of our subsequent decisions addressing Bouie -type claims interpreted Bouie as extending so far. Those decisions instead have uniformly viewed Bouie as restricted to its traditional due process roots. In doing so, they have applied Bouie ``s check on retroactive judicial decision making not by reference to the ex post facto categories set out in Calder, but, rather, in accordance with the more basic and general principle of fair warning that Bouie so clearly articulated. See, e. g., United States v. Lanier, 520 U.S. 259, 266 (1997) ("[D]ue process bars courts from applying a novel construction of a criminal statute to conduct that neither the statute nor any prior judicial decision has fairly disclosed to be within its scope"); Marks v. United States, 430 U. S., at 191-192 (Due process protects against judicial infringement of the "right to fair warning" that certain conduct will give rise to criminal penalties); Rose v. Locke, 423 U.S. 48, 53 (1975) (per curiam) (upholding defendant's conviction under statute prohibiting "crimes against nature" because, unlike in Bouie, the defendant "[could] make no claim that [the statute] afforded no notice that his conduct might be within its scope"); Douglas v. Buder, 412 U.S. 430, 432 (1973) (per curiam) (trial court's construction of the term "arrest" as including a traffic citation, and application of that construction to defendant to revoke his probation, was unforeseeable and thus violated due process); Rabe v. Washington, 405 U.S. 313, 316 (1972) (per curiam) (reversing conviction under state obscenity law because it did "not giv[e] fair notice" that the location of the allegedly obscene exhibition was a vital element of the offense).

Petitioner observes that the Due Process and Ex Post Facto Clauses safeguard common interests—in particular, the interests in fundamental fairness (through notice and fair warning) and the prevention of the arbitrary and vindictive use of the laws. Brief for Petitioner 12-18. While this is undoubtedly correct, see, e. g., Lynce v. Mathis, 519 U.S. 433, 439-440, and n. 12 (1997), petitioner is mistaken to suggest that these considerations compel extending the strictures of the Ex Post Facto Clause to the context of common law judging. The Ex Post Facto Clause, by its own terms, does not apply to courts. Extending the Clause to courts through the rubric of due process thus would circumvent the clear constitutional text. It also would evince too little regard for the important institutional and contextual differences between legislating, on the one hand, and common law decisionmaking, on the other.

Petitioner contends that state courts acting in their common law capacity act much like legislatures in the exercise of their lawmaking function, and indeed may in some cases even be subject to the same kinds of political influences and pressures that justify ex post facto limitations upon legislatures. Brief for Petitioner 12-18; Reply Brief for Petitioner 15. A court's "opportunity for discrimination," however, "is more limited than [a] legislature's, in that [it] can only act in construing existing law in actual litigation." James v. United States, 366 U.S. 213, 247, n. 3 (1961) (Harlan, J., concurring in part and dissenting in part). Moreover, "[g]iven the divergent pulls of flexibility and precedent in our case law system," ibid., incorporation of the Calder categories into due process limitations on judicial decision making would place an unworkable and unacceptable restraint on normal judicial processes and would be incompatible with the resolution of uncertainty that marks any evolving legal system.

That is particularly so where, as here, the allegedly impermissible judicial application of a rule of law involves not the interpretation of a statute but an act of common law judging. In the context of common law doctrines (such as the year and a day rule), there often arises a need to clarify or even to reevaluate prior opinions as new circumstances and fact patterns present themselves. Such judicial acts, whether they be characterized as "making" or "finding" the law, are a necessary part of the judicial business in States in which the criminal law retains some of its common law elements. Strict application of ex post facto principles in that context would unduly impair the incremental and reasoned development of precedent that is the foundation of the common law system. The common law, in short, presupposes a measure of evolution that is incompatible with stringent application of ex post facto principles. It was on account of concerns such as these that Bouie restricted due process limitations on the retroactive application of judicial interpretations of criminal statutes to those that are "unexpected and indefensible by reference to the law which had been expressed prior to the conduct in issue." Bouie v. City of Columbia, 378 U. S., at 354 (internal quotation marks omitted).

We believe this limitation adequately serves the common law context as well. It accords common law courts the substantial leeway they must enjoy as they engage in the daily task of formulating and passing upon criminal defenses and interpreting such doctrines as causation and intent, reevaluating and refining them as may be necessary to bring the common law into conformity with logic and common sense. It also adequately respects the due process concern with fundamental fairness and protects against vindictive or arbitrary judicial lawmaking by safeguarding defendants against unjustified and unpredictable breaks with prior law. Accordingly, we conclude that a judicial alteration of a common law doctrine of criminal law violates the principle of fair warning, and hence must not be given retroactive effect, only where it is "unexpected and indefensible by reference to the law which had been expressed prior to the conduct in issue." Ibid.

Justice Scalia makes much of the fact that, at the time of the framing of the Constitution, it was widely accepted that courts could not "change" the law, see post, at 472-473, 477-478 (dissenting opinion), and that (according to Justice Scalia) there is no doubt that the Ex Post Facto Clause would have prohibited a legislative decision identical to the Tennessee court's decision here, see post, at 469-471, 478. This latter argument seeks at bottom merely to reopen what has long been settled by the constitutional text and our own decisions: that the Ex Post Facto Clause does not apply to judicial decision making. The former argument is beside the point. Common law courts at the time of the framing undoubtedly believed that they were finding rather than making law. But, however one characterizes their actions, the fact of the matter is that common law courts then, as now, were deciding cases, and in doing so were fashioning and refining the law as it then existed in light of reason and experience. Due process clearly did not prohibit this process of judicial evolution at the time of the framing, and it does not do so today.

III

Turning to the particular facts of the instant case, the Tennessee court's abolition of the year and a day rule was not unexpected and indefensible. The year and a day rule is widely viewed as an outdated relic of the common law. Petitioner does not even so much as hint that good reasons exist for retaining the rule, and so we need not delve too deeply into the rule and its history here. Suffice it to say that the rule is generally believed to date back to the 13th century, when it served as a statute of limitations governing the time in which an individual might initiate a private action for murder known as an "appeal of death"; that by the 18th century the rule had been extended to the law governing public prosecutions for murder; that the primary and most frequently cited justification for the rule is that 13th century medical science was incapable of establishing causation beyond a reasonable doubt when a great deal of time had elapsed between the injury to the victim and his death; and that, as practically every court recently to have considered the rule has noted, advances in medical and related science have so undermined the usefulness of the rule as to render it without question obsolete. See, e. g., People v. Carrillo, 164 Ill. 2d 144, 150, 646 N.E.2d 582, 585 (1995); Commonwealth v. Lewis, 381 Mass. 411, 414-415, 409 N.E.2d 771, 772-773 (1980); People v. Stevenson, 416 Mich. 383, 391-392, 331 N.W.2d 143, 146 (1982); State v. Hefler, 310 N. C. 135, 138-140, 310 S.E.2d 310, 313 (1984); see generally Comment, 59 U. Chi. L. Rev. 1337 (1992) (tracing the history of the rule).

For this reason, the year and a day rule has been legislatively or judicially abolished in the vast majority of jurisdictions recently to have addressed the issue. See 992 S.W.2d, at 397, n. 4 (reviewing cases and statutes). Citing Bouie, petitioner contends that the judicial abolition of the rule in other jurisdictions is irrelevant to whether he had fair warning that the rule in Tennessee might similarly be abolished and, hence, to whether the Tennessee court's decision was unexpected and indefensible as applied to him. Brief for Petitioner 28-30. In discussing the apparent meaning of the South Carolina statute in Bouie, we noted that "[i]t would be a rare situation in which the meaning of a statute of another State sufficed to afford a person ``fair warning' that his own State's statute meant something quite different from what its words said." 378 U.S., at 359-360. This case, however, involves not the precise meaning of the words of a particular statute, but rather the continuing viability of a common law rule. Common law courts frequently look to the decisions of other jurisdictions in determining whether to alter or modify a common law rule in light of changed circumstances, increased knowledge, and general logic and experience. Due process, of course, does not require a person to apprise himself of the common law of all 50 States in order to guarantee that his actions will not subject him to punishment in light of a developing trend in the law that has not yet made its way to his State. At the same time, however, the fact that a vast number of jurisdictions have abolished a rule that has so clearly outlived its purpose is surely relevant to whether the abolition of the rule in a particular case can be said to be unexpected and indefensible by reference to the law as it then existed.

Finally, and perhaps most importantly, at the time of petitioner's crime the year and a day rule had only the most tenuous foothold as part of the criminal law of the State of Tennessee. The rule did not exist as part of Tennessee's statutory criminal code. And while the Supreme Court of Tennessee concluded that the rule persisted at common law, it also pointedly observed that the rule had never once served as a ground of decision in any prosecution for murder in the State. Indeed, in all the reported Tennessee cases, the rule has been mentioned only three times, and each time in dicta.

The first mention of the rule in Tennessee, and the only mention of it by the Supreme Court of that State, was in 1907 in Percer v. State, 118 Tenn. 765, 103 S.W. 780. In Percer, the court reversed the defendant's conviction for second degree murder because the defendant was not present in court when the verdict was announced and because the proof failed to show that the murder occurred prior to the finding of the indictment. In discussing the latter ground for its decision, the court quoted the rule that "``it is . . . for the State to show that the crime was committed before the indictment was found, and, where it fails to do so, a conviction will be reversed.' " Id., at 777, 103 S.W., at 783 (quoting 12 Cyclopedia of Law and Procedure 382 (1904)). The court then also quoted the rule that "``[i]n murder, the death must be proven to have taken place within a year and a day from the date of the injury received.' " 118 Tenn., at 777, 103 S.W., at 783 (quoting F. Wharton, Law of Homicide § 18 (3d ed. 1907)).

While petitioner relies on this case for the proposition that the year and a day rule was firmly entrenched in the common law of Tennessee, we agree with the Supreme Court of Tennessee that the case cannot establish nearly so much. After reciting the rules just mentioned, the court in Percer went on to point out that the indictment was found on July 6, 1906; that it charged that the murder was committed sometime in May 1906; and that the only evidence of when the victim died was testimony from a witness stating that he thought the death occurred sometime in July, but specifying neither a date nor a year. From this, the court concluded that it did "not affirmatively appear" from the evidence "whether the death occurred before or after the finding of the indictment." 118 Tenn., at 777, 103 S.W., at 783. The court made no mention of the year and a day rule anywhere in its legal analysis or, for that matter, anywhere else in its opinion. Thus, whatever the import of the court's earlier quoting of the rule, it is clear that the rule did not serve as the basis for the Percer court's decision.

The next two references to the rule both were by the Tennessee Court of Criminal Appeals in cases in which the date of the victim's death was not even in issue. Sixty-seven years after Percer, the court in Cole v. State, 512 S.W.2d 598 (Tenn. Crim. App. 1974), noted the existence of the rule in rejecting the defendants' contentions that insufficient evidence existed to support the jury's conclusion that they had caused the victim's death in a drag-racing crash. Id., at 601. Twenty-one years after that, in State v. Ruane, 912 S.W.2d 766 (Tenn. Crim. App. 1995), a defendant referred to the rule in arguing that the operative cause of his victim's death was removal of life support rather than a gunshot wound at the defendant's hand. The victim had died within 10 days of receiving the wound. The Court of Criminal Appeals rejected the defendant's argument, concluding, as it had in this case, that the year and a day rule had been abolished by the 1989 Act. It went on to hold that the evidence of causation was sufficient to support the conviction. Id., at 773-777. Ruane, of course, was decided after petitioner committed his crime, and it concluded that the year and a day rule no longer existed in Tennessee for a reason that the high court of that State ultimately rejected. But we note the case nonetheless to complete our account of the few appearances of the common law rule in the decisions of the Tennessee courts.

These cases hardly suggest that the Tennessee court's decision was "unexpected and indefensible" such that it offended the due process principle of fair warning articulated in Bouie and its progeny. This is so despite the fact that, as Justice Scalia correctly points out, the court viewed the year and a day rule as a "substantive principle" of the common law of Tennessee. See post, at 480. As such, however, it was a principle in name only, having never once been enforced in the State. The Supreme Court of Tennessee also emphasized this fact in its opinion, see 992 S.W.2d, at 402, and rightly so, for it is surely relevant to whether the court's abolition of the rule in petitioner's case violated due process limitations on retroactive judicial decision making. And while we readily agree with Justice Scalia that fundamental due process prohibits the punishment of conduct that cannot fairly be said to have been criminal at the time the conduct occurred, see, e. g., post, at 470, 478, 480, nothing suggests that is what took place here.

There is, in short, nothing to indicate that the Tennessee court's abolition of the rule in petitioner's case represented an exercise of the sort of unfair and arbitrary judicial action against which the Due Process Clause aims to protect. Far from a marked and unpredictable departure from prior precedent, the court's decision was a routine exercise of common law decision making in which the court brought the law into conformity with reason and common sense. It did so by laying to rest an archaic and outdated rule that had never been relied upon as a ground of decision in any reported Tennessee case.

The judgment of the Supreme Court of Tennessee is accordingly affirmed.

It is so ordered.

Justice Stevens, dissenting.

While I have joined Justice Scalia's entire dissent, I must add this brief caveat. The perception that commonlaw judges had no power to change the law was unquestionably an important aspect of our judicial heritage in the 17th century but, as he has explained, that perception has played a role of diminishing importance in later years. Whether the most significant changes in that perception occurred before the end of the 18th century or early in the 19th century is, in my judgment, a tangential question that need not be resolved in order to decide this case correctly. For me, far more important than the historical issue is the fact that the majority has undervalued the threat to liberty that is posed whenever the criminal law is changed retroactively.

Justice Scalia, with whom Justice Stevens and Justice Thomas join, and with whom Justice Breyer joins as to Part II, dissenting.

The Court today approves the conviction of a man for a murder that was not murder (but only manslaughter) when the offense was committed. It thus violates a principle— encapsulated in the maxim nulla poena sine lege —which "dates from the ancient Greeks" and has been described as one of the most "widely held value-judgment[s] in the entire history of human thought." J. Hall, General Principles of Criminal Law 59 (2d ed. 1960). Today's opinion produces, moreover, a curious constitution that only a judge could love. One in which (by virtue of the Ex Post Facto Clause) the elected representatives of all the people cannot retroactively make murder what was not murder when the act was committed; but in which unelected judges can do precisely that. One in which the predictability of parliamentary lawmaking cannot validate the retroactive creation of crimes, but the predictability of judicial lawmaking can do so. I do not believe this is the system that the Framers envisioned—or, for that matter, that any reasonable person would imagine.

I

A

To begin with, let us be clear that the law here was altered after the fact. Petitioner, whatever else he was guilty of, was innocent of murder under the law as it stood at the time of the stabbing, because the victim did not die until after a year and a day had passed. The requisite condition subsequent of the murder victim's death within a year and a day is no different from the requisite condition subsequent of the rape victim's raising a "hue and cry" which we held could not retroactively be eliminated in Carmell v. Texas, 529 U.S. 513 (2000). Here, as there, it operates to bar conviction. Indeed, if the present condition differs at all from the one involved in Carmell it is in the fact that it does not merely pertain to the "quantum of evidence" necessary to corroborate a charge, id. , at 530, but is an actual element of the crime—a "substantive principle of law," 992 S.W.2d 393, 399 (Tenn. 1999), the failure to establish which "entirely precludes a murder prosecution," id., at 400. Though the Court spends some time questioning whether the year-and-a-day rule was ever truly established in Tennessee, see ante, at 464-466, the Supreme Court of Tennessee said it was, see 992 S.W.2d, at 396, 400, and this reasonable reading of state law by the State's highest court is binding upon us.

Petitioner's claim is that his conviction violated the Due Process Clause of the Fourteenth Amendment, insofar as that Clause contains the principle applied against the legislature by the Ex Post Facto Clause of Article I. We first discussed the relationship between these two Clauses in Bouie v. City of Columbia, 378 U.S. 347 (1964). There, we considered Justice Chase to have spoken for the Court in Calder v. Bull, 3 Dall. 386, 390 (1798), when he defined an ex post facto law as, inter alia, one that "aggravates a crime, or makes it greater than it was, when committed." 378 U.S., at 353 (emphasis deleted). We concluded that, "[i]f a state legislature is barred by the Ex Post Facto Clause from passing such a law, it must follow that a State Supreme Court is barred by the Due Process Clause from achieving precisely the same result by judicial construction." Id., at 353-354. The Court seeks to avoid the obvious import of this language by characterizing it as mere dicta. See ante, at 459. Only a concept of dictum that includes the very reasoning of the opinion could support this characterization. The ratio decidendi of Bouie was that the principle applied to the legislature though the Ex Post Facto Clause was contained in the Due Process Clause insofar as judicial action is concerned. I cannot understand why the Court derives such comfort from the fact that later opinions applying Bouie have referred to the Due Process Clause rather than the Ex Post Facto Clause, see ante, at 459-460; that is entirely in accord with the rationale of the case, which I follow and which the Court discards.

The Court attempts to cabin Bouie by reading it to prohibit only "``unexpected and indefensible' " judicial law revision, and to permit retroactive judicial changes so long as the defendant has had "fair warning" that the changes might occur. Ante, at 462. This reading seems plausible because Bouie does indeed use those quoted terms; but they have been wrenched entirely out of context. The "fair warning" to which Bouie and subsequent cases referred was not "fair warning that the law might be changed," but fair warning of what constituted the crime at the time of the offense. And Bouie did not express disapproval of "unexpected and indefensible changes in the law" (and thus implicitly approve "expected or defensible changes"). It expressed disapproval of "judicial construction of a criminal statute" that is "unexpected and indefensible by reference to the law which had been expressed prior to the conduct in issue. " 378 U.S., at 354 (emphasis added; internal quotation marks omitted). It thus implicitly approved only a judicial construction that was an expected or defensible application of prior cases interpreting the statute. Extending this principle from statutory crimes to common-law crimes would result in the approval of retroactive holdings that accord with prior cases expounding the common law, and the disapproval of retroactive holdings that clearly depart from prior cases expounding the common law. According to Bouie, not just "unexpected and indefensible" retroactive changes in the common law of crimes are bad, but all retroactive changes.

Bouie rested squarely upon "[t]he fundamental principle that ``the required criminal law must have existed when the conduct in issue occurred,' " ibid. (Nulla poena sine lege.) Proceeding from that principle, Bouie said that "a State Supreme Court is barred by the Due Process Clause from achieving precisely the same result [prohibited by the Ex Post Facto Clause] by judicial construction." Id., at 353— 354. There is no doubt that "fair warning" of the legislature's intent to change the law does not insulate retroactive legislative criminalization. Such a statute violates the Ex Post Facto Clause, no matter that, at the time the offense was committed, the bill enacting the change was pending and assured of passage—or indeed, had already been passed but not yet signed by the President whose administration had proposed it. It follows from the analysis of Bouie that "fair warning" of impending change cannot insulate retroactive judicial criminalization either.

Nor is there any reason in the nature of things why it should. According to the Court, the exception is necessary because prohibiting retroactive judicial criminalization would "place an unworkable and unacceptable restraint on normal judicial processes," would be "incompatible with the resolution of uncertainty that marks any evolving legal system," and would "unduly impair the incremental and reasoned development of precedent that is the foundation of the common law system." Ante, at 461. That assessment ignores the crucial difference between simply applying a law to a new set of circumstances and changing the law that has previously been applied to the very circumstances before the court. Many criminal cases present some factual nuance that arguably distinguishes them from cases that have come before; a court applying the penal statute to the new fact pattern does not purport to change the law. That, however, is not the action before us here, but rather, a square, head-on overruling of prior law—or, more accurately, something even more extreme than that: a judicial opinion acknowledging that under prior law, for reasons that used to be valid, the accused could not be convicted, but decreeing that, because of changed circumstances, "we hereby abolish the common law rule," 992 S.W.2d, at 401, and upholding the conviction by applying the new rule to conduct that occurred before the change in law was announced. Even in civil cases, and even in modern times, such retroactive revision of a concededly valid legal rule is extremely rare. With regard to criminal cases, I have no hesitation in affirming that it was unheard of at the time the original Due Process Clause was adopted. As I discuss in detail in the following section, proceeding in that fashion would have been regarded as contrary to the judicial traditions embraced within the concept of due process of law.

B

The Court's opinion considers the judgment at issue here "a routine exercise of common law decision making," whereby the Tennessee court "brought the law into conformity with reason and common sense," by "laying to rest an archaic and outdated rule." Ante, at 467. This is an accurate enough description of what modern "common law decision making" consists of—but it is not an accurate description of the theoretical model of common-law decision making accepted by those who adopted the Due Process Clause. At the time of the framing, common-law jurists believed (in the words of Sir Francis Bacon) that the judge's "office is jus dicere, and not jus dare; to interpret law, and not to make law, or give law." Bacon, Essays, Civil and Moral, in 3 Harvard Classics 130 (C. Eliot ed. 1909) (1625). Or as described by Blackstone, whose Commentaries were widely read and "accepted [by the framing generation] as the most satisfactory exposition of the common law of England," see Schick v. United States, 195 U.S. 65, 69 (1904), "judicial decisions are the principal and most authoritative evidence, that can be given, of the existence of such a custom as shall form a part of the common law," 1 W. Blackstone, Commentaries *69 (hereinafter Blackstone) (emphasis added).

Blackstone acknowledged that the courts' exposition of what the law was could change. Stare decisis, he said, "admits of exception, where the former determination is most evidently contrary to reason . . . ." Ibid. But "in such cases the subsequent judges do not pretend to make a new law, but to vindicate the old one from misrepresentation." Id., at *70. To fit within this category of bad law, a law must be "manifestly absurd or unjust." It would not suffice, he said, that "the particular reason [for the law] can at this distance of time [not be] precisely assigned." "[F]or though [its] reason be not obvious at first view, yet we owe such a deference to former times as not to suppose they acted wholly without consideration." Ibid.^[1] By way of example, Blackstone pointed to the seemingly unreasonable rule that one cannot inherit the estate of one's half brother. Though he accepted that the feudal reason behind the law was no longer obvious, he wrote "yet it is not in [a common law judge's] power to alter it." Id., at *70—*71 (emphasis added).^[2] Moreover, "the unreasonableness of a custom in modern circumstances will not affect its validity if the Court is satisfied of a reasonable origin." Allen 140-141. "A custome once reasonable and tolerable, if after it become grievous, and not answerable to the reason, whereupon it was grounded, yet is to be . . . taken away by act of parliament." 2 E. Coke, Institutes of the Laws of England *664 (hereinafter Institutes); see also id., at *97 ("No law, or custome of England can be taken away, abrogated, or adnulled, but by authority of parliament"); Of Oaths before an Ecclesiastical Judge Ex Officio, 12 Co. Rep. *26, *29 (1655) ("[T]he law and custom of England is the inheritance of the subject, which he cannot be deprived of without his assent in Parliament").

There are, of course, stray statements and doctrines found in the historical record that—read out of context—could be thought to support the modern-day proposition that the common law was always meant to evolve. Take, for instance, Lord Coke's statement in the Institutes that "the reason of the law ceasing, the law itself ceases." This maxim is often cited by modern devotees of a turbulently changing common law—often in its Latin form (cessante ratione legis, cessat ipse lex) to create the impression of great venerability. In its original context, however, it had nothing to do with the power of common-law courts to change the law. At the point at which it appears in the Institutes, Coke was discussing the exception granted abbots and mayors from the obligation of military service to the King which attached to land ownership. Such service would be impracticable for a man of the cloth or a mayor. But, said Coke, "if they convey over the lands to any naturall man and his heires," the immunity "by the conveyance over ceaseth." 1 Institutes *70. In other words, the service which attached to the land would apply to any subsequent owner not cloaked in a similar immunity. It was in describing this change that Coke employed the Latin maxim cessante ratione legis, cessat ipse lex. It had to do, not with a changing of the common-law rule, but with a change of circumstances that rendered the common-law rule no longer applicable to the case.

The same is true of the similar quotation from Coke: "[R]atio legis est anima legis, et mutata legis ratione, mutatur et lex"— reason is the soul of the law; the reason of the law being changed, the law is also changed. This is taken from Coke's report of Milborn's Case, 7 Co. Rep. 6b, 7a (1587), a suit involving a town's responsibility for a murder committed within its precincts. The common-law rule had been that a town could be amerced for failure to apprehend a murderer who committed his crime on its streets during the day, but not a murderer who struck after nightfall, when its citizens were presumably asleep. Parliament, however, enacted a statute requiring towns to close their gates at night, and the court reasoned that thereafter a town that left its gates open could be amerced for the nocturnal homicide as well, since the town's violation of the Act was negligence that facilitated the escape. This perhaps partakes more of a new right of action implied from legislation than of any common-law rule. But to the extent it involved the common law, it assuredly did not change the prior rule: A town not in violation of the statute would continue to be immune. Milborn's Case simply held that the rule would not be extended to towns that wrongfully failed to close their gates— which involves no overruling, but nothing more than normal, case-by-case common-law adjudication.

It is true that framing-era judges in this country considered themselves authorized to reject English common-law precedent they found "barbarous" and "ignorant," see 1 Z. Swift, A System of the Laws of the State of Connecticut 46 (1795) (hereinafter Swift); N. Chipman, A Dissertation on the Act Adopting the Common and Statute Laws of England, in Reports and Dissertations 117, 128 (1793) (hereinafter Chipman). That, however, was not an assertion of judges' power to change the common law. For, as Blackstone wrote, the common law was a law for England, and did not automatically transfer to the American Colonies; rather, it had to be adopted. See 1 Blackstone *107—*108 (observing that "the common law of England, as such, has no allowance or authority" in "[o]ur American plantations"); see also 1 Swift 46 ("The English common law is not in itself binding in this state"); id., at 44-45 ("The English common law has never been considered to be more obligatory here, than the Roman law has been in England"). In short, the colonial courts felt themselves perfectly free to pick and choose which parts of the English common law they would adopt.^[3] As stated by Chipman 128: "If no reason can be assigned, in support of rules, or precedents, not already adopted in practice, to adopt such rules, is certainly contrary to the principles of our government." (Emphasis added.) This discretion not to adopt would not presuppose, or even support, the power of colonial courts subsequently to change the accumulated colonial common law. The absence of belief in that power is demonstrated by the following passage from 1 Horwitz 5: "Massachusetts Chief Justice Hutchison could declare in 1767 that ``laws should be established, else Judges and Juries must go according to their Reason, that is, their Will.' It was also imperative ``that the Judge should never be the Legislator: Because, then the Will of the Judge would be the Law: and this tends to a State of Slavery.' " Or, as Judge Swift put it, courts "ought never to be allowed to depart from the well known boundaries of express law, into the wide fields of discretion." 2 Swift 366.

Nor is the framing era's acceptance of common-law crimes support for the proposition that the Framers accepted an evolving common law. The acknowledgment of a new crime, not thitherto rejected by judicial decision, was not a changing of the common law, but an application of it. At the time of the framing, common-law crimes were considered unobjectionable, for "``a law founded on the law of nature may be retrospective, because it always existed,' " 1 Horwitz 7, quoting Blackwell v. Wilkinson, Jefferson's Rep. 73, 77 (Va. 1768) (argument of then-Attorney General John Randolph). Of course, the notion of a common-law crime is utterly anathema today, which leads one to wonder why that is so. The obvious answer is that we now agree with the perceptive chief justice of Connecticut, who wrote in 1796 that common-law crimes "partak[e] of the odious nature of an ex post facto law." 2 Swift 365-366. But, as Horwitz makes clear, a widespread sharing of Swift's "preoccupation with the unfairness of administering a system of judge-made criminal law was a distinctly post-revolutionary phenomenon, reflecting a profound change in sensibility. For the inarticulate premise that lay behind Swift's warnings against the danger of judicial discretion was a growing perception that judges no longer merely discovered law; they also made it. " 1 Horwitz 14-15 (emphases added). In other words, the connection between ex post facto lawmaking and common-law judging would not have become widely apparent until common-law judging became lawmaking, not (as it had been) law declaring. This did not happen, see id., at 1-4, until the 19th century, after the framing.

What occurred in the present case, then, is precisely what Blackstone said—and the Framers believed—would not suffice. The Tennessee Supreme Court made no pretense that the year-and-a-day rule was "bad" law from the outset; rather, it asserted, the need for the rule, as a means of assuring causality of the death, had disappeared with time. Blackstone—and the Framers who were formed by Blackstone—would clearly have regarded that change in law as a matter for the legislature, beyond the power of the court. It may well be that some common-law decisions of the era in fact changed the law while purporting not to. But that is beside the point. What is important here is that it was an undoubted point of principle, at the time the Due Process Clause was adopted, that courts could not "change" the law. That explains why the Constitution restricted only the legislature from enacting ex post facto laws. Under accepted norms of judicial process, an ex post facto law (in the sense of a judicial holding, not that a prior decision was erroneous, but that the prior valid law is hereby retroactively changed ) was simply not an option for the courts. This attitude subsisted, I may note, well beyond the founding era, and beyond the time when due process guarantees were extended against the States by the Fourteenth Amendment. In an 1886 admiralty case, for example, this Court said the following: "The rights of persons in this particular under the maritime law of this country are not different from those under the common law, and as it is the duty of courts to declare the law, not to make it, we cannot change this rule." The Harrisburg, 119 U.S. 199, 213-214 (1886), overruled by Moragne v. States Marine Lines, Inc., 398 U.S. 375 (1970).

It is not a matter, therefore, of "[e]xtending the [Ex Post Facto] Clause to courts through the rubric of due process," and thereby "circumvent[ing] the clear constitutional text," ante, at 460. It is simply a matter of determining what due judicial process consists of—and it does not consist of retroactive creation of crimes. The Ex Post Facto Clause is relevant only because it demonstrates beyond doubt that, however much the acknowledged and accepted role of common-law courts could evolve (as it has) in other respects, retroactive revision of the criminal law was regarded as so fundamentally unfair that an alteration of the judicial role which permits that will be a denial of due process. Madison wrote that "ex-post-facto laws . . . are contrary to the first principles of the social compact, and to every principle of social legislation." The Federalist No. 44, p. 282 (C. Rossiter ed. 1961). I find it impossible to believe, as the Court does, that this strong sentiment attached only to retroactive laws passed by the legislature, and would not apply equally (or indeed with even greater force) to a court's production of the same result through disregard of the traditional limits upon judicial power. Insofar as the "first principles of the social compact" are concerned, what possible difference does it make that "[a] court's opportunity for discrimination" by retroactively changing a law "is more limited than a legislature's, in that it can only act in construing existing law in actual litigation"? Ante, at 460-461 (internal quotation marks and citation omitted). The injustice to the individuals affected is no less.

II

Even if I agreed with the Court that the Due Process Clause is violated only when there is lack of "fair warning" of the impending retroactive change, I would not find such fair warning here. It is not clear to me, in fact, what the Court believes the fair warning consisted of. Was it the mere fact that "[t]he year and a day rule is widely viewed as an outdated relic of the common law"? Ante, at 462. So are many of the elements of common-law crimes, such as "breaking the close" as an element of burglary, or "asportation" as an element of larceny. See W. LaFave & A. Scott, Criminal Law 631-633, 708-710 (1972). Are all of these "outdated relics" subject to retroactive judicial rescission? Or perhaps the fair warning consisted of the fact that "the year and a day rule has been legislatively or judicially abolished in the vast majority of jurisdictions recently to have addressed the issue." Ante, at 463. But why not count in petitioner's favor (as giving him no reason to expect a change in law) those even more numerous jurisdictions that have chosen not "recently to have addressed the issue"? And why not also count in petitioner's favor (rather than against him) those jurisdictions that have abolished the rule legislatively, and those jurisdictions that have abolished it through prospective rather than retroactive judicial rulings (together, a large majority of the abolitions, see 992 S.W.2d, at 397, n. 4, 402 (listing statutes and cases))? That is to say, even if it was predictable that the rule would be changed, it was not predictable that it would be changed retroactively, rather than in the prospective manner to which legislatures are restricted by the Ex Post Facto Clause, or in the prospective manner that most other courts have employed.

In any event, as the Court itself acknowledges, "[d]ue process . . . does not require a person to apprise himself of the common law of all 50 States in order to guarantee that his actions will not subject him to punishment in light of a developing trend in the law that has not yet made its way to his State." Ante, at 464. The Court tries to counter this self-evident point with the statement that "[a]t the same time, however, the fact that a vast number of jurisdictions have abolished a rule that has so clearly outlived its purpose is surely relevant to whether the abolition of the rule in a particular case can be said to be unexpected and indefensible by reference to the law as it then existed," ibid. This retort rests upon the fallacy that I discussed earlier: that "expected or defensible" "abolition" of prior law was approved by Bouie. It was not—and according such conclusive effect to the "defensibility" (by which I presume the Court means the "reasonableness") of the change in law will validate the retroactive creation of many new crimes.

Finally, the Court seeks to establish fair warning by discussing at great length, ante, at 464-466, how unclear it was that the year-and-a-day rule was ever the law in Tennessee. As I have already observed, the Supreme Court of Tennessee is the authoritative expositor of Tennessee law, and has said categorically that the year-and-a-day rule was the law. Does the Court mean to establish the principle that fair warning of impending change exists—or perhaps fair warning can be dispensed with—when the prior law is not crystal clear? Yet another boon for retroactively created crimes.

I reiterate that the only "fair warning" discussed in our precedents, and the only "fair warning" relevant to the issue before us here, is fair warning of what the law is. That warning, unlike the new one that today's opinion invents, goes well beyond merely "safeguarding defendants against unjustified and unpredictable breaks with prior law," ante, at 462 (emphasis added). It safeguards them against changes in the law after the fact. But even accepting the Court's novel substitute, the opinion's conclusion that this watered-down standard has been met seems to me to proceed on the principle that a large number of almost-valid arguments makes a solid case. As far as I can tell, petitioner had nothing that could fairly be called a "warning" that the Supreme Court of Tennessee would retroactively eliminate one of the elements of the crime of murder.

* * *

To decide this case, we need only conclude that due process prevents a court from (1) acknowledging the validity, when they were rendered, of prior decisions establishing a particular element of a crime; (2) changing the prior law so as to eliminate that element; and (3) applying that change to conduct that occurred under the prior regime. A court would remain free to apply common-law criminal rules to new fact patterns, see ante, at 461-462, so long as that application is consistent with a fair reading of prior cases. It would remain free to conclude that a prior decision or series of decisions establishing a particular element of a crime was in error, and to apply that conclusion retroactively (so long as the "fair notice" requirement of Bouie is satisfied). It would even remain free, insofar as the ex post facto element of the Due Process Clause is concerned, to "reevaluat[e] and refin[e]" the elements of common-law crimes to its heart's content, so long as it does so prospectively. (The majority of state courts that have abolished the year-and-a-day rule have done so in this fashion.) And, of course (as Blackstone and the Framers envisioned), legislatures would be free to eliminate outmoded elements of common-law crimes for the future by law. But what a court cannot do, consistent with due process, is what the Tennessee Supreme Court did here: avowedly change (to the defendant's disadvantage) the criminal law governing past acts.

For these reasons, I would reverse the judgment of the Supreme Court of Tennessee.

Justice Breyer, dissenting.

I agree with the Court's basic approach. Justice Cardozo pointed out that retroactivity should be determined "not by metaphysical conceptions of the nature of judge-made law, . . . but by considerations of convenience, of utility, and of the deepest sentiments of justice." The Nature of the Judicial Process 148-149 (1921). Similarly, the Due Process Clause asks us to consider the basic fairness or unfairness of retroactive application of the Tennessee court's change in the law. That Clause provides protection against after-the-fact changes in criminal law that deprive defendants of fair warning of the nature and consequences of their actions. It does not enshrine Blackstone's "ancient dogma that the law declared by . . . courts had a Platonic or ideal existence before the act of declaration," Great Northern R. Co. v. Sunburst Oil & Refining Co., 287 U.S. 358, 365 (1932) (Cardozo, J.). Cf. ante, at 473-474 (Scalia, J., dissenting).

I also agree with the Court that, in applying the Due Process Clause, we must ask whether the judicial ruling in question was "unexpected and indefensible by reference to the law which had been expressed prior to the conduct in issue." Bouie v. City of Columbia, 378 U.S. 347, 354 (1964) (quoting J. Hall, General Principles of Criminal Law 61 (2d ed. 1960) (internal quotation marks omitted)).

I cannot agree, however, with the majority's application of that due process principle to this case. As Justice Scalia well explains, Rogers did not have fair warning that the Tennessee courts would abolish the year and a day rule or that they would retroactively apply the new law to the circumstances of his case, thereby upgrading the crime those circumstances revealed from attempted murder to murder. I therefore join Part II of Justice Scalia's dissenting opinion.

NOTES

[*] Paula R. Voss filed a brief for the Tennessee Association of Criminal Defense Attorneys as amicus curiae urging reversal.

[1] Inquiring into a law's original reasonableness was perhaps tantamount to questioning whether it existed at all. "In holding the origin to have been unreasonable, the Court nearly always doubts or denies the actual origin and continuance of the custom in fact. " C. Allen, Law in the Making 140 (3d ed. 1939) (hereinafter Allen).

[2] The near-dispositive strength Blackstone accorded stare decisis was not some mere personal predilection. Chancellor Kent was of the same view: "If a decision has been made upon solemn argument and mature deliberation, the presumption is in favor of its correctness; and the community have a right to regard it as a just declaration or exposition of the law, and to regulate their actions and contracts by it." 1 J. Kent, Commentaries *475—*476 (emphasis added). See also Hamilton's statement in The Federalist: "To avoid an arbitrary discretion in the courts, it is indispensable that they should be bound down by strict rules and precedents which serve to define and point out their duty in every particular case that comes before them." The Federalist No. 78, p. 471 (C. Rossiter ed. 1961).

[3] In fact, however, "most of the basic departures [from English common law] were accomplished not by judicial decision but by local statute, so that by the time of the American Revolution one hears less and less about the unsuitability of common law principles to the American environment." 1 M. Horwitz, Transformation of American Law 1780-1860, p. 5 (1977) (hereinafter Horwitz).