Skip navigation
× You have 2 more free articles available this month. Subscribe today.

Habeas Hints

by Kent Russell

This column is intended to provide "habeas hints" to prisoners who are considering or handling habeas corpus petitions as their own attorneys ("in pro per"). The focus of the column is habeas corpus practice under the AEDPA, the 1996 habeas corpus law which now governs habeas corpus practice throughout the U.S.


2004 TERM

During the 2004 term, there were two U.S. Supreme Court decisions which significantly expanded the rights of criminal defendants by broadly interpreting the Sixth Amendment guarantees to the right of Confrontation and the right to a jury trial. In this column I discuss these two cases, Crawford v. Washington, 124 S.Ct. 1354 (2004) and Blakely v. Washington, 124 S. Ct. 1493 (2004), and provide some Habeas Hints for petitioners attempting to use the decisions to their advantage on habeas corpus.

Crawford v. Washington

Crawford held that "testimonial hearsay" cannot be introduced at a defendant's trial if the defendant did not have a prior opportunity to cross-examine the person making the statement.

In Crawford, a defendant was tried for attempted murder. The police had taken a tape-recorded statement from the defendant's wife, Sylvia, in which Sylvia related comments the defendant had made to her, which conflicted with the account the defendant later testified to at trial. Sylvia was "unavailable" for trial because she claimed the marital privilege, by which a spouse can refuse to testify against the other spouse. However, the prosecution was permitted to introduce Sylvia's statement to the police on the basis of a rule which allows a hearsay statement from an unavailable witness to be introduced in evidence if it is found to be "trustworthy" by the judge, and the defendant was convicted.

The Supreme Court reversed the conviction on the grounds that the admission of Sylvia's statement violated the Sixth Amendment. Undertaking an exhaustive historical analysis, the Supreme Court found that the Confrontation Clause of the 6th Amendment was intended to prohibit "testimonial hearsay" from being admitted, unless (a) the witness was unavailable for trial and (b) the defendant had a prior opportunity to cross-examine the witness. Applying these historical principles as a guide, the Court found that Sylvia's statement to the police was "testimonial" because it was a "formal statement to a governmental officer", and thus something that the declarant (Sylvia) could reasonably expect to be used in court. Hence, and because the defense hadn't had a prior opportunity to cross-examine Sylvia when she made her statement, the Court found it inadmissible because part (b) of the Court's test hadn't been satisfied.

In so holding, the Court overruled Ohio v. Roberts, 448 U.S. 56 (1980), a relatively recent Supreme Court case, which had permitted testimonial hearsay to come in upon application of a list of "reliability" factors that were usually found to be present by the trial judge. Thus, by holding that all testimonial hearsay which hasn't been subjected to cross-examination is inadmissible regardless of "reliability", Crawford effectively guts the prevailing practice in many states, especially in domestic violence cases, whereby prosecutors would put on their cases while shielding the victims from cross-examination.

Blakely v. Washington

In Blakely, the Court held that a judge cannot impose a sentence unless each fact necessary to justify that sentence has been found to be true by a jury, applying the beyond a reasonable doubt standard of proof, rather than by a judge, applying the much lower preponderance (51%) standard.

Blakely pled guilty to kidnapping his estranged wife in Washington state. The facts admitted in his plea justified a maximum sentence of 53 months. However, the sentencing judge made an additional finding _ that the defendant had acted with "deliberate cruelty" _ on which basis the judge sentenced Blakely to 90 months in prison. Blakely attacked his sentence in the Washington courts. Relying on the rule first announced in Apprendi v. New Jersey, 530 U.S. 466 (2000), he claimed that, because he had never admitted acting with "deliberate cruelty", nor had a jury found him guilty of that, the maximum sentence he could receive was 53 months. The Washington courts rejected the claim on the basis that Apprendi only applies where the sentence meted out is greater than the "statutory maximum"; that Washington law allowed a sentence of up to 10 years for aggravated kidnapping; and that Blakely's sentence was therefore less than the statutory maximum and, as a result, immune from an Apprendi attack. The Supreme Court reversed, holding that: "... The statutory maximum for Apprendi purposes is the maximum sentence a judge may impose solely on the basis of the facts reflected in the jury verdict or admitted by the defendant ... In other words, the relevant `statutory maximum' is not the maximum sentence a judge may impose after finding additional facts, but the maximum he may impose without any additional findings." Thus, even though Blakely had gotten a sentence substantially less than the 10-year max, the Court found that Blakely's sentence violated the 6th Amendment because it was based on facts found by a judge rather than by a jury.

Blakely is important because it takes the Apprendi focus off the maximum sentence allowable by statute _ which is usually so high that even a very harsh sentence will be lower than that _ and puts it on the facts actually found by the jury to be true. Under many sentencing systems, the prevailing practice has been for judges to ratchet up sentences based on facts that the jury never was asked to determine and never did determine. Thus, although Blakely by its terms only applies in Washington State, its implications are potentially so far-reaching that some federal courts have now questioned the constitutionality of the U.S. Sentencing Guidelines, which are applied to every sentence in federal court. See, e.g., United States v. Booker, 2004 U.S. App. Lexis 14223 (7th Cir. 7/9/04) [holding that sentence imposed in excess of maximum that Guidelines authorized violates 6th Amendment]; United States v. Penaranda, 2004 U.S. App. Lexis 14268 (2nd Cir 7/12/04) (en banc) [certifying constitutionality of Guidelines to Supreme Court for emergency review].

Habeas Hints

1. Make sure your case is covered by Crawford or Blakely before relying on either decision in a habeas corpus petition.

Crawford and Blakely are both very defendant-friendly decisions, and their potential effect is so broad that there has been considerable hoopla about them in the media. However, each decision has strict limitations, so look before you leap into basing a habeas corpus claim on either of these cases. For example, Crawford excludes only "testimonial hearsay", which requires a formal statement such as one given to police or investigators; it does not prohibit, for instance, admissions a defendant makes to friends or acquaintances. As for Blakely, it has no application at all to enhancements for prior convictions or "strikes" (which are not covered by Apprendi in the first place), and it will have very limited application in states like California, where the facts which normally enhance a sentence (for example, great bodily injury, use of a weapon, etc.) are already required to be submitted to the jury. In short, it's important to know that both of these cases have come down, but there's still a good chance that neither applies to your own situation.

2. If your case does involve a Crawford or a Blakely issue, be prepared to prevent or deal with retroactivity problems on habeas corpus.

The "Teague bar", see Teague v. Lane, 489 U.S. 288 (1989), generally prohibits a habeas corpus petitioner from benefiting from a "new" rule of law which was first announced by the Supreme Court after the petitioner's conviction and sentence became final on direct appeal. There are a couple of exceptions to the general rule, but they are applicable only in very rare situations, Therefore, I recommend the following strategies on state habeas corpus or on a § 2255 motion to vacate sentence in federal court:

a. Do whatever you can to attempt to raise your claim before your appeal becomes final.

The Teague bar only kicks in where the new rule relied upon on habeas came into being after the defendant's conviction became final on direct appeal. A conviction becomes final after the defendant has gone through one round of appeals in state court, has lost there, and has either been denied review on Certiorari ("Cert") in the U. S. Supreme Court, or the time has run out (90 days) within which to apply for Cert. Therefore, if your case hasn't become final yet, you should do everything you can to get a Crawford or Blakely claim on file before finality occurs and the Teague bar sets in. Taking California state prisoners as an example, even if your case has been fully briefed in the Court of Appeal and you're waiting for a decision there, ask your lawyer to file a Supplemental Brief and make your Crawford or Blakely claim in the Court of Appeal. If your conviction has been affirmed in the Court of Appeal, raise your claim in a Petition for Review to the California Supreme Court. If your Petition for Review has already been filed and is pending, file a Supplemental Petition for Review. If your Petition for Review has been denied and there's still time to file a Cert petition, make your claim there. Bottom line: If you've got a Crawford or Blakely issue and there's still time to raise it before your conviction becomes final, do so.

b. Trying to get around the non-retroactivity rule.

If your conviction has already become final and you have to try and overcome the Teague bar against retroactive application of new rules, you'll be fighting an uphill battle to have a court consider any claims based on Crawford or Blakely. However, you might consider the following:

First, if you've got a claim based on Blakely, and if your conviction became final between August of 2000 (when Apprendi came down) and June of 2004 (when Blakely was announced), you could label your claim as one based on Apprendi rather than on Blakely, on the basis that Blakely didn't really announce a "new" rule, but rather "clarified" the correct meaning of Apprendi. That way, for example, if your conviction became final in 2001, 2002, or 2003, you would be able to rely on Apprendi, since Apprendi was already the law when finality occurred in your case. Meanwhile, when you make your argument now on habeas, you would argue that the correct way to interpret Apprendi is the way the Supreme Court interpreted it in Blakely.

Second, on a Crawford claim, you could try to argue that Crawford isn't really a "new" rule, but rather a clarification of a very old rule that goes back centuries, but was only recently misinterpreted in Ohio v. Roberts. In other words, you would be arguing that the 6th Amendment has always meant that testimonial hearsay was inadmissible absent a prior opportunity to cross-examine an unavailable witness, and that the Teague bar doesn't apply because the rule in Crawford is not "new".

Third, you can monitor Crawford and Blakely and see whether some court down the road declares them retroactive. A habeas claim can be based on a new rule if the rule is declared to be "retroactive", and at least on a first .petition for habeas corpus, the declaration of retroactivity can come from any court; it doesn't have to be the Supreme Court, as would be required if you were filing a second or third ("successive") petition.

Finally, you can try and get around the Teague bar by arguing that your claim is an exception to Teague: either because the new rule you are relying on is "substantive" rather than procedural, or because the rule is one that effects a "watershed" change in criminal procedure. Unfortunately, these arguments have already been rejected in the Apprendi context by the Supreme Court in Schiro v. Summerlin, 124 S. Ct. 2519 (2004); and Crawford claims will probably meet the same fate. However, note that in Summerlin, the Arizona judge had already made the finding in question beyond a reasonable doubt. Hence, where findings are made only by a preponderance, you can still argue, based on existing precedent, that the requirement that all sentencing findings be made beyond a reasonable doubt should be retroactive because it is essential to accurate fact-finding. See, e.g., Ivan v. City of New York, 407 U.S. 203 (1972) and Hankerson v. North Carolina, 432 U.S. 233 (1977).

Kent Russell specializes in habeas corpus and § 2255 motions. He is the author of the California Habeas Handbook (4th Edition: 2003), which explains habeas corpus and the AEDPA. The book can be purchased with a check or money order for $29 (complete, includes 1st class postage, no order form needed), directly from the Law Offices of Russell and Russell, 2299 Sutter Street, San Francisco, CA 94115.

As a digital subscriber to Prison Legal News, you can access full text and downloads for this and other premium content.

Subscribe today

Already a subscriber? Login

Related legal cases

Crawford v. Washington

Crawford v. Washington, 124 S.Ct. 1354, 541 U.S. 36, 158 L.Ed.2d 177 (U.S. 03/08/2004)


[2] No. 02-9410

[3] 124 S.Ct. 1354, 541 U.S. 36, 158 L.Ed.2d 177, 63 Fed. R. Evid. Serv. 1077, 72 USLW 4229, 2004 Daily Journal D.A.R. 2949, 04 Cal. Daily Op. Serv. 2017, 4 Cal. Daily Op. Serv. 2017,

[4] March 08, 2004



[7] OCTOBER TERM, 2003

[8] Argued November 10, 2003

[9] Petitioner was tried for assault and attempted murder. The State sought to introduce a recorded statement that petitioner's wife Sylvia had made during police interrogation, as evidence that the stabbing was not in self-defense. Sylvia did not testify at trial because of Washington's marital privilege. Petitioner argued that admitting the evidence would violate his Sixth Amendment right to be "confronted with the witnesses against him." Under Ohio v. Roberts, 448 U. S. 56, that right does not bar admission of an unavailable witness's statement against a criminal defendant if the statement bears "adequate `indicia of reliability,' " a test met when the evidence either falls within a "firmly rooted hearsay exception" or bears "particularized guarantees of trustworthiness." Id., at 66. The trial court admitted the statement on the latter ground. The State Supreme Court upheld the conviction, deeming the statement reliable because it was nearly identical to, i.e., interlocked with, petitioner's own statement to the police, in that both were ambiguous as to whether the victim had drawn a weapon before petitioner assaulted him.

[10] Held: The State's use of Sylvia's statement violated the Confrontation Clause because, where testimonial statements are at issue, the only indicium of reliability sufficient to satisfy constitutional demands is confrontation. Pp. 5-33.

[11] (a) The Confrontation Clause's text does not alone resolve this case, so this Court turns to the Clause's historical background. That history supports two principles. First, the principal evil at which the Clause was directed was the civil-law mode of criminal procedure, particularly the use of ex parte examinations as evidence against the accused. The Clause's primary object is testimonial hearsay, and interrogations by law enforcement officers fall squarely within that class. Second, the Framers would not have allowed admission of testimonial statements of a witness who did not appear at trial unless he was unavailable to testify and the defendant had had a prior opportunity for cross-examination. English authorities and early state cases indicate that this was the common law at the time of the founding. And the "right ... to be confronted with the witnesses against him," Amdt. 6, is most naturally read as a reference to the common-law right of confrontation, admitting only those exceptions established at the time of the founding. See Mattox v. United States, 156 U. S. 237, 243. Pp. 5-21.

[12] (b) This Court's decisions have generally remained faithful to the Confrontation Clause's original meaning. See, e.g., Mattox, supra. Pp. 21-23.

[13] (c) However, the same cannot be said of the rationales of this Court's more recent decisions. See Roberts, supra, at 66. The Roberts test departs from historical principles because it admits statements consisting of ex parte testimony upon a mere reliability finding. Pp. 24-25.

[14] (d) The Confrontation Clause commands that reliability be assessed in a particular manner: by testing in the crucible of cross-examination. Roberts allows a jury to hear evidence, untested by the adversary process, based on a mere judicial determination of reliability, thus replacing the constitutionally prescribed method of assessing reliability with a wholly foreign one. Pp. 25-27.

[15] (e) Roberts' framework is unpredictable. Whether a statement is deemed reliable depends on which factors a judge considers and how much weight he accords each of them. However, the unpardonable vice of the Roberts test is its demonstrated capacity to admit core testimonial statements that the Confrontation Clause plainly meant to exclude. Pp. 27-30.

[16] (f) The instant case is a self-contained demonstration of Roberts' unpredictable and inconsistent application. It also reveals Roberts' failure to interpret the Constitution in a way that secures its intended constraint on judicial discretion. The Constitution prescribes the procedure for determining the reliability of testimony in criminal trials, and this Court, no less than the state courts, lacks authority to replace it with one of its own devising. Pp. 30-32.

[17] 147 Wash. 2d 424, 54 P. 3d 656, reversed and remanded.

[18] Scalia, J., delivered the opinion of the Court, in which Stevens, Kennedy, Souter, Thomas, Ginsburg, and Breyer, JJ., joined. Rehnquist, C. J., filed an opinion concurring in the judgment, in which O'Connor, J., joined.

[19] On Writ Of Certiorari To The Supreme Court Of Washington Court Below: 147 Wash. 2d 424, 54 P. 3d 656

[20] Jeffrey L. Fisher, by appointment of the Court, 540 U. S. 807, argued the cause for petitioner. With him on the briefs was Bruce E. H. Johnson.

[21] Steven C. Sherman argued the cause for respondent. With him on the brief was John Michael Jones.

[22] Deputy Solicitor General Dreeben argued the cause for the United States as amicus curiae urging affirmance. With him on the brief were Solicitor General Olson, Acting Assistant Attorney General Wray, Sri Srinivasan, and Joel M. Gershowitz.

[23] Briefs of amici curiae urging reversal were filed for the National Association of Criminal Defense Lawyers et al. by Jeffrey T. Green, David M. Porter, and Steven R. Shapiro; and for Sherman J. Clark et al. by Richard D. Friedman and David A. Moran.

[24] The opinion of the court was delivered by: Justice Scalia

[25] 541 U. S. ____ (2004)

[26] Petitioner Michael Crawford stabbed a man who allegedly tried to rape his wife, Sylvia. At his trial, the State played for the jury Sylvia's tape-recorded statement to the police describing the stabbing, even though he had no opportunity for cross-examination. The Washington Supreme Court upheld petitioner's conviction after determining that Sylvia's statement was reliable. The question presented is whether this procedure complied with the Sixth Amendment's guarantee that, "[i]n all criminal prosecutions, the accused shall enjoy the right ... to be confronted with the witnesses against him."

[27] I.

[28] On August 5, 1999, Kenneth Lee was stabbed at his apartment. Police arrested petitioner later that night. After giving petitioner and his wife Miranda warnings, detectives interrogated each of them twice. Petitioner eventually confessed that he and Sylvia had gone in search of Lee because he was upset over an earlier incident in which Lee had tried to rape her. The two had found Lee at his apartment, and a fight ensued in which Lee was stabbed in the torso and petitioner's hand was cut.

[29] Petitioner gave the following account of the fight:

[30] "Q. Okay. Did you ever see anything in [Lee's] hands?

[31] "A. I think so, but I'm not positive.

[32] "Q. Okay, when you think so, what do you mean by that?

[33] "A. I coulda swore I seen him goin' for somethin' before, right before everything happened. He was like reachin', fiddlin' around down here and stuff ... and I just ... I don't know, I think, this is just a possibility, but I think, I think that he pulled somethin' out and I grabbed for it and that's how I got cut ... but I'm not positive. I, I, my mind goes blank when things like this happen. I mean, I just, I remember things wrong, I remember things that just doesn't, don't make sense to me later." App. 155 (punctuation added).

[34] Sylvia generally corroborated petitioner's story about the events leading up to the fight, but her account of the fight itself was arguably different -- particularly with respect to whether Lee had drawn a weapon before petitioner assaulted him:

[35] "Q. Did Kenny do anything to fight back from this assault?

[36] "A. (pausing) I know he reached into his pocket ... or somethin' ... I don't know what.

[37] "Q. After he was stabbed?

[38] "A. He saw Michael coming up. He lifted his hand ... his chest open, he might [have] went to go strike his hand out or something and then (inaudible).

[39] "Q. Okay, you, you gotta speak up.

[40] "A. Okay, he lifted his hand over his head maybe to strike Michael's hand down or something and then he put his hands in his ... put his right hand in his right pocket ... took a step back ... Michael proceeded to stab him ... then his hands were like ... how do you explain this ... open arms ... with his hands open and he fell down ... and we ran (describing subject holding hands open, palms toward assailant).

[41] "Q. Okay, when he's standing there with his open hands, you're talking about Kenny, correct?

[42] "A. Yeah, after, after the fact, yes.

[43] "Q. Did you see anything in his hands at that point?

[44] "A. (pausing) um um (no)." Id., at 137 (punctuation added).

[45] The State charged petitioner with assault and attempted murder. At trial, he claimed self-defense. Sylvia did not testify because of the state marital privilege, which generally bars a spouse from testifying without the other spouse's consent. See Wash. Rev. Code §5.60.060(1) (1994). In Washington, this privilege does not extend to a spouse's out-of-court statements admissible under a hearsay exception, see State v. Burden, 120 Wash. 2d 371, 377, 841 P. 2d 758, 761 (1992), so the State sought to introduce Sylvia's tape-recorded statements to the police as evidence that the stabbing was not in self-defense. Noting that Sylvia had admitted she led petitioner to Lee's apartment and thus had facilitated the assault, the State invoked the hearsay exception for statements against penal interest, Wash. Rule Evid. 804(b)(3) (2003).

[46] Petitioner countered that, state law notwithstanding, admitting the evidence would violate his federal constitutional right to be "confronted with the witnesses against him." Amdt. 6. According to our description of that right in Ohio v. Roberts, 448 U. S. 56 (1980), it does not bar admission of an unavailable witness's statement against a criminal defendant if the statement bears "adequate `indicia of reliability.' " Id., at 66. To meet that test, evidence must either fall within a "firmly rooted hearsay exception" or bear "particularized guarantees of trustworthiness." Ibid. The trial court here admitted the statement on the latter ground, offering several reasons why it was trustworthy: Sylvia was not shifting blame but rather corroborating her husband's story that he acted in self-defense or "justified reprisal"; she had direct knowledge as an eyewitness; she was describing recent events; and she was being questioned by a "neutral" law enforcement officer. App. 76-77. The prosecution played the tape for the jury and relied on it in closing, arguing that it was "damning evidence" that "completely refutes [petitioner's] claim of self-defense." Tr. 468 (Oct. 21, 1999). The jury convicted petitioner of assault.

[47] The Washington Court of Appeals reversed. It applied a nine-factor test to determine whether Sylvia's statement bore particularized guarantees of trustworthiness, and noted several reasons why it did not: The statement contradicted one she had previously given; it was made in response to specific questions; and at one point she admitted she had shut her eyes during the stabbing. The court considered and rejected the State's argument that Sylvia's statement was reliable because it coincided with petitioner's to such a degree that the two "interlocked." The court determined that, although the two statements agreed about the events leading up to the stabbing, they differed on the issue crucial to petitioner's self-defense claim: "[Petitioner's] version asserts that Lee may have had something in his hand when he stabbed him; but Sylvia's version has Lee grabbing for something only after he has been stabbed." App. 32.

[48] The Washington Supreme Court reinstated the conviction, unanimously concluding that, although Sylvia's statement did not fall under a firmly rooted hearsay exception, it bore guarantees of trustworthiness: " `[W]hen a co-defendant's confession is virtually identical [to, i.e., interlocks with,] that of a defendant, it may be deemed reliable.' " 147 Wash. 2d 424, 437, 54 P. 3d 656, 663 (2002) (quoting State v. Rice, 120 Wash. 2d 549, 570, 844 P. 2d 416, 427 (1993)). The court explained:

[49] "Although the Court of Appeals concluded that the statements were contradictory, upon closer inspection they appear to overlap... .

[50] "[B]oth of the Crawfords' statements indicate that Lee was possibly grabbing for a weapon, but they are equally unsure when this event may have taken place. They are also equally unsure how Michael received the cut on his hand, leading the court to question when, if ever, Lee possessed a weapon. In this respect they overlap.

[51] "[N]either Michael nor Sylvia clearly stated that Lee had a weapon in hand from which Michael was simply defending himself. And it is this omission by both that interlocks the statements and makes Sylvia's statement reliable." 147 Wash. 2d, at 438- 439, 54 P. 3d, at 664 (internal quotation marks omitted).*fn1

[52] We granted certiorari to determine whether the State's use of Sylvia's statement violated the Confrontation Clause. 539 U. S. 914 (2003).

[53] II.

[54] The Sixth Amendment's Confrontation Clause provides that, "[i]n all criminal prosecutions, the accused shall enjoy the right ... to be confronted with the witnesses against him." We have held that this bedrock procedural guarantee applies to both federal and state prosecutions. Pointer v. Texas, 380 U. S. 400, 406 (1965). As noted above, Roberts says that an unavailable witness's out-of-court statement may be admitted so long as it has adequate indicia of reliability -- i.e., falls within a "firmly rooted hearsay exception" or bears "particularized guarantees of trustworthiness." 448 U. S., at 66. Petitioner argues that this test strays from the original meaning of the Confrontation Clause and urges us to reconsider it.

[55] A.

[56] The Constitution's text does not alone resolve this case. One could plausibly read "witnesses against" a defendant to mean those who actually testify at trial, cf. Woodsides v. State, 3 Miss. 655, 664-665 (1837), those whose statements are offered at trial, see 3 J. Wigmore, Evidence §1397, p. 104 (2d ed. 1923) (hereinafter Wigmore), or something in-between, see infra, at 15-16. We must therefore turn to the historical background of the Clause to understand its meaning.

[57] The right to confront one's accusers is a concept that dates back to Roman times. See Coy v. Iowa, 487 U. S. 1012, 1015 (1988); Herrmann & Speer, Facing the Accuser: Ancient and Medieval Precursors of the Confrontation Clause, 34 Va. J. Int'l L. 481 (1994). The founding generation's immediate source of the concept, however, was the common law. English common law has long differed from continental civil law in regard to the manner in which witnesses give testimony in criminal trials. The common-law tradition is one of live testimony in court subject to adversarial testing, while the civil law condones examination in private by judicial officers. See 3 W. Blackstone, Commentaries on the Laws of England 373-374 (1768).

[58] Nonetheless, England at times adopted elements of the civil-law practice. Justices of the peace or other officials examined suspects and witnesses before trial. These examinations were sometimes read in court in lieu of live testimony, a practice that "occasioned frequent demands by the prisoner to have his `accusers,' i.e. the witnesses against him, brought before him face to face." 1 J. Stephen, History of the Criminal Law of England 326 (1883). In some cases, these demands were refused. See 9 W. Holdsworth, History of English Law 216-217, 228 (3d ed. 1944); e.g., Raleigh's Case, 2 How. St. Tr. 1, 15-16, 24 (1603); Throckmorton's Case, 1 How. St. Tr. 869, 875-876 (1554); cf. Lilburn's Case, 3 How. St. Tr. 1315, 1318-1322, 1329 (Star Chamber 1637).

[59] Pretrial examinations became routine under two statutes passed during the reign of Queen Mary in the 16th century, 1 & 2 Phil. & M., c. 13 (1554), and 2 & 3 id., c. 10 (1555). These Marian bail and committal statutes required justices of the peace to examine suspects and witnesses in felony cases and to certify the results to the court. It is doubtful that the original purpose of the examinations was to produce evidence admissible at trial. See J. Langbein, Prosecuting Crime in the Renaissance 21-34 (1974). Whatever the original purpose, however, they came to be used as evidence in some cases, see 2 M. Hale, Pleas of the Crown 284 (1736), resulting in an adoption of continental procedure. See 4 Holdsworth, supra, at 528-530.

[60] The most notorious instances of civil-law examination occurred in the great political trials of the 16th and 17th centuries. One such was the 1603 trial of Sir Walter Raleigh for treason. Lord Cobham, Raleigh's alleged accomplice, had implicated him in an examination before the Privy Council and in a letter. At Raleigh's trial, these were read to the jury. Raleigh argued that Cobham had lied to save himself: "Cobham is absolutely in the King's mercy; to excuse me cannot avail him; by accusing me he may hope for favour." 1 D. Jardine, Criminal Trials 435 (1832). Suspecting that Cobham would recant, Raleigh demanded that the judges call him to appear, arguing that "[t]he Proof of the Common Law is by witness and jury: let Cobham be here, let him speak it. Call my accuser before my face ... ." 2 How. St. Tr., at 15-16. The judges refused, id., at 24, and, despite Raleigh's protestations that he was being tried "by the Spanish Inquisition," id., at 15, the jury convicted, and Raleigh was sentenced to death.

[61] One of Raleigh's trial judges later lamented that " `the justice of England has never been so degraded and injured as by the condemnation of Sir Walter Raleigh.' " 1 Jardine, supra, at 520. Through a series of statutory and judicial reforms, English law developed a right of confrontation that limited these abuses. For example, treason statutes required witnesses to confront the accused "face to face" at his arraignment. E.g., 13 Car. 2, c. 1, §5 (1661); see 1 Hale, supra, at 306. Courts, meanwhile, developed relatively strict rules of unavailability, admitting examinations only if the witness was demonstrably unable to testify in person. See Lord Morley's Case, 6 How. St. Tr. 769, 770-771 (H. L. 1666); 2 Hale, supra, at 284; 1 Stephen, supra, at 358. Several authorities also stated that a suspect's confession could be admitted only against himself, and not against others he implicated. See 2 W. Hawkins, Pleas of the Crown c. 46, §3, pp. 603-604 (T. Leach 6th ed. 1787); 1 Hale, supra, at 585, n. (k); 1 G. Gilbert, Evidence 216 (C. Lofft ed. 1791); cf. Tong's Case, Kel. J. 17, 18, 84 Eng. Rep. 1061, 1062 (1662) (treason). But see King v. Westbeer, 1 Leach 12, 168 Eng. Rep. 108, 109 (1739).

[62] One recurring question was whether the admissibility of an unavailable witness's pretrial examination depended on whether the defendant had had an opportunity to cross-examine him. In 1696, the Court of King's Bench answered this question in the affirmative, in the widely reported misdemeanor libel case of King v. Paine, 5 Mod. 163, 87 Eng. Rep. 584. The court ruled that, even though a witness was dead, his examination was not admissible where "the defendant not being present when [it was] taken before the mayor ... had lost the benefit of a cross-examination." Id., at 165, 87 Eng. Rep., at 585. The question was also debated at length during the infamous proceedings against Sir John Fenwick on a bill of attainder. Fenwick's counsel objected to admitting the examination of a witness who had been spirited away, on the ground that Fenwick had had no opportunity to cross-examine. See Fenwick's Case, 13 How. St. Tr. 537, 591-592 (H. C. 1696) (Powys) ("[T]hat which they would offer is something that Mr. Goodman hath sworn when he was examined ... ; sir J. F. not being present or privy, and no opportunity given to cross-examine the person; and I conceive that cannot be offered as evidence ..."); id., at 592 (Shower) ("[N]o deposition of a person can be read, though beyond sea, unless in cases where the party it is to be read against was privy to the examination, and might have cross-examined him ... . [O]ur constitution is, that the person shall see his accuser"). The examination was nonetheless admitted on a closely divided vote after several of those present opined that the common-law rules of procedure did not apply to parliamentary attainder proceedings --one speaker even admitting that the evidence would normally be inadmissible. See id., at 603-604 (Williamson); id., at 604-605 (Chancellor of the Exchequer); id., at 607; 3 Wigmore §1364, at 22-23, n. 54. Fenwick was condemned, but the proceedings "must have burned into the general consciousness the vital importance of the rule securing the right of cross-examination." Id., §1364, at 22; cf. Carmell v. Texas, 529 U. S. 513, 526-530 (2000).

[63] Paine had settled the rule requiring a prior opportunity for cross-examination as a matter of common law, but some doubts remained over whether the Marian statutes prescribed an exception to it in felony cases. The statutes did not identify the circumstances under which examinations were admissible, see 1 & 2 Phil. & M., c. 13 (1554); 2 & 3 id., c. 10 (1555), and some inferred that no prior opportunity for cross-examination was required. See Westbeer, supra, at 12, 168 Eng. Rep., at 109; compare Fenwick's Case, 13 How. St. Tr., at 596 (Sloane), with id., at 602 (Musgrave). Many who expressed this view acknowledged that it meant the statutes were in derogation of the common law. See King v. Eriswell, 3 T. R. 707, 710, 100 Eng. Rep. 815, 817 (K. B. 1790) (Grose, J.) (dicta); id., at 722-723, 100 Eng. Rep., at 823-824 (Kenyon, C. J.) (same); compare 1 Gilbert, Evidence, at 215 (admissible only "by Force `of the Statute' "), with id., at 65. Nevertheless, by 1791 (the year the Sixth Amendment was ratified), courts were applying the cross-examination rule even to examinations by justices of the peace in felony cases. See King v. Dingler, 2 Leach 561, 562-563, 168 Eng. Rep. 383, 383-384 (1791); King v. Woodcock, 1 Leach 500, 502-504, 168 Eng. Rep. 352, 353 (1789); cf. King v. Radbourne, 1 Leach 457, 459-461, 168 Eng. Rep. 330, 331-332 (1787); 3 Wigmore §1364, at 23. Early 19th-century treatises confirm that requirement. See 1 T. Starkie, Evidence 95 (1826); 2 id., at 484-492; T. Peake, Evidence 63-64 (3d ed. 1808). When Parliament amended the statutes in 1848 to make the requirement explicit, see 11 & 12 Vict., c. 42, §17, the change merely "introduced in terms" what was already afforded the defendant "by the equitable construction of the law." Queen v. Beeston, 29 Eng. L. & Eq. R. 527, 529 (Ct. Crim. App. 1854) (Jervis, C. J.).*fn2

[64] B.

[65] Controversial examination practices were also used in the Colonies. Early in the 18th century, for example, the Virginia Council protested against the Governor for having "privately issued several commissions to examine witnesses against particular men ex parte," complaining that "the person accused is not admitted to be confronted with, or defend himself against his defamers." A Memorial Concerning the Maladministrations of His Excellency Francis Nicholson, reprinted in 9 English Historical Documents 253, 257 (D. Douglas ed. 1955). A decade before the Revolution, England gave jurisdiction over Stamp Act offenses to the admiralty courts, which followed civil-law rather than common-law procedures and thus routinely took testimony by deposition or private judicial examination. See 5 Geo. 3, c. 12, §57 (1765); Pollitt, The Right of Confrontation: Its History and Modern Dress, 8 J. Pub. L. 381, 396-397 (1959). Colonial representatives protested that the Act subverted their rights "by extending the jurisdiction of the courts of admiralty beyond its ancient limits." Resolutions of the Stamp Act Congress §8th (Oct. 19, 1765), reprinted in Sources of Our Liberties 270, 271 (R. Perry & J. Cooper eds. 1959). John Adams, defending a merchant in a high-profile admiralty case, argued: "Examinations of witnesses upon Interrogatories, are only by the Civil Law. Interrogatories are unknown at common Law, and Englishmen and common Lawyers have an aversion to them if not an Abhorrence of them." Draft of Argument in Sewall v. Hancock (1768-1769), in 2 Legal Papers of John Adams 194, 207 (K. Wroth & H. Zobel eds. 1965).

[66] Many declarations of rights adopted around the time of the Revolution guaranteed a right of confrontation. See Virginia Declaration of Rights §8 (1776); Pennsylvania Declaration of Rights §IX (1776); Delaware Declaration of Rights §14 (1776); Maryland Declaration of Rights §XIX (1776); North Carolina Declaration of Rights §VII (1776); Vermont Declaration of Rights Ch. I, §X (1777); Massachusetts Declaration of Rights §XII (1780); New Hampshire Bill of Rights §XV (1783), all reprinted in 1 B. Schwartz, The Bill of Rights: A Documentary History 235, 265, 278, 282, 287, 323, 342, 377 (1971). The proposed Federal Constitution, however, did not. At the Massachusetts ratifying convention, Abraham Holmes objected to this omission precisely on the ground that it would lead to civil-law practices: "The mode of trial is altogether indetermined; ... whether [the defendant] is to be allowed to confront the witnesses, and have the advantage of cross-examination, we are not yet told... . [W]e shall find Congress possessed of powers enabling them to institute judicatories little less inauspicious than a certain tribunal in Spain, ... the Inquisition." 2 Debates on the Federal Constitution 110-111 (J. Elliot 2d ed. 1863). Similarly, a prominent Antifederalist writing under the pseudonym Federal Farmer criticized the use of "written evidence" while objecting to the omission of a vicinage right: "Nothing can be more essential than the cross examining [of] witnesses, and generally before the triers of the facts in question... . [W]ritten evidence ... [is] almost useless; it must be frequently taken ex parte, and but very seldom leads to the proper discovery of truth." R. Lee, Letter IV by the Federal Farmer (Oct. 15, 1787), reprinted in 1 Schwartz, supra, at 469, 473. The First Congress responded by including the Confrontation Clause in the proposal that became the Sixth Amendment.

[67] Early state decisions shed light upon the original understanding of the common-law right. State v. Webb, 2 N. C. 103 (1794) (per curiam), decided a mere three years after the adoption of the Sixth Amendment, held that depositions could be read against an accused only if they were taken in his presence. Rejecting a broader reading of the English authorities, the court held: "[I]t is a rule of the common law, founded on natural justice, that no man shall be prejudiced by evidence which he had not the liberty to cross examine." Id., at 104.

[68] Similarly, in State v. Campbell, 1 S. C. 124 (1844), South Carolina's highest law court excluded a deposition taken by a coroner in the absence of the accused. It held: "[I]f we are to decide the question by the established rules of the common law, there could not be a dissenting voice. For, notwithstanding the death of the witness, and whatever the respectability of the court taking the depositions, the solemnity of the occasion and the weight of the testimony, such depositions are ex parte, and, therefore, utterly incompetent." Id., at 125. The court said that one of the "indispensable conditions" implicitly guaranteed by the State Constitution was that "prosecutions be carried on to the conviction of the accused, by witnesses confronted by him, and subjected to his personal examination." Ibid.

[69] Many other decisions are to the same effect. Some early cases went so far as to hold that prior testimony was inadmissible in criminal cases even if the accused had a previous opportunity to cross-examine. See Finn v. Commonwealth, 26 Va. 701, 708 (1827); State v. Atkins, 1 Tenn. 229 (1807) (per curiam). Most courts rejected that view, but only after reaffirming that admissibility depended on a prior opportunity for cross-examination. See United States v. Macomb, 26 F. Cas. 1132, 1133 (No. 15,702) (CC Ill. 1851); State v. Houser, 26 Mo. 431, 435-436 (1858); Kendrick v. State, 29 Tenn. 479, 485-488 (1850); Bostick v. State, 22 Tenn. 344, 345-346 (1842); Commonwealth v. Richards, 35 Mass. 434, 437 (1837); State v. Hill, 2 Hill 607, 608-610 (S. C. 1835); Johnston v. State, 10 Tenn. 58, 59 (1821). Nineteenth-century treatises confirm the rule. See 1 J. Bishop, Criminal Procedure §1093, p. 689 (2d ed. 1872); T. Cooley, Constitutional Limitations *318.

[70] III.

[71] This history supports two inferences about the meaning of the Sixth Amendment.

[72] A.

[73] First, the principal evil at which the Confrontation Clause was directed was the civil-law mode of criminal procedure, and particularly its use of ex parte examinations as evidence against the accused. It was these practices that the Crown deployed in notorious treason cases like Raleigh's; that the Marian statutes invited; that English law's assertion of a right to confrontation was meant to prohibit; and that the founding-era rhetoric decried. The Sixth Amendment must be interpreted with this focus in mind.

[74] Accordingly, we once again reject the view that the Confrontation Clause applies of its own force only to in-court testimony, and that its application to out-of-court statements introduced at trial depends upon "the law of Evidence for the time being." 3 Wigmore §1397, at 101; accord, Dutton v. Evans, 400 U. S. 74, 94 (1970) (Harlan, J., concurring in result). Leaving the regulation of out-of-court statements to the law of evidence would render the Confrontation Clause powerless to prevent even the most flagrant inquisitorial practices. Raleigh was, after all, perfectly free to confront those who read Cobham's confession in court.

[75] This focus also suggests that not all hearsay implicates the Sixth Amendment's core concerns. An off-hand, overheard remark might be unreliable evidence and thus a good candidate for exclusion under hearsay rules, but it bears little resemblance to the civil-law abuses the Confrontation Clause targeted. On the other hand, ex parte examinations might sometimes be admissible under modern hearsay rules, but the Framers certainly would not have condoned them.

[76] The text of the Confrontation Clause reflects this focus. It applies to "witnesses" against the accused -- in other words, those who "bear testimony." 1 N. Webster, An American Dictionary of the English Language (1828). "Testimony," in turn, is typically "[a] solemn declaration or affirmation made for the purpose of establishing or proving some fact." Ibid. An accuser who makes a formal statement to government officers bears testimony in a sense that a person who makes a casual remark to an acquaintance does not. The constitutional text, like the history underlying the common-law right of confrontation, thus reflects an especially acute concern with a specific type of out-of-court statement.

[77] Various formulations of this core class of "testimonial" statements exist: "ex parte in-court testimony or its functional equivalent -- that is, material such as affidavits, custodial examinations, prior testimony that the defendant was unable to cross-examine, or similar pretrial statements that declarants would reasonably expect to be used prosecutorially," Brief for Petitioner 23; "extra-judicial statements ... contained in formalized testimonial materials, such as affidavits, depositions, prior testimony, or confessions," White v. Illinois, 502 U. S. 346, 365 (1992) (Thomas, J., joined by Scalia, J., concurring in part and concurring in judgment); "statements that were made under circumstances which would lead an objective witness reasonably to believe that the statement would be available for use at a later trial," Brief for National Association of Criminal Defense Lawyers et al. as Amici Curiae 3. These formulations all share a common nucleus and then define the Clause's coverage at various levels of abstraction around it. Regardless of the precise articulation, some statements qualify under any definition -- for example, ex parte testimony at a preliminary hearing.

[78] Statements taken by police officers in the course of interrogations are also testimonial under even a narrow standard. Police interrogations bear a striking resemblance to examinations by justices of the peace in England. The statements are not sworn testimony, but the absence of oath was not dispositive. Cobham's examination was unsworn, see 1 Jardine, Criminal Trials, at 430, yet Raleigh's trial has long been thought a paradigmatic confrontation violation, see, e.g., Campbell, 1 S. C., at 130. Under the Marian statutes, witnesses were typically put on oath, but suspects were not. See 2 Hale, Pleas of the Crown, at 52. Yet Hawkins and others went out of their way to caution that such unsworn confessions were not admissible against anyone but the confessor. See supra, at 8.*fn3

[79] That interrogators are police officers rather than magistrates does not change the picture either. Justices of the peace conducting examinations under the Marian statutes were not magistrates as we understand that office today, but had an essentially investigative and prosecutorial function. See 1 Stephen, Criminal Law of England, at 221; Langbein, Prosecuting Crime in the Renaissance, at 34-45. England did not have a professional police force until the 19th century, see 1 Stephen, supra, at 194-200, so it is not surprising that other government officers performed the investigative functions now associated primarily with the police. The involvement of government officers in the production of testimonial evidence presents the same risk, whether the officers are police or justices of the peace.

[80] In sum, even if the Sixth Amendment is not solely concerned with testimonial hearsay, that is its primary object, and interrogations by law enforcement officers fall squarely within that class.*fn4

[81] B.

[82] The historical record also supports a second proposition: that the Framers would not have allowed admission of testimonial statements of a witness who did not appear at trial unless he was unavailable to testify, and the defendant had had a prior opportunity for cross-examination. The text of the Sixth Amendment does not suggest any open-ended exceptions from the confrontation requirement to be developed by the courts. Rather, the "right ... to be confronted with the witnesses against him," Amdt. 6, is most naturally read as a reference to the right of confrontation at common law, admitting only those exceptions established at the time of the founding. See Mattox v. United States, 156 U. S. 237, 243 (1895); cf. Houser, 26 Mo., at 433-435. As the English authorities above reveal, the common law in 1791 conditioned admissibility of an absent witness's examination on unavailability and a prior opportunity to cross-examine. The Sixth Amendment therefore incorporates those limitations. The numerous early state decisions applying the same test confirm that these principles were received as part of the common law in this country.*fn5

[83] We do not read the historical sources to say that a prior opportunity to cross-examine was merely a sufficient, rather than a necessary, condition for admissibility of testimonial statements. They suggest that this requirement was dispositive, and not merely one of several ways to establish reliability. This is not to deny, as The Chief Justice notes, that "[t]here were always exceptions to the general rule of exclusion" of hearsay evidence. Post, at 5. Several had become well established by 1791. See 3 Wigmore §1397, at 101; Brief for United States as Amicus Curiae 13, n. 5. But there is scant evidence that exceptions were invoked to admit testimonial statements against the accused in a criminal case.*fn6 Most of the hearsay exceptions covered statements that by their nature were not testimonial -- for example, business records or statements in furtherance of a conspiracy. We do not infer from these that the Framers thought exceptions would apply even to prior testimony. Cf. Lilly v. Virginia, 527 U. S. 116, 134 (1999) (plurality opinion) ("[A]ccomplices' confessions that inculpate a criminal defendant are not within a firmly rooted exception to the hearsay rule").*fn7

[84] IV.

[85] Our case law has been largely consistent with these two principles. Our leading early decision, for example, involved a deceased witness's prior trial testimony. Mattox v. United States, 156 U. S. 237 (1895). In allowing the statement to be admitted, we relied on the fact that the defendant had had, at the first trial, an adequate opportunity to confront the witness: "The substance of the constitutional protection is preserved to the prisoner in the advantage he has once had of seeing the witness face to face, and of subjecting him to the ordeal of a cross-examination. This, the law says, he shall under no circumstances be deprived of ... ." Id., at 244.

[86] Our later cases conform to Mattox's holding that prior trial or preliminary hearing testimony is admissible only if the defendant had an adequate opportunity to cross-examine. See Mancusi v. Stubbs, 408 U. S. 204, 213-216 (1972); California v. Green, 399 U. S. 149, 165-168 (1970); Pointer v. Texas, 380 U. S., at 406-408; cf. Kirby v. United States, 174 U. S. 47, 55-61 (1899). Even where the defendant had such an opportunity, we excluded the testimony where the government had not established unavailability of the witness. See Barber v. Page, 390 U. S. 719, 722-725 (1968); cf. Motes v. United States, 178 U. S. 458, 470-471 (1900). We similarly excluded accomplice confessions where the defendant had no opportunity to cross-examine. See Roberts v. Russell, 392 U. S. 293, 294-295 (1968) (per curiam); Bruton v. United States, 391 U. S. 123, 126-128 (1968); Douglas v. Alabama, 380 U. S. 415, 418-420 (1965). In contrast, we considered reliability factors beyond prior opportunity for cross-examination when the hearsay statement at issue was not testimonial. See Dutton v. Evans, 400 U. S., at 87-89 (plurality opinion).

[87] Even our recent cases, in their outcomes, hew closely to the traditional line. Ohio v. Roberts, 448 U. S., at 67-70, admitted testimony from a preliminary hearing at which the defendant had examined the witness. Lilly v. Virginia, supra, excluded testimonial statements that the defendant had had no opportunity to test by cross-examination. And Bourjaily v. United States, 483 U. S. 171, 181-184 (1987), admitted statements made unwittingly to an FBI informant after applying a more general test that did not make prior cross-examination an indispensable requirement.*fn8

[88] Lee v. Illinois, 476 U. S. 530 (1986), on which the State relies, is not to the contrary. There, we rejected the State's attempt to admit an accomplice confession. The State had argued that the confession was admissible because it "interlocked" with the defendant's. We dealt with the argument by rejecting its premise, holding that "when the discrepancies between the statements are not insignificant, the co-defendant's confession may not be admitted." Id., at 545. Respondent argues that "[t]he logical inference of this statement is that when the discrepancies between the statements are insignificant, then the co-defendant's statement may be admitted." Brief for Respondent 6. But this is merely a possible inference, not an inevitable one, and we do not draw it here. If Lee had meant authoritatively to announce an exception -- previously unknown to this Court's jurisprudence -- for interlocking confessions, it would not have done so in such an oblique manner. Our only precedent on interlocking confessions had addressed the entirely different question whether a limiting instruction cured prejudice to co-defendants from admitting a defendant's own confession against him in a joint trial. See Parker v. Randolph, 442 U. S. 62, 69-76 (1979) (plurality opinion), abrogated by Cruz v. New York, 481 U. S. 186 (1987).

[89] Our cases have thus remained faithful to the Framers' understanding: Testimonial statements of witnesses absent from trial have been admitted only where the declarant is unavailable, and only where the defendant has had a prior opportunity to cross-examine.*fn9

[90] V.

[91] Although the results of our decisions have generally been faithful to the original meaning of the Confrontation Clause, the same cannot be said of our rationales. Roberts conditions the admissibility of all hearsay evidence on whether it falls under a "firmly rooted hearsay exception" or bears "particularized guarantees of trustworthiness." 448 U. S., at 66. This test departs from the historical principles identified above in two respects. First, it is too broad: It applies the same mode of analysis whether or not the hearsay consists of ex parte testimony. This often results in close constitutional scrutiny in cases that are far removed from the core concerns of the Clause. At the same time, however, the test is too narrow: It admits statements that do consist of ex parte testimony upon a mere finding of reliability. This malleable standard often fails to protect against paradigmatic confrontation violations.

[92] Members of this Court and academics have suggested that we revise our doctrine to reflect more accurately the original understanding of the Clause. See, e.g., Lilly, 527 U. S., at 140-143 (Breyer, J., concurring); White, 502 U. S., at 366 (Thomas, J., joined by Scalia, J., concurring in part and concurring in judgment); A. Amar, The Constitution and Criminal Procedure 125-131 (1997); Friedman, Confrontation: The Search for Basic Principles, 86 Geo. L. J. 1011 (1998). They offer two proposals: First, that we apply the Confrontation Clause only to testimonial statements, leaving the remainder to regulation by hearsay law -- thus eliminating the overbreadth referred to above. Second, that we impose an absolute bar to statements that are testimonial, absent a prior opportunity to cross-examine -- thus eliminating the excessive narrowness referred to above.

[93] In White, we considered the first proposal and rejected it. 502 U. S., at 352-353. Although our analysis in this case casts doubt on that holding, we need not definitively resolve whether it survives our decision today, because Sylvia Crawford's statement is testimonial under any definition. This case does, however, squarely implicate the second proposal.

[94] A.

[95] Where testimonial statements are involved, we do not think the Framers meant to leave the Sixth Amendment's protection to the vagaries of the rules of evidence, much less to amorphous notions of "reliability." Certainly none of the authorities discussed above acknowledges any general reliability exception to the common-law rule. Admitting statements deemed reliable by a judge is fundamentally at odds with the right of confrontation. To be sure, the Clause's ultimate goal is to ensure reliability of evidence, but it is a procedural rather than a substantive guarantee. It commands, not that evidence be reliable, but that reliability be assessed in a particular manner: by testing in the crucible of cross-examination. The Clause thus reflects a judgment, not only about the desirability of reliable evidence (a point on which there could be little dissent), but about how reliability can best be determined. Cf. 3 Blackstone, Commentaries, at 373 ("This open examination of witnesses . . . is much more conducive to the clearing up of truth"); M. Hale, History and Analysis of the Common Law of England 258 (1713) (adversarial testing "beats and bolts out the Truth much better").

[96] The Roberts test allows a jury to hear evidence, untested by the adversary process, based on a mere judicial determination of reliability. It thus replaces the constitutionally prescribed method of assessing reliability with a wholly foreign one. In this respect, it is very different from exceptions to the Confrontation Clause that make no claim to be a surrogate means of assessing reliability. For example, the rule of forfeiture by wrongdoing (which we accept) extinguishes confrontation claims on essentially equitable grounds; it does not purport to be an alternative means of determining reliability. See Reynolds v. United States, 98 U. S. 145, 158-159 (1879).

[97] The Raleigh trial itself involved the very sorts of reliability determinations that Roberts authorizes. In the face of Raleigh's repeated demands for confrontation, the prosecution responded with many of the arguments a court applying Roberts might invoke today: that Cobham's statements were self-inculpatory, 2 How. St. Tr., at 19, that they were not made in the heat of passion, id., at 14, and that they were not "extracted from [him] upon any hopes or promise of Pardon," id., at 29. It is not plausible that the Framers' only objection to the trial was that Raleigh's judges did not properly weigh these factors before sentencing him to death. Rather, the problem was that the judges refused to allow Raleigh to confront Cobham in court, where he could cross-examine him and try to expose his accusation as a lie.

[98] Dispensing with confrontation because testimony is obviously reliable is akin to dispensing with jury trial because a defendant is obviously guilty. This is not what the Sixth Amendment prescribes.

[99] B.

[100] The legacy of Roberts in other courts vindicates the Framers' wisdom in rejecting a general reliability exception. The framework is so unpredictable that it fails to provide meaningful protection from even core confrontation violations.

[101] Reliability is an amorphous, if not entirely subjective, concept. There are countless factors bearing on whether a statement is reliable; the nine-factor balancing test applied by the Court of Appeals below is representative. See, e.g., People v. Farrell, 34 P. 3d 401, 406-407 (Colo. 2001) (eight-factor test). Whether a statement is deemed reliable depends heavily on which factors the judge considers and how much weight he accords each of them. Some courts wind up attaching the same significance to opposite facts. For example, the Colorado Supreme Court held a statement more reliable because its inculpation of the defendant was "detailed," id., at 407, while the Fourth Circuit found a statement more reliable because the portion implicating another was "fleeting," United States v. Photogrammetric Data Servs., Inc., 259 F. 3d 229, 245 (2001). The Virginia Court of Appeals found a statement more reliable because the witness was in custody and charged with a crime (thus making the statement more obviously against her penal interest), see Nowlin v. Commonwealth, 40 Va. App. 327, 335-338, 579 S. E. 2d 367, 371-372 (2003), while the Wisconsin Court of Appeals found a statement more reliable because the witness was not in custody and not a suspect, see State v. Bintz, 2002 WI App. 204, ¶ ;13, 257 Wis. 2d 177, 187, 650 N. W. 2d 913, 918. Finally, the Colorado Supreme Court in one case found a statement more reliable because it was given "immediately after" the events at issue, Farrell, supra, at 407, while that same court, in another case, found a statement more reliable because two years had elapsed, Stevens v. People, 29 P. 3d 305, 316 (2001).

[102] The unpardonable vice of the Roberts test, however, is not its unpredictability, but its demonstrated capacity to admit core testimonial statements that the Confrontation Clause plainly meant to exclude. Despite the plurality's speculation in Lilly, 527 U. S., at 137, that it was "highly unlikely" that accomplice confessions implicating the accused could survive Roberts, courts continue routinely to admit them. See Photogrammetric Data Servs., supra, at 245-246; Farrell, supra, at 406-408; Stevens, supra, at 314-318; Taylor v. Commonwealth, 63 S. W. 3d 151, 166-168 (Ky. 2001); State v. Hawkins, No. 2001-P-0060, 2002 WL 31895118, ¶ ;¶ ;34-37, *6 (Ohio App., Dec. 31, 2002); Bintz, supra, ¶ ;¶ ;7-14, 257 Wis. 2d, at 183-188, 650 N. W. 2d, at 916-918; People v. Lawrence, 55 P. 3d 155, 160-161 (Colo. App. 2001); State v. Jones, 171 Ore. App. 375, 387-391, 15 P. 3d 616, 623-625 (2000); State v. Marshall, 136 Ohio App. 3d 742, 747-748, 737 N. E. 2d 1005, 1009 (2000); People v. Schutte, 240 Mich. App. 713, 718-721, 613 N. W. 2d 370, 376-377 (2000); People v. Thomas, 313 Ill. App. 3d 998, 1005-1007, 730 N. E. 2d 618, 625-626 (2000); cf. Nowlin, supra, at 335-338, 579 S. E. 2d, at 371-372 (witness confessed to a related crime); People v. Campbell, 309 Ill. App. 3d 423, 431-432, 721 N. E. 2d 1225, 1230 (1999) (same). One recent study found that, after Lilly, appellate courts admitted accomplice statements to the authorities in 25 out of 70 cases -- more than one-third of the time. Kirst, Appellate Court Answers to the Confrontation Questions in Lilly v. Virginia, 53 Syracuse L. Rev. 87, 105 (2003). Courts have invoked Roberts to admit other sorts of plainly testimonial statements despite the absence of any opportunity to cross-examine. See United States v. Aguilar, 295 F. 3d 1018, 1021-1023 (CA9 2002) (plea allocution showing existence of a conspiracy); United States v. Centracchio, 265 F. 3d 518, 527-530 (CA7 2001) (same); United States v. Dolah, 245 F. 3d 98, 104-105 (CA2 2001) (same); United States v. Petrillo, 237 F. 3d 119, 122-123 (CA2 2000) (same); United States v. Moskowitz, 215 F. 3d 265, 268-269 (CA2 2000) (same); United States v. Gallego, 191 F. 3d 156, 166-168 (CA2 1999) (same); United States v. Papajohn, 212 F. 3d 1112, 1118-1120 (CA8 2000) (grand jury testimony); United States v. Thomas, 30 Fed. Appx. 277, 279 (CA4 2002) (same); Bintz, supra, ¶ ;¶ ;15-22, 257 Wis. 2d, at 188-191, 650 N. W. 2d, at 918-920 (prior trial testimony); State v. McNeill, 140 N. C. App. 450, 457-460, 537 S. E. 2d 518, 523-524 (2000) (same).

[103] To add insult to injury, some of the courts that admit untested testimonial statements find reliability in the very factors that make the statements testimonial. As noted earlier, one court relied on the fact that the witness's statement was made to police while in custody on pending charges -- the theory being that this made the statement more clearly against penal interest and thus more reliable. Nowlin, supra, at 335-338, 579 S. E. 2d, at 371-372. Other courts routinely rely on the fact that a prior statement is given under oath in judicial proceedings. E.g., Gallego, supra, at 168 (plea allocution); Papajohn, supra, at 1120 (grand jury testimony). That inculpating statements are given in a testimonial setting is not an antidote to the confrontation problem, but rather the trigger that makes the Clause's demands most urgent. It is not enough to point out that most of the usual safeguards of the adversary process attend the statement, when the single safeguard missing is the one the Confrontation Clause demands.

[104] C.

[105] Roberts' failings were on full display in the proceedings below. Sylvia Crawford made her statement while in police custody, herself a potential suspect in the case. Indeed, she had been told that whether she would be released "depend[ed] on how the investigation continues." App. 81. In response to often leading questions from police detectives, she implicated her husband in Lee's stabbing and at least arguably undermined his self-defense claim. Despite all this, the trial court admitted her statement, listing several reasons why it was reliable. In its opinion reversing, the Court of Appeals listed several other reasons why the statement was not reliable. Finally, the State Supreme Court relied exclusively on the interlocking character of the statement and disregarded every other factor the lower courts had considered. The case is thus a self-contained demonstration of Roberts' unpredictable and inconsistent application.

[106] Each of the courts also made assumptions that cross-examination might well have undermined. The trial court, for example, stated that Sylvia Crawford's statement was reliable because she was an eyewitness with direct knowledge of the events. But Sylvia at one point told the police that she had "shut [her] eyes and ... didn't really watch" part of the fight, and that she was "in shock." App. 134. The trial court also buttressed its reliability finding by claiming that Sylvia was "being questioned by law enforcement, and, thus, the [questioner] is ... neutral to her and not someone who would be inclined to advance her interests and shade her version of the truth unfavorably toward the defendant." Id., at 77. The Framers would be astounded to learn that ex parte testimony could be admitted against a criminal defendant because it was elicited by "neutral" government officers. But even if the court's assessment of the officer's motives was accurate, it says nothing about Sylvia's perception of her situation. Only cross-examination could reveal that.

[107] The State Supreme Court gave dispositive weight to the interlocking nature of the two statements -- that they were both ambiguous as to when and whether Lee had a weapon. The court's claim that the two statements were equally ambiguous is hard to accept. Petitioner's statement is ambiguous only in the sense that he had lingering doubts about his recollection: "A. I coulda swore I seen him goin' for somethin' before, right before everything happened... . [B]ut I'm not positive." Id., at 155. Sylvia's statement, on the other hand, is truly inscrutable, since the key timing detail was simply assumed in the leading question she was asked: "Q. Did Kenny do anything to fight back from this assault?" Id., at 137. Moreover, Sylvia specifically said Lee had nothing in his hands after he was stabbed, while petitioner was not asked about that.

[108] The prosecutor obviously did not share the court's view that Sylvia's statement was ambiguous -- he called it "damning evidence" that "completely refutes [petitioner's] claim of self-defense." Tr. 468 (Oct. 21, 1999). We have no way of knowing whether the jury agreed with the prosecutor or the court. Far from obviating the need for cross-examination, the "interlocking" ambiguity of the two statements made it all the more imperative that they be tested to tease out the truth.

[109] We readily concede that we could resolve this case by simply reweighing the "reliability factors" under Roberts and finding that Sylvia Crawford's statement falls short. But we view this as one of those rare cases in which the result below is so improbable that it reveals a fundamental failure on our part to interpret the Constitution in a way that secures its intended constraint on judicial discretion. Moreover, to reverse the Washington Supreme Court's decision after conducting our own reliability analysis would perpetuate, not avoid, what the Sixth Amendment condemns. The Constitution prescribes a procedure for determining the reliability of testimony in criminal trials, and we, no less than the state courts, lack authority to replace it with one of our own devising.

[110] We have no doubt that the courts below were acting in utmost good faith when they found reliability. The Framers, however, would not have been content to indulge this assumption. They knew that judges, like other government officers, could not always be trusted to safeguard the rights of the people; the likes of the dread Lord Jeffreys were not yet too distant a memory. They were loath to leave too much discretion in judicial hands. Cf. U. S. Const., Amdt. 6 (criminal jury trial); Amdt. 7 (civil jury trial); Ring v. Arizona, 536 U. S. 584, 611-612 (2002) (Scalia, J., concurring). By replacing categorical constitutional guarantees with open-ended balancing tests, we do violence to their design. Vague standards are manipulable, and, while that might be a small concern in run-of-the-mill assault prosecutions like this one, the Framers had an eye toward politically charged cases like Raleigh's -- great state trials where the impartiality of even those at the highest levels of the judiciary might not be so clear. It is difficult to imagine Roberts' providing any meaningful protection in those circumstances.

[111] Where non-testimonial hearsay is at issue, it is wholly consistent with the Framers' design to afford the States flexibility in their development of hearsay law -- as does Roberts, and as would an approach that exempted such statements from Confrontation Clause scrutiny altogether. Where testimonial evidence is at issue, however, the Sixth Amendment demands what the common law required: unavailability and a prior opportunity for cross-examination. We leave for another day any effort to spell out a comprehensive definition of "testimonial."*fn10 Whatever else the term covers, it applies at a minimum to prior testimony at a preliminary hearing, before a grand jury, or at a former trial; and to police interrogations. These are the modern practices with closest kinship to the abuses at which the Confrontation Clause was directed.

[112] In this case, the State admitted Sylvia's testimonial statement against petitioner, despite the fact that he had no opportunity to cross-examine her. That alone is sufficient to make out a violation of the Sixth Amendment. Roberts notwithstanding, we decline to mine the record in search of indicia of reliability. Where testimonial statements are at issue, the only indicium of reliability sufficient to satisfy constitutional demands is the one the Constitution actually prescribes: confrontation.

[113] The judgment of the Washington Supreme Court is reversed, and the case is remanded for further proceedings not inconsistent with this opinion.

[114] It is so ordered.

[115] Rehnquist, C. J., concurring in judgment

[116] Chief Justice Rehnquist, with whom Justice O'Connor joins, concurring in the judgment.

[117] I dissent from the Court's decision to overrule Ohio v. Roberts, 448 U. S. 56 (1980). I believe that the Court's adoption of a new interpretation of the Confrontation Clause is not backed by sufficiently persuasive reasoning to overrule long-established precedent. Its decision casts a mantle of uncertainty over future criminal trials in both federal and state courts, and is by no means necessary to decide the present case.

[118] The Court's distinction between testimonial and non-testimonial statements, contrary to its claim, is no better rooted in history than our current doctrine. Under the common law, although the courts were far from consistent, out-of-court statements made by someone other than the accused and not taken under oath, unlike ex parte depositions or affidavits, were generally not considered substantive evidence upon which a conviction could be based.*fn11 See, e.g., King v. Brasier, 1 Leach 199, 200, 168 Eng. Rep. 202 (K. B. 1779); see also J. Langbein, Origins of Adversary Criminal Trial 235-242 (2003); G. Gilbert, Evidence 152 (3d ed 1769).*fn12 Testimonial statements such as accusatory statements to police officers likely would have been disapproved of in the 18th century, not necessarily because they resembled ex parte affidavits or depositions as the Court reasons, but more likely than not because they were not made under oath.*fn13 See King v. Woodcock, 1 Leach 500, 503, 168 Eng. Rep. 352, 353 (1789) (noting that a statement taken by a justice of the peace may not be admitted into evidence unless taken under oath). Without an oath, one usually did not get to the second step of whether confrontation was required.

[119] Thus, while I agree that the Framers were mainly concerned about sworn affidavits and depositions, it does not follow that they were similarly concerned about the Court's broader category of testimonial statements. See 1 N. Webster, An American Dictionary of the English Language (1828) (defining "Testimony" as "[a] solemn declaration or affirmation made for the purpose of establishing or proving some fact. Such affirmation in judicial proceedings, may be verbal or written, but must be under oath" (emphasis added)). As far as I can tell, unsworn testimonial statements were treated no differently at common law than were non-testimonial statements, and it seems to me any classification of statements as testimonial beyond that of sworn affidavits and depositions will be somewhat arbitrary, merely a proxy for what the Framers might have intended had such evidence been liberally admitted as substantive evidence like it is today.*fn14

[120] I therefore see no reason why the distinction the Court draws is preferable to our precedent. Starting with Chief Justice Marshall's interpretation as a Circuit Justice in 1807, 16 years after the ratification of the Sixth Amendment, United States v. Burr, 25 F. Cas. 187, 193 (No. 14,694) (CC Va. 1807), continuing with our cases in the late 19th century, Mattox v. United States, 156 U. S. 237, 243-244 (1895); Kirby v. United States, 174 U. S. 47, 54-57 (1899), and through today, e.g., White v. Illinois, 502 U. S. 346, 352-353 (1992), we have never drawn a distinction between testimonial and non-testimonial statements. And for that matter, neither has any other court of which I am aware. I see little value in trading our precedent for an imprecise approximation at this late date.

[121] I am also not convinced that the Confrontation Clause categorically requires the exclusion of testimonial statements. Although many States had their own Confrontation Clauses, they were of recent vintage and were not interpreted with any regularity before 1791. State cases that recently followed the ratification of the Sixth Amendment were not uniform; the Court itself cites state cases from the early 19th century that took a more stringent view of the right to confrontation than does the Court, prohibiting former testimony even if the witness was subjected to cross-examination. See ante, at 13 (citing Finn v. Commonwealth, 26 Va. 701, 708 (1827); State v. Atkins, 1 Tenn. 229 (1807) (per curiam)).

[122] Nor was the English law at the time of the framing entirely consistent in its treatment of testimonial evidence. Generally ex parte affidavits and depositions were excluded as the Court notes, but even that proposition was not universal. See King v. Eriswell, 3 T. R. 707, 100 Eng. Rep. 815 (K. B. 1790) (affirming by an equally divided court the admission of an ex parte examination because the declarant was unavailable to testify); King v. Westbeer, 1 Leach 12, 13, 168 Eng. Rep. 108, 109 (1739) (noting the admission of an ex parte affidavit); see also 1 M. Hale, Pleas of the Crown 585-586 (1736) (noting that statements of "accusers and witnesses" which were taken under oath could be admitted into evidence if the declarant was "dead or not able to travel"). Wigmore notes that sworn examinations of witnesses before justices of the peace in certain cases would not have been excluded until the end of the 1700's, 5 Wigmore §1364, at 26-27, and sworn statements of witnesses before coroners became excluded only by statute in the 1800's, see ibid.; id., §1374, at 59. With respect to unsworn testimonial statements, there is no indication that once the hearsay rule was developed courts ever excluded these statements if they otherwise fell within a firmly rooted exception. See, e.g., Eriswell, supra, at 715-719 (Buller, J.), 720 (Ashhurst, J.), 100 Eng. Rep., at 819-822 (concluding that an ex parte examination was admissible as an exception to the hearsay rule because it was a declaration by a party of his state and condition). Dying declarations are one example. See, e.g., Woodcock, supra, at 502-504, 168 Eng. Rep., at 353-354; King v. Reason, 16 How. St. Tr. 1, 22-23 (K. B. 1722).

[123] Between 1700 and 1800 the rules regarding the admissibility of out-of-court statements were still being developed. See n. 1, supra. There were always exceptions to the general rule of exclusion, and it is not clear to me that the Framers categorically wanted to eliminate further ones. It is one thing to trace the right of confrontation back to the Roman Empire; it is quite another to conclude that such a right absolutely excludes a large category of evidence. It is an odd conclusion indeed to think that the Framers created a cut-and-dried rule with respect to the admissibility of testimonial statements when the law during their own time was not fully settled.

[124] To find exceptions to exclusion under the Clause is not to denigrate it as the Court suggests. Chief Justice Marshall stated of the Confrontation Clause: "I know of no principle in the preservation of which all are more concerned. I know none, by undermining which, life, liberty and property, might be more endangered. It is therefore incumbent on courts to be watchful of every inroad on a principle so truly important." Burr, 25 F. Cas., at 193. Yet, he recognized that such a right was not absolute, acknowledging that exceptions to the exclusionary component of the hearsay rule, which he considered as an "inroad" on the right to confrontation, had been introduced. See ibid.

[125] Exceptions to confrontation have always been derived from the experience that some out-of-court statements are just as reliable as cross-examined in-court testimony due to the circumstances under which they were made. We have recognized, for example, that co-conspirator statements simply "cannot be replicated, even if the declarant testifies to the same matters in court." United States v. Inadi, 475 U. S. 387, 395 (1986). Because the statements are made while the declarant and the accused are partners in an illegal enterprise, the statements are unlikely to be false and their admission "actually furthers the `Confrontation Clause's very mission' which is to `advance the accuracy of the truth-determining process in criminal trials.' " Id., at 396 (quoting Tennessee v. Street, 471 U. S. 409, 415 (1985) (some internal quotation marks omitted)). Similar reasons justify the introduction of spontaneous declarations, see White, 502 U. S., at 356, statements made in the course of procuring medical services, see ibid., dying declarations, see Kirby, supra, at 61, and countless other hearsay exceptions. That a statement might be testimonial does nothing to undermine the wisdom of one of these exceptions.

[126] Indeed, cross-examination is a tool used to flesh out the truth, not an empty procedure. See Kentucky v. Stincer, 482 U. S. 730, 737 (1987) ("The right to cross-examination, protected by the Confrontation Clause, thus is essentially a `functional' right designed to promote reliability in the truth-finding functions of a criminal trial"); see also Maryland v. Craig, 497 U. S. 836, 845 (1990) ("The central concern of the Confrontation Clause is to ensure the reliability of the evidence against a criminal defendant by subjecting it to rigorous testing in the context of an adversary proceeding before the trier of fact"). "[I]n a given instance [cross-examination may] be superfluous; it may be sufficiently clear, in that instance, that the statement offered is free enough from the risk of inaccuracy and untrustworthiness, so that the test of cross-examination would be a work of supererogation." 5 Wigmore §1420, at 251. In such a case, as we noted over 100 years ago, "The law in its wisdom declares that the rights of the public shall not be wholly sacrificed in order that an incidental benefit may be preserved to the accused." Mattox, 156 U. S., at 243; see also Salinger v. United States, 272 U. S. 542, 548 (1926). By creating an immutable category of excluded evidence, the Court adds little to a trial's truth-finding function and ignores this longstanding guidance.

[127] In choosing the path it does, the Court of course overrules Ohio v. Roberts, 448 U. S. 56 (1980), a case decided nearly a quarter of a century ago. Stare decisis is not an inexorable command in the area of constitutional law, see Payne v. Tennessee, 501 U. S. 808, 828 (1991), but by and large, it "is the preferred course because it promotes the evenhanded, predictable, and consistent development of legal principles, fosters reliance on judicial decisions, and contributes to the actual and perceived integrity of the judicial process," id., at 827. And in making this appraisal, doubt that the new rule is indeed the "right" one should surely be weighed in the balance. Though there are no vested interests involved, unresolved questions for the future of everyday criminal trials throughout the country surely counsel the same sort of caution. The Court grandly declares that "[w]e leave for another day any effort to spell out a comprehensive definition of `testimonial,' " ante, at 33. But the thousands of federal prosecutors and the tens of thousands of state prosecutors need answers as to what beyond the specific kinds of "testimony" the Court lists, see ibid., is covered by the new rule. They need them now, not months or years from now. Rules of criminal evidence are applied every day in courts throughout the country, and parties should not be left in the dark in this manner.

[128] To its credit, the Court's analysis of "testimony" excludes at least some hearsay exceptions, such as business records and official records. See ante, at 20. To hold otherwise would require numerous additional witnesses without any apparent gain in the truth-seeking process. Likewise to the Court's credit is its implicit recognition that the mistaken application of its new rule by courts which guess wrong as to the scope of the rule is subject to harmless-error analysis. See ante, at 5, n. 1.

[129] But these are palliatives to what I believe is a mistaken change of course. It is a change of course not in the least necessary to reverse the judgment of the Supreme Court of Washington in this case. The result the Court reaches follows inexorably from Roberts and its progeny without any need for overruling that line of cases. In Idaho v. Wright, 497 U. S. 805, 820-824 (1990), we held that an out-of-court statement was not admissible simply because the truthfulness of that statement was corroborated by other evidence at trial. As the Court notes, ante, at 31, the Supreme Court of Washington gave decisive weight to the "interlocking nature of the two statements." No re-weighing of the "reliability factors," which is hypothesized by the Court, ante, at 31, is required to reverse the judgment here. A citation to Idaho v. Wright, supra, would suffice. For the reasons stated, I believe that this would be a far preferable course for the Court to take here.


Opinion Footnotes


[130] *fn1 The court rejected the State's argument that guarantees of trustworthiness were unnecessary since petitioner waived his confrontation rights by invoking the marital privilege. It reasoned that "forcing the defendant to choose between the marital privilege and confronting his spouse presents an untenable Hobson's choice." 147 Wash. 2d, at 432, 54 P. 3d, at 660. The State has not challenged this holding here. The State also has not challenged the Court of Appeals' conclusion (not reached by the State Supreme Court) that the confrontation violation, if it occurred, was not harmless. We express no opinion on these matters.

[131] *fn2 There is some question whether the requirement of a prior opportunity for cross-examination applied as well to statements taken by a coroner, which were also authorized by the Marian statutes. See 3 Wigmore §1364, at 23 (requirement "never came to be conceded at all in England"); T. Peake, Evidence 64, n. (m) (3d ed. 1808) (not finding the point "expressly decided in any reported case"); State v. Houser, 26 Mo. 431, 436 (1858) ("there may be a few cases ... but the authority of such cases is questioned, even in [England], by their ablest writers on common law"); State v. Campbell, 1 S. C. 124, 130 (1844) (point "has not ... been plainly adjudged, even in the English cases"). Whatever the English rule, several early American authorities flatly rejected any special status for coroner statements. See Houser, supra, at 436; Campbell, supra, at 130; T. Cooley, Constitutional Limitations *318.

[132] *fn3 These sources -- especially Raleigh's trial -- refute The Chief Justice's assertion, post, at 3 (opinion concurring in judgment), that the right of confrontation was not particularly concerned with unsworn testimonial statements. But even if, as he claims, a general bar on unsworn hearsay made application of the Confrontation Clause to unsworn testimonial statements a moot point, that would merely change our focus from direct evidence of original meaning of the Sixth Amendment to reasonable inference. We find it implausible that a provision which concededly condemned trial by sworn ex parte affidavit thought trial by unsworn ex parte affidavit perfectly OK. (The claim that unsworn testimony was self-regulating because jurors would disbelieve it, cf. post, at 2, n. 1, is belied by the very existence of a general bar on unsworn testimony.) Any attempt to determine the application of a constitutional provision to a phenomenon that did not exist at the time of its adoption (here, allegedly, admissible unsworn testimony) involves some degree of estimation -- what The Chief Justice calls use of a "proxy," post, at 3 -- but that is hardly a reason not to make the estimation as accurate as possible. Even if, as The Chief Justice mistakenly asserts, there were no direct evidence of how the Sixth Amendment originally applied to unsworn testimony, there is no doubt what its application would have been.

[133] *fn4 We use the term "interrogation" in its colloquial, rather than any technical legal, sense. Cf. Rhode Island v. Innis, 446 U. S. 291, 300-301 (1980). Just as various definitions of "testimonial" exist, one can imagine various definitions of "interrogation," and we need not select among them in this case. Sylvia's recorded statement, knowingly given in response to structured police questioning, qualifies under any conceivable definition.

[134] *fn5 The Chief Justice claims that English law's treatment of testimonial statements was inconsistent at the time of the framing, post, at 4-5, but the examples he cites relate to examinations under the Marian statutes. As we have explained, to the extent Marian examinations were admissible, it was only because the statutes derogated from the common law. See supra, at 10. Moreover, by 1791 even the statutory-derogation view had been rejected with respect to justice-of-the-peace examinations -- explicitly in King v. Woodcock, 1 Leach 500, 502-504, 168 Eng. Rep. 352, 353 (1789), and King v. Dingler, 2 Leach 561, 562-563, 168 Eng. Rep. 383, 383-384 (1791), and by implication in King v. Radbourne, 1 Leach 457, 459-461, 168 Eng. Rep. 330, 331-332 (1787). None of The Chief Justice's citations proves otherwise. King v. Westbeer, 1 Leach 12, 168 Eng. Rep. 108 (1739), was decided a half-century earlier and cannot be taken as an accurate statement of the law in 1791 given the directly contrary holdings of Woodcock and Dingler. Hale's treatise is older still, and far more ambiguous on this point, see 1 M. Hale, Pleas of the Crown 585-586 (1736); some who espoused the requirement of a prior opportunity for cross-examination thought it entirely consistent with Hale's views. See Fenwick's Case, 13 How. St. Tr. 537, 602 (H. C. 1696) (Musgrave). The only timely authority The Chief Justice cites is King v. Eriswell, 3 T. R. 707, 100 Eng. Rep. 815 (K. B. 1790), but even that decision provides no substantial support. Eriswell was not a criminal case at all, but a Crown suit against the inhabitants of a town to charge them with care of an insane pauper. Id., at 707-708, 100 Eng. Rep., at 815-816. It is relevant only because the judges discuss the Marian statutes in dicta. One of them, Buller, J., defended admission of the pauper's statement of residence on the basis of authorities that purportedly held ex parte Marian examinations admissible. Id., at 713-714, 100 Eng. Rep., at 819. As evidence writers were quick to point out, however, his authorities said no such thing. See Peake, Evidence, at 64, n. (m) ("Mr. J. Buller is reported to have said that it was so settled in 1 Lev. 180, and Kel. 55; certainly nothing of the kind appears in those books"); 2 T. Starkie, Evidence 487-488, n. (c) (1826) ("Buller, J. ... refers to Radbourne's case ... ; but in that case the deposition was taken in the hearing of the prisoner, and of course the question did not arise" (citation omitted)). Two other judges, Grose, J., and Kenyon, C. J., responded to Buller's argument by distinguishing Marian examinations as a statutory exception to the common-law rule, but the context and tenor of their remarks suggest they merely assumed the accuracy of Buller's premise without independent consideration, at least with respect to examinations by justices of the peace. See 3 T. R., at 710, 100 Eng. Rep., at 817 (Grose, J.); id., at 722-723, 100 Eng. Rep., at 823-824 (Kenyon, C. J.). In fact, the case reporter specifically notes in a footnote that their assumption was erroneous. See id., at 710, n. (c), 100 Eng. Rep., at 817, n. (c). Notably, Buller's position on pauper examinations was resoundingly rejected only a decade later in King v. Ferry Frystone, 2 East 54, 55, 102 Eng. Rep. 289 (K. B. 1801) ("The point ... has been since considered to be so clear against the admissibility of the evidence ... that it was abandoned by the counsel ... without argument"), further suggesting that his views on evidence were not mainstream at the time of the framing. In short, none of The Chief Justice's sources shows that the law in 1791 was unsettled even as to examinations by justices of the peace under the Marian statutes. More importantly, however, even if the statutory rule in 1791 were in doubt, the numerous early state-court decisions make abundantly clear that the Sixth Amendment incorporated the common-law right of confrontation and not any exceptions the Marian statutes supposedly carved out from it. See supra, at 13-14; see also supra, at 11, n. 2 (coroner statements). The common-law rule had been settled since Paine in 1696. See King v. Paine, 5 Mod. 163, 165, 87 Eng. Rep. 584, 585 (K. B.).

[135] *fn6 The one deviation we have found involves dying declarations. The existence of that exception as a general rule of criminal hearsay law cannot be disputed. See, e.g., Mattox v. United States, 156 U. S. 237, 243-244 (1895); King v. Reason, 16 How. St. Tr. 1, 24-38 (K. B. 1722); 1 D. Jardine, Criminal Trials 435 (1832); Cooley, Constitutional Limitations, at *318; 1 G. Gilbert, Evidence 211 (C. Lofft ed. 1791); see also F. Heller, The Sixth Amendment 105 (1951) (asserting that this was the only recognized criminal hearsay exception at common law). Although many dying declarations may not be testimonial, there is authority for admitting even those that clearly are. See Woodcock, supra, at 501-504, 168 Eng. Rep., at 353-354; Reason, supra, at 24-38; Peake, Evidence, at 64; cf. Radbourne, supra, at 460-462, 168 Eng. Rep., at 332-333. We need not decide in this case whether the Sixth Amendment incorporates an exception for testimonial dying declarations. If this exception must be accepted on historical grounds, it is sui generis.

[136] *fn7 We cannot agree with The Chief Justice that the fact "[t]hat a statement might be testimonial does nothing to undermine the wisdom of one of these [hearsay] exceptions." Post, at 6. Involvement of government officers in the production of testimony with an eye toward trial presents unique potential for prosecutorial abuse -- a fact borne out time and again throughout a history with which the Framers were keenly familiar. This consideration does not evaporate when testimony happens to fall within some broad, modern hearsay exception, even if that exception might be justifiable in other circumstances.

[137] *fn8 One case arguably in tension with the rule requiring a prior opportunity for cross-examination when the proffered statement is testimonial is White v. Illinois, 502 U. S. 346 (1992), which involved, inter alia, statements of a child victim to an investigating police officer admitted as spontaneous declarations. Id., at 349-351. It is questionable whether testimonial statements would ever have been admissible on that ground in 1791; to the extent the hearsay exception for spontaneous declarations existed at all, it required that the statements be made "immediat[ely] upon the hurt received, and before [the declarant] had time to devise or contrive any thing for her own advantage." Thompson v. Trevanion, Skin. 402, 90 Eng. Rep. 179 (K. B. 1694). In any case, the only question presented in White was whether the Confrontation Clause imposed an unavailability requirement on the types of hearsay at issue. See 502 U. S., at 348-349. The holding did not address the question whether certain of the statements, because they were testimonial, had to be excluded even if the witness was unavailable. We "[took] as a given ... that the testimony properly falls within the relevant hearsay exceptions." Id., at 351, n. 4.

[138] *fn9 The Chief Justice complains that our prior decisions have "never drawn a distinction" like the one we now draw, citing in particular Mattox v. United States, 156 U. S. 237 (1895), Kirby v. United States, 174 U. S. 47 (1899), and United States v. Burr, 25 F. Cas. 187 (No. 14,694) (CC Va. 1807) (Marshall, C. J.). Post, at 4-6. But nothing in these cases contradicts our holding in any way. Mattox and Kirby allowed or excluded evidence depending on whether the defendant had had an opportunity for cross-examination. Mattox, supra, at 242-244; Kirby, supra, at 55-61. That the two cases did not extrapolate a more general class of evidence to which that criterion applied does not prevent us from doing so now. As to Burr, we disagree with The Chief Justice's reading of the case. Although Chief Justice Marshall made one passing reference to the Confrontation Clause, the case was fundamentally about the hearsay rules governing statements in furtherance of a conspiracy. The "principle so truly important" on which "inroad[s]" had been introduced was the "rule of evidence which rejects mere hearsay testimony." See 25 F. Cas., at 193. Nothing in the opinion concedes exceptions to the Confrontation Clause's exclusion of testimonial statements as we use the term. The Chief Justice fails to identify a single case (aside from one minor, arguable exception, see supra, at 22, n. 8), where we have admitted testimonial statements based on indicia of reliability other than a prior opportunity for cross-examination. If nothing else, the test we announce is an empirically accurate explanation of the results our cases have reached. Finally, we reiterate that, when the declarant appears for cross-examination at trial, the Confrontation Clause places no constraints at all on the use of his prior testimonial statements. See California v. Green, 399 U. S. 149, 162 (1970). It is therefore irrelevant that the reliability of some out-of-court statements " `cannot be replicated, even if the declarant testifies to the same matters in court.' " Post, at 6 (quoting United States v. Inadi, 475 U. S. 387, 395 (1986)). The Clause does not bar admission of a statement so long as the declarant is present at trial to defend or explain it. (The Clause also does not bar the use of testimonial statements for purposes other than establishing the truth of the matter asserted. See Tennessee v. Street, 471 U. S. 409, 414 (1985).)

[139] *fn10 We acknowledge The Chief Justice's objection, post, at 7-8, that our refusal to articulate a comprehensive definition in this case will cause interim uncertainty. But it can hardly be any worse than the status quo. See supra, at 27-30, and cases cited. The difference is that the Roberts test is inherently, and therefore permanently, unpredictable.

[140] *fn11 Modern scholars have concluded that at the time of the founding the law had yet to fully develop the exclusionary component of the hearsay rule and its attendant exceptions, and thus hearsay was still often heard by the jury. See Gallanis, The Rise of Modern Evidence Law, 84 Iowa L. Rev. 499, 534-535 (1999); Mosteller, Remaking Confrontation Clause and Hearsay Doctrine Under the Challenge of Child Sexual Abuse Prosecutions, 1993 U. Ill. L. Rev. 691, 738-746. In many cases, hearsay alone was generally not considered sufficient to support a conviction; rather, it was used to corroborate sworn witness testimony. See 5 J. Wigmore, Evidence, §1364, pp. 17, 19-20, 19, n. 33 (J. Chadbourn rev. 1974) (hereinafter Wigmore) (noting in the 1600's and early 1700's testimonial and non-testimonial hearsay was permissible to corroborate direct testimony); see also J. Langbein, Origins of Adversary Criminal Trial 238-239 (2003). Even when unsworn hearsay was proffered as substantive evidence, however, because of the predominance of the oath in society, juries were largely skeptical of it. See Landsman, Rise of the Contentious Spirit: Adversary Procedure in Eighteenth Century England, 75 Cornell L. Rev. 497, 506 (1990) (describing late 17th-century sentiments); Langbein, Criminal Trial before the Lawyers, 45 U. Chi. L. Rev. 263, 291-293 (1978). In the 18th century, unsworn hearsay was simply held to be of much lesser value than were sworn affidavits or depositions.

[141] *fn12 Gilbert's noted in 1769: "Hearsay is no Evidence ... though a Person Testify what he hath heard upon Oath, yet the Person who spake it was not upon Oath; and if a Man had been in Court and said the same Thing and had not sworn it, he had not been believed in a Court of Justice; for all Credit being derived from Attestation and Evidence, it can rise no higher than the Fountain from whence it flows, and if the first Speech was without Oath, an Oath that there was such a Speech makes it no more than a bare speaking, and so of no Value in a Court of Justice, where all Things were determined under the Solemnities of an Oath ... ."

[142] *fn13 Confessions not taken under oath were admissible against a confessor because " `the most obvious Principles of Justice, Policy, and Humanity' " prohibited an accused from attesting to his statements. 1 G. Gilbert, Evidence 216 (C. Lofft ed. 1791). Still, these unsworn confessions were considered evidence only against the confessor as the Court points out, see ante, at 16, and in cases of treason, were insufficient to support even the conviction of the confessor, 2 W. Hawkins, Pleas of the Crown, C. 46, §4, p. 604, n. 3 (T. Leach 6th ed. 1787).

[143] *fn14 The fact that the prosecution introduced an unsworn examination in 1603 at Sir Walter Raleigh's trial, as the Court notes, see ante, at 16, says little about the Court's distinction between testimonial and non-testimonial statements. Our precedent indicates that unsworn testimonial statements, as do some non-testimonial statements, raise confrontation concerns once admitted into evidence, see, e.g., Lilly v. Virginia, 527 U. S. 116 (1999); Lee v. Illinois, 476 U. S. 530 (1986), and I do not contend otherwise. My point is not that the Confrontation Clause does not reach these statements, but rather that it is far from clear that courts in the late 18th century would have treated unsworn statements, even testimonial ones, the same as sworn statements.

Blakely v. Washington

Blakely v. Washington, 124 S.Ct. 2531, 542 U.S. 296, 159 L.Ed.2d 403 (U.S. 06/24/2004)


[2] No. 02-1632

[3] 124 S.Ct. 2531, 542 U.S. 296, 159 L.Ed.2d 403, 72 USLW 4546, 2004 Daily Journal D.A.R. 7581, 04 Cal. Daily Op. Serv. 5539, 4 Cal. Daily Op. Serv. 5539,

[4] June 24, 2004



[7] OCTOBER TERM, 2003 Argued March 23, 2004

[8] Petitioner pleaded guilty to kidnaping his estranged wife. The facts admitted in his plea, standing alone, supported a maximum sentence of 53 months, but the judge imposed a 90-month sentence after finding that petitioner had acted with deliberate cruelty, a statutorily enumerated ground for departing from the standard range. The Washington Court of Appeals affirmed, rejecting petitioner's argument that the sentencing procedure deprived him of his federal constitutional right to have a jury determine beyond a reasonable doubt all facts legally essential to his sentence.

[9] Held: Because the facts supporting petitioner's exceptional sentence were neither admitted by petitioner nor found by a jury, the sentence violated his Sixth Amendment right to trial by jury. Pp. 5-18.

[10] (a) This case requires the Court to apply the rule of Apprendi v. New Jersey, 530 U. S. 466, 490, that, "[o]ther than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt." The relevant statutory maximum for Apprendi purposes is the maximum a judge may impose based solely on the facts reflected in the jury verdict or admitted by the defendant. Here, the judge could not have imposed the 90-month sentence based solely on the facts admitted in the guilty plea, because Washington law requires an exceptional sentence to be based on factors other than those used in computing the standard-range sentence. Petitioner's sentence is not analogous to those upheld in McMillan v. Pennsylvania, 477 U. S. 79, and Williams v. New York, 337 U. S. 241, which were not greater than what state law authorized based on the verdict alone. Regardless of whether the judge's authority to impose the enhanced sentence depends on a judge's finding a specified fact, one of several specified facts, or any aggravating fact, it remains the case that the jury's verdict alone does not authorize the sentence. Pp. 5-9.

[11] (b) This Court's commitment to Apprendi in this context reflects not just respect for longstanding precedent, but the need to give intelligible content to the fundamental constitutional right of jury trial. Pp. 9-12.

[12] (c) This case is not about the constitutionality of determinate sentencing, but only about how it can be implemented in a way that respects the Sixth Amendment. The Framers' paradigm for criminal justice is the common-law ideal of limited state power accomplished by strict division of authority between judge and jury. That can be preserved without abandoning determinate sentencing and at no sacrifice of fairness to the defendant. Pp. 12-17.

[13] 111 Wash. App. 851, 47 P. 3d 149, reversed and remanded.

[14] Scalia, J., delivered the opinion of the Court, in which Stevens, Souter, Thomas, and Ginsburg, JJ., joined. O'Connor, J., filed a dissenting opinion, in which Breyer, J., joined, and in which Rehnquist, C. J., and Kennedy, J., joined except as to Part IV-B. Kennedy, J., filed a dissenting opinion, in which Breyer, J., joined. Breyer, J., filed a dissenting opinion, in which O'Connor, J., joined.

[15] On Writ Of Certiorari To The Court Of Appeals Of Washington, Division 3 Court Below: 111 Wash. App. 851, 47 P. 3d 149

[16] Jeffrey L. Fisher argued the cause and filed briefs for petitioner.

[17] John D. Knodell III argued the cause and filed a brief for respondent. Deputy Solicitor General Dreeben argued the cause for the United States as amicus curiae urging affirmance. With him on the brief were Solicitor General Olson, Assistant Attorney General Wray, Matthew D. Roberts, and Nina Goodman.

[18] Briefs of amici curiae urging reversal were filed for the American Civil Liberties Union et al. by James E. Lobsenz, Aaron H. Caplan, and Steven R. Shapiro; for the Kansas Appellate Defender Office by Randall L. Hodgkinson; and for the National Association of Criminal Defense Lawyers et al. by David M. Porter and Sheryl Gordon McCloud.

[19] Briefs of amici curiae urging affirmance were filed for the State of Alabama et al. by William H. Pryor, Jr., Attorney General of Alabama, Kevin C. Newsom, Solicitor General, Michael B. Billingsley, Deputy Solicitor General, and Nathan A. Forrester, and by the Attorneys General for their respective States as follows: M. Jane Brady of Delaware, Charles J. Crist, Jr., of Florida, Mark J. Bennett of Hawaii, Lisa Madigan of Illinois, Jon Bruning of Nebraska, Hardy Myers of Oregon, Greg Abbott of Texas, Mark L. Shurtleff of Utah, and Jerry W. Kilgore of Virginia; and for the Criminal Justice Legal Foundation by Kent S. Scheidegger.

[20] The opinion of the court was delivered by: Justice Scalia

[21] 542 U. S. ____ (2004)

[22] Petitioner Ralph Howard Blakely, Jr., pleaded guilty to the kidnaping of his estranged wife. The facts admitted in his plea, standing alone, supported a maximum sentence of 53 months. Pursuant to state law, the court imposed an "exceptional" sentence of 90 months after making a judicial determination that he had acted with "deliberate cruelty." App. 40, 49. We consider whether this violated petitioner's Sixth Amendment right to trial by jury.

[23] I.

[24] Petitioner married his wife Yolanda in 1973. He was evidently a difficult man to live with, having been diagnosed at various times with psychological and personality disorders including paranoid schizophrenia. His wife ultimately filed for divorce. In 1998, he abducted her from their orchard home in Grant County, Washington, binding her with duct tape and forcing her at knifepoint into a wooden box in the bed of his pickup truck. In the process, he implored her to dismiss the divorce suit and related trust proceedings.

[25] When the couple's 13-year-old son Ralphy returned home from school, petitioner ordered him to follow in another car, threatening to harm Yolanda with a shotgun if he did not do so. Ralphy escaped and sought help when they stopped at a gas station, but petitioner continued on with Yolanda to a friend's house in Montana. He was finally arrested after the friend called the police.

[26] The State charged petitioner with first-degree kidnaping, Wash. Rev. Code Ann. §9A.40.020(1) (2000).*fn1 Upon reaching a plea agreement, however, it reduced the charge to second-degree kidnaping involving domestic violence and use of a firearm, see §§9A.40.030(1), 10.99.020(3)(p), 9.94A.125.*fn2 Petitioner entered a guilty plea admitting the elements of second-degree kidnaping and the domestic-violence and firearm allegations, but no other relevant facts.

[27] The case then proceeded to sentencing. In Washington, second-degree kidnaping is a class B felony. §9A.40.030(3). State law provides that "[n]o person convicted of a [class B] felony shall be punished by confinement ... exceeding ... a term of ten years." §9A.20.021(1)(b). Other provisions of state law, however, further limit the range of sentences a judge may impose. Washington's Sentencing Reform Act specifies, for petitioner's offense of second-degree kidnaping with a firearm, a "standard range" of 49 to 53 months. See §9.94A.320 (seriousness level V for second-degree kidnaping); App. 27 (offender score 2 based on §9.94A.360); §9.94A.310(1), box 2-V (standard range of 13-17 months); §9.94A.310(3)(b) (36-month firearm enhancement).*fn3 A judge may impose a sentence above the standard range if he finds "substantial and compelling reasons justifying an exceptional sentence." §9.94A.120(2). The Act lists aggravating factors that justify such a departure, which it recites to be illustrative rather than exhaustive. §9.94A.390. Nevertheless, "[a] reason offered to justify an exceptional sentence can be considered only if it takes into account factors other than those which are used in computing the standard range sentence for the offense." State v. Gore, 143 Wash. 2d 288, 315-316, 21 P. 3d 262, 277 (2001). When a judge imposes an exceptional sentence, he must set forth findings of fact and conclusions of law supporting it. §9.94A.120(3). A reviewing court will reverse the sentence if it finds that "under a clearly erroneous standard there is insufficient evidence in the record to support the reasons for imposing an exceptional sentence." Gore, supra, at 315, 21 P. 3d, at 277 (citing §9.94A.210(4)).

[28] Pursuant to the plea agreement, the State recommended a sentence within the standard range of 49 to 53 months. After hearing Yolanda's description of the kidnaping, however, the judge rejected the State's recommendation and imposed an exceptional sentence of 90 months -- 37 months beyond the standard maximum. He justified the sentence on the ground that petitioner had acted with "deliberate cruelty," a statutorily enumerated ground for departure in domestic-violence cases. §9.94A.390(2)(h)(iii).*fn4

[29] Faced with an unexpected increase of more than three years in his sentence, petitioner objected. The judge accordingly conducted a 3-day bench hearing featuring testimony from petitioner, Yolanda, Ralphy, a police officer, and medical experts. After the hearing, he issued 32 findings of fact, concluding:

[30] "The defendant's motivation to commit kidnapping was complex, contributed to by his mental condition and personality disorders, the pressures of the divorce litigation, the impending trust litigation trial and anger over his troubled interpersonal relationships with his spouse and children. While he misguidedly intended to forcefully reunite his family, his attempt to do so was subservient to his desire to terminate lawsuits and modify title ownerships to his benefit.

[31] "The defendant's methods were more homogeneous than his motive. He used stealth and surprise, and took advantage of the victim's isolation. He immediately employed physical violence, restrained the victim with tape, and threatened her with injury and death to herself and others. He immediately coerced the victim into providing information by the threatening application of a knife. He violated a subsisting restraining order." App. 48-49.

[32] The judge adhered to his initial determination of deliberate cruelty.

[33] Petitioner appealed, arguing that this sentencing procedure deprived him of his federal constitutional right to have a jury determine beyond a reasonable doubt all facts legally essential to his sentence. The State Court of Appeals affirmed, 111 Wash. App. 851, 870-871, 47 P. 3d 149, 159 (2002), relying on the Washington Supreme Court's rejection of a similar challenge in Gore, supra, at 311-315, 21 P. 3d, at 275-277. The Washington Supreme Court denied discretionary review. 148 Wash. 2d 1010, 62 P. 3d 889 (2003). We granted certiorari. 540 U. S. 965 (2003).

[34] II.

[35] This case requires us to apply the rule we expressed in Apprendi v. New Jersey, 530 U. S. 466, 490 (2000): "Other than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt." This rule reflects two longstanding tenets of common-law criminal jurisprudence: that the "truth of every accusation" against a defendant "should afterwards be confirmed by the unanimous suffrage of twelve of his equals and neighbours," 4 W. Blackstone, Commentaries on the Laws of England 343 (1769), and that "an accusation which lacks any particular fact which the law makes essential to the punishment is ... no accusation within the requirements of the common law, and it is no accusation in reason," 1 J. Bishop, Criminal Procedure §87, p. 55 (2d ed. 1872).*fn5 These principles have been acknowledged by courts and treatises since the earliest days of graduated sentencing; we compiled the relevant authorities in Apprendi, see 530 U. S., at 476-483, 489-490, n. 15; id., at 501-518 (Thomas, J., concurring), and need not repeat them here.*fn6

[36] Apprendi involved a New Jersey hate-crime statute that authorized a 20-year sentence, despite the usual 10-year maximum, if the judge found the crime to have been committed " `with a purpose to intimidate ... because of race, color, gender, handicap, religion, sexual orientation or ethnicity.' " Id., at 468-469 (quoting N. J. Stat. Ann. §2C:44-3(e) (West Supp. 1999-2000)). In Ring v. Arizona, 536 U. S. 584, 592-593, and n. 1 (2002), we applied Apprendi to an Arizona law that authorized the death penalty if the judge found one of ten aggravating factors. In each case, we concluded that the defendant's constitutional rights had been violated because the judge had imposed a sentence greater than the maximum he could have imposed under state law without the challenged factual finding. Apprendi, supra, at 491-497; Ring, supra, at 603-609.

[37] In this case, petitioner was sentenced to more than three years above the 53-month statutory maximum of the standard range because he had acted with "deliberate cruelty." The facts supporting that finding were neither admitted by petitioner nor found by a jury. The State nevertheless contends that there was no Apprendi violation because the relevant "statutory maximum" is not 53 months, but the 10-year maximum for class B felonies in §9A.20.021(1)(b). It observes that no exceptional sentence may exceed that limit. See §9.94A.420. Our precedents make clear, however, that the "statutory maximum" for Apprendi purposes is the maximum sentence a judge may impose solely on the basis of the facts reflected in the jury verdict or admitted by the defendant. See Ring, supra, at 602 (" `the maximum he would receive if punished according to the facts reflected in the jury verdict alone' " (quoting Apprendi, supra, at 483)); Harris v. United States, 536 U. S. 545, 563 (2002) (plurality opinion) (same); cf. Apprendi, supra, at 488 (facts admitted by the defendant). In other words, the relevant "statutory maximum" is not the maximum sentence a judge may impose after finding additional facts, but the maximum he may impose without any additional findings. When a judge inflicts punishment that the jury's verdict alone does not allow, the jury has not found all the facts "which the law makes essential to the punishment," Bishop, supra, §87, at 55, and the judge exceeds his proper authority.

[38] The judge in this case could not have imposed the exceptional 90-month sentence solely on the basis of the facts admitted in the guilty plea. Those facts alone were insufficient because, as the Washington Supreme Court has explained, "[a] reason offered to justify an exceptional sentence can be considered only if it takes into account factors other than those which are used in computing the standard range sentence for the offense," Gore, 143 Wash. 2d, at 315-316, 21 P. 3d, at 277, which in this case included the elements of second-degree kidnaping and the use of a firearm, see §§9.94A.320, 9.94A.310(3)(b).*fn7 Had the judge imposed the 90-month sentence solely on the basis of the plea, he would have been reversed. See §9.94A.210(4). The "maximum sentence" is no more 10 years here than it was 20 years in Apprendi (because that is what the judge could have imposed upon finding a hate crime) or death in Ring (because that is what the judge could have imposed upon finding an aggravator).

[39] The State defends the sentence by drawing an analogy to those we upheld in McMillan v. Pennsylvania, 477 U. S. 79 (1986), and Williams v. New York, 337 U. S. 241 (1949). Neither case is on point. McMillan involved a sentencing scheme that imposed a statutory minimum if a judge found a particular fact. 477 U. S., at 81. We specifically noted that the statute "does not authorize a sentence in excess of that otherwise allowed for [the underlying] offense." Id., at 82; cf. Harris, supra, at 567. Williams involved an indeterminate-sentencing regime that allowed a judge (but did not compel him) to rely on facts outside the trial record in determining whether to sentence a defendant to death. 337 U. S., at 242-243, and n. 2. The judge could have "sentenced [the defendant] to death giving no reason at all." Id., at 252. Thus, neither case involved a sentence greater than what state law authorized on the basis of the verdict alone.

[40] Finally, the State tries to distinguish Apprendi and Ring by pointing out that the enumerated grounds for departure in its regime are illustrative rather than exhaustive. This distinction is immaterial. Whether the judge's authority to impose an enhanced sentence depends on finding a specified fact (as in Apprendi), one of several specified facts (as in Ring), or any aggravating fact (as here), it remains the case that the jury's verdict alone does not authorize the sentence. The judge acquires that authority only upon finding some additional fact.*fn8

[41] Because the State's sentencing procedure did not comply with the Sixth Amendment, petitioner's sentence is invalid.*fn9

[42] III.

[43] Our commitment to Apprendi in this context reflects not just respect for longstanding precedent, but the need to give intelligible content to the right of jury trial. That right is no mere procedural formality, but a fundamental reservation of power in our constitutional structure. Just as suffrage ensures the people's ultimate control in the legislative and executive branches, jury trial is meant to ensure their control in the judiciary. See Letter XV by the Federal Farmer (Jan. 18, 1788), reprinted in 2 The Complete Anti-Federalist 315, 320 (H. Storing ed. 1981) (describing the jury as "secur[ing] to the people at large, their just and rightful controul in the judicial department"); John Adams, Diary Entry (Feb. 12, 1771), reprinted in 2 Works of John Adams 252, 253 (C. Adams ed. 1850) ("[T]he common people, should have as complete a control ... in every judgment of a court of judicature" as in the legislature); Letter from Thomas Jefferson to the Abbé Arnoux (July 19, 1789), reprinted in 15 Papers of Thomas Jefferson 282, 283 (J. Boyd ed. 1958) ("Were I called upon to decide whether the people had best be omitted in the Legislative or Judiciary department, I would say it is better to leave them out of the Legislative"); Jones v. United States, 526 U. S. 227, 244-248 (1999). Apprendi carries out this design by ensuring that the judge's authority to sentence derives wholly from the jury's verdict. Without that restriction, the jury would not exercise the control that the Framers intended.

[44] Those who would reject Apprendi are resigned to one of two alternatives. The first is that the jury need only find whatever facts the legislature chooses to label elements of the crime, and that those it labels sentencing factors -- no matter how much they may increase the punishment -- may be found by the judge. This would mean, for example, that a judge could sentence a man for committing murder even if the jury convicted him only of illegally possessing the firearm used to commit it -- or of making an illegal lane change while fleeing the death scene. Not even Apprendi's critics would advocate this absurd result. Cf. 530 U. S., at 552-553 (O'Connor, J., dissenting). The jury could not function as circuitbreaker in the State's machinery of justice if it were relegated to making a determination that the defendant at some point did something wrong, a mere preliminary to a judicial inquisition into the facts of the crime the State actually seeks to punish.*fn10

[45] The second alternative is that legislatures may establish legally essential sentencing factors within limits -- limits crossed when, perhaps, the sentencing factor is a "tail which wags the dog of the substantive offense." McMillan, 477 U. S., at 88. What this means in operation is that the law must not go too far -- it must not exceed the judicial estimation of the proper role of the judge.

[46] The subjectivity of this standard is obvious. Petitioner argued below that second-degree kidnaping with deliberate cruelty was essentially the same as first-degree kidnaping, the very charge he had avoided by pleading to a lesser offense. The court conceded this might be so but held it irrelevant. See 111 Wash. App., at 869, 47 P. 3d, at 158.*fn11 Petitioner's 90-month sentence exceeded the 53-month standard maximum by almost 70%; the Washington Supreme Court in other cases has upheld exceptional sentences 15 times the standard maximum. See State v. Oxborrow, 106 Wash. 2d 525, 528, 533, 723 P. 2d 1123, 1125, 1128 (1986) (15-year exceptional sentence; 1-year standard maximum sentence); State v. Branch, 129 Wash. 2d 635, 650, 919 P. 2d 1228, 1235 (1996) (4-year exceptional sentence; 3-month standard maximum sentence). Did the court go too far in any of these cases? There is no answer that legal analysis can provide. With too far as the yardstick, it is always possible to disagree with such judgments and never to refute them.

[47] Whether the Sixth Amendment incorporates this manipulable standard rather than Apprendi's bright-line rule depends on the plausibility of the claim that the Framers would have left definition of the scope of jury power up to judges' intuitive sense of how far is too far. We think that claim not plausible at all, because the very reason the Framers put a jury-trial guarantee in the Constitution is that they were unwilling to trust government to mark out the role of the jury.

[48] IV.

[49] By reversing the judgment below, we are not, as the State would have it, "find[ing] determinate sentencing schemes unconstitutional." Brief for Respondent 34. This case is not about whether determinate sentencing is constitutional, only about how it can be implemented in a way that respects the Sixth Amendment. Several policies prompted Washington's adoption of determinate sentencing, including proportionality to the gravity of the offense and parity among defendants. See Wash. Rev. Code Ann. §9.94A.010 (2000). Nothing we have said impugns those salutary objectives.

[50] Justice O'Connor argues that, because determinate sentencing schemes involving judicial factfinding entail less judicial discretion than indeterminate schemes, the constitutionality of the latter implies the constitutionality of the former. Post, at 1-10. This argument is flawed on a number of levels. First, the Sixth Amendment by its terms is not a limitation on judicial power, but a reservation of jury power. It limits judicial power only to the extent that the claimed judicial power infringes on the province of the jury. Indeterminate sentencing does not do so. It increases judicial discretion, to be sure, but not at the expense of the jury's traditional function of finding the facts essential to lawful imposition of the penalty. Of course indeterminate schemes involve judicial factfinding, in that a judge (like a parole board) may implicitly rule on those facts he deems important to the exercise of his sentencing discretion. But the facts do not pertain to whether the defendant has a legal right to a lesser sentence -- and that makes all the difference insofar as judicial impingement upon the traditional role of the jury is concerned. In a system that says the judge may punish burglary with 10 to 40 years, every burglar knows he is risking 40 years in jail. In a system that punishes burglary with a 10-year sentence, with another 30 added for use of a gun, the burglar who enters a home unarmed is entitled to no more than a 10-year sentence -- and by reason of the Sixth Amendment the facts bearing upon that entitlement must be found by a jury.

[51] But even assuming that restraint of judicial power unrelated to the jury's role is a Sixth Amendment objective, it is far from clear that Apprendi disserves that goal. Determinate judicial-factfinding schemes entail less judicial power than indeterminate schemes, but more judicial power than determinate jury-factfinding schemes. Whether Apprendi increases judicial power overall depends on what States with determinate judicial-factfinding schemes would do, given the choice between the two alternatives. Justice O'Connor simply assumes that the net effect will favor judges, but she has no empirical basis for that prediction. Indeed, what evidence we have points exactly the other way: When the Kansas Supreme Court found Apprendi infirmities in that State's determinate-sentencing regime in State v. Gould, 271 Kan. 394, 404-414, 23 P. 3d 801, 809-814 (2001), the legislature responded not by reestablishing indeterminate sentencing but by applying Apprendi's requirements to its current regime. See Act of May 29, 2002, ch. 170, 2002 Kan. Sess. Laws pp. 1018-1023 (codified at Kan. Stat. Ann. §21-4718 (2003 Cum. Supp.)); Brief for Kansas Appellate Defender Office as Amicus Curiae 3-7. The result was less, not more, judicial power.

[52] Justice Breyer argues that Apprendi works to the detriment of criminal defendants who plead guilty by depriving them of the opportunity to argue sentencing factors to a judge. Post, at 4-5. But nothing prevents a defendant from waiving his Apprendi rights. When a defendant pleads guilty, the State is free to seek judicial sentence enhancements so long as the defendant either stipulates to the relevant facts or consents to judicial factfinding. See Apprendi, 530 U. S., at 488; Duncan v. Louisiana, 391 U. S. 145, 158 (1968). If appropriate waivers are procured, States may continue to offer judicial factfinding as a matter of course to all defendants who plead guilty. Even a defendant who stands trial may consent to judicial factfinding as to sentence enhancements, which may well be in his interest if relevant evidence would prejudice him at trial. We do not understand how Apprendi can possibly work to the detriment of those who are free, if they think its costs outweigh its benefits, to render it inapplicable.*fn12

[53] Nor do we see any merit to Justice Breyer's contention that Apprendi is unfair to criminal defendants because, if States respond by enacting "17-element robbery crime[s]," prosecutors will have more elements with which to bargain. Post, at 4-5, 9 (citing Bibas, Judicial Fact-Finding and Sentence Enhancements in a World of Guilty Pleas, 110 Yale L. J. 1097 (2001)). Bargaining already exists with regard to sentencing factors because defendants can either stipulate or contest the facts that make them applicable. If there is any difference between bargaining over sentencing factors and bargaining over elements, the latter probably favors the defendant. Every new element that a prosecutor can threaten to charge is also an element that a defendant can threaten to contest at trial and make the prosecutor prove beyond a reasonable doubt. Moreover, given the sprawling scope of most criminal codes, and the power to affect sentences by making (even non-binding) sentencing recommendations, there is already no shortage of in terrorem tools at prosecutors' disposal. See King & Klein, Apprendi and Plea Bargaining, 54 Stan. L. Rev. 295, 296 (2001) ("Every prosecutorial bargaining chip mentioned by Professor Bibas existed pre-Apprendi exactly as it does post-Apprendi").

[54] Any evaluation of Apprendi's "fairness" to criminal defendants must compare it with the regime it replaced, in which a defendant, with no warning in either his indictment or plea, would routinely see his maximum potential sentence balloon from as little as five years to as much as life imprisonment, see 21 U. S. C. §§841(b)(1)(A), (D),*fn13 based not on facts proved to his peers beyond a reasonable doubt, but on facts extracted after trial from a report compiled by a probation officer who the judge thinks more likely got it right than got it wrong. We can conceive of no measure of fairness that would find more fault in the utterly speculative bargaining effects Justice Breyer identifies than in the regime he champions. Suffice it to say that, if such a measure exists, it is not the one the Framers left us with.

[55] The implausibility of Justice Breyer's contention that Apprendi is unfair to criminal defendants is exposed by the lineup of amici in this case. It is hard to believe that the National Association of Criminal Defense Lawyers was somehow duped into arguing for the wrong side. Justice Breyer's only authority asking that defendants be protected from Apprendi is an article written not by a criminal defense lawyer but by a law professor and former prosecutor. See post, at 4-5 (citing Bibas, supra); Association of American Law Schools Directory of Law Teachers 2003-2004, p. 319.

[56] Justice Breyer also claims that Apprendi will attenuate the connection between "real criminal conduct and real punishment" by encouraging plea bargaining and by restricting alternatives to adversarial factfinding. Post, at 7-8, 11-12. The short answer to the former point (even assuming the questionable premise that Apprendi does encourage plea bargaining, but see supra, at 14, and n. 12) is that the Sixth Amendment was not written for the benefit of those who choose to forgo its protection. It guarantees the right to jury trial. It does not guarantee that a particular number of jury trials will actually take place. That more defendants elect to waive that right (because, for example, government at the moment is not particularly oppressive) does not prove that a constitutional provision guaranteeing availability of that option is disserved.

[57] Justice Breyer's more general argument -- that Apprendi undermines alternatives to adversarial factfinding -- is not so much a criticism of Apprendi as an assault on jury trial generally. His esteem for "non-adversarial" truth-seeking processes, post, at 12, supports just as well an argument against either. Our Constitution and the common-law traditions it entrenches, however, do not admit the contention that facts are better discovered by judicial inquisition than by adversarial testing before a jury. See 3 Blackstone, Commentaries, at 373-374, 379-381. Justice Breyer may be convinced of the equity of the regime he favors, but his views are not the ones we are bound to uphold.

[58] Ultimately, our decision cannot turn on whether or to what degree trial by jury impairs the efficiency or fairness of criminal justice. One can certainly argue that both these values would be better served by leaving justice entirely in the hands of professionals; many nations of the world, particularly those following civil-law traditions, take just that course. There is not one shred of doubt, however, about the Framers' paradigm for criminal justice: not the civil-law ideal of administrative perfection, but the common-law ideal of limited state power accomplished by strict division of authority between judge and jury. As Apprendi held, every defendant has the right to insist that the prosecutor prove to a jury all facts legally essential to the punishment. Under the dissenters' alternative, he has no such right. That should be the end of the matter.

[59] Petitioner was sentenced to prison for more than three years beyond what the law allowed for the crime to which he confessed, on the basis of a disputed finding that he had acted with "deliberate cruelty." The Framers would not have thought it too much to demand that, before depriving a man of three more years of his liberty, the State should suffer the modest inconvenience of submitting its accusation to "the unanimous suffrage of twelve of his equals and neighbours," 4 Blackstone, Commentaries, at 343, rather than a lone employee of the State.

[60] The judgment of the Washington Court of Appeals is reversed, and the case is remanded for further proceedings not inconsistent with this opinion.

[61] It is so ordered.

[62] O'Connor, J., dissenting

[63] Justice O'Connor, with whom Justice Breyer joins, and with whom The Chief Justice and Justice Kennedy join as to all but Part IV-B, dissenting.

[64] The legacy of today's opinion, whether intended or not, will be the consolidation of sentencing power in the State and Federal Judiciaries. The Court says to Congress and state legislatures: If you want to constrain the sentencing discretion of judges and bring some uniformity to sentencing, it will cost you -- dearly. Congress and States, faced with the burdens imposed by the extension of Apprendi to the present context, will either trim or eliminate altogether their sentencing guidelines schemes and, with them, 20 years of sentencing reform. It is thus of little moment that the majority does not expressly declare guidelines schemes unconstitutional, ante, at 12; for, as residents of "Apprendi-land" are fond of saying, "the relevant inquiry is one not of form, but of effect." Apprendi v. New Jersey, 530 U. S. 466, 494 (2000); Ring v. Arizona, 536 U. S. 584, 613 (2002) (Scalia, J., concurring). The "effect" of today's decision will be greater judicial discretion and less uniformity in sentencing. Because I find it implausible that the Framers would have considered such a result to be required by the Due Process Clause or the Sixth Amendment, and because the practical consequences of today's decision may be disastrous, I respectfully dissent.

[65] I.

[66] One need look no further than the history leading up to and following the enactment of Washington's guidelines scheme to appreciate the damage that today's decision will cause. Prior to 1981, Washington, like most other States and the Federal Government, employed an indeterminate sentencing scheme. Washington's criminal code separated all felonies into three broad categories: "class A," carrying a sentence of 20 years to life; "class B," carrying a sentence of 0 to 10 years; and "class C," carrying a sentence of 0 to 5 years. Wash. Rev. Code Ann. §9A.20.020 (2000); see also Sentencing Reform Act of 1981, 1981 Wash. Laws, ch. 137, p. 534. Sentencing judges, in conjunction with parole boards, had virtually unfettered discretion to sentence defendants to prison terms falling anywhere within the statutory range, including probation -- i.e., no jail sentence at all. Wash. Rev. Code Ann. §§9.95.010-.011; Boerner & Lieb, Sentencing Reform in the Other Washington, 28 Crime and Justice 71, 73 (M. Tonry ed. 2001) (hereinafter Boerner & Lieb) ("Judges were authorized to choose between prison and probation with few exceptions, subject only to review for abuse of discretion"). See also D. Boerner, Sentencing in Washington §2.4, pp. 2-27 to 2-28 (1985).

[67] This system of unguided discretion inevitably resulted in severe disparities in sentences received and served by defendants committing the same offense and having similar criminal histories. Boerner & Lieb 126-127; cf. S. Rep. No. 98-225, p. 38 (1983) (Senate Report on precursor to federal Sentencing Reform Act of 1984) ("[E]very day Federal judges mete out an unjustifiably wide range of sentences to offenders with similar histories, convicted of similar crimes, committed under similar circumstances. ... These disparities, whether they occur at the time of the initial sentencing or at the parole stage, can be traced directly to the unfettered discretion the law confers on those judges and parole authorities responsible for imposing and implementing the sentence"). Indeed, rather than reflect legally relevant criteria, these disparities too often were correlated with constitutionally suspect variables such as race. Boerner & Lieb 126-128. See also Breyer, The Federal Sentencing Guidelines and Key Compromises Upon Which They Rest, 17 Hofstra L. Rev. 1, 5 (1988) (elimination of racial disparity one reason behind Congress' creation of the Federal Sentencing Commission).

[68] To counteract these trends, the state legislature passed the Sentencing Reform Act of 1981. The Act had the laudable purposes of "mak[ing] the criminal justice system accountable to the public," and "[e]nsur[ing] that the punishment for a criminal offense is proportionate to the seriousness of the offense ... [and] commensurate with the punishment imposed on others committing similar offenses." Wash. Rev. Code Ann. §9.94A.010 (2000). The Act neither increased any of the statutory sentencing ranges for the three types of felonies (though it did eliminate the statutory mandatory minimum for class A felonies), nor reclassified any substantive offenses. 1981 Wash. Laws ch. 137, p. 534. It merely placed meaningful constraints on discretion to sentence offenders within the statutory ranges, and eliminated parole. There is thus no evidence that the legislature was attempting to manipulate the statutory elements of criminal offenses or to circumvent the procedural protections of the Bill of Rights. Rather, lawmakers were trying to bring some much-needed uniformity, transparency, and accountability to an otherwise " `labyrinthine' sentencing and corrections system that `lack[ed] any principle except unguided discretion.' " Boerner & Lieb 73 (quoting F. Zimring, Making the Punishment Fit the Crime: A Consumers' Guide to Sentencing Reform, Occasional Paper No. 12, p. 6 (1977)).

[69] II.

[70] Far from disregarding principles of due process and the jury trial right, as the majority today suggests, Washington's reform has served them. Before passage of the Act, a defendant charged with second degree kidnaping, like petitioner, had no idea whether he would receive a 10-year sentence or probation. The ultimate sentencing determination could turn as much on the idiosyncracies of a particular judge as on the specifics of the defendant's crime or background. A defendant did not know what facts, if any, about his offense or his history would be considered relevant by the sentencing judge or by the parole board. After passage of the Act, a defendant charged with second degree kidnaping knows what his presumptive sentence will be; he has a good idea of the types of factors that a sentencing judge can and will consider when deciding whether to sentence him outside that range; he is guaranteed meaningful appellate review to protect against an arbitrary sentence. Boerner & Lieb 93 ("By consulting one sheet, practitioners could identify the applicable scoring rules for criminal history, the sentencing range, and the available sentencing options for each case"). Criminal defendants still face the same statutory maximum sentences, but they now at least know, much more than before, the real consequences of their actions.

[71] Washington's move to a system of guided discretion has served equal protection principles as well. Over the past 20 years, there has been a substantial reduction in racial disparity in sentencing across the State. Id., at 126 (Racial disparities that do exist "are accounted for by differences in legally relevant variables -- the offense of conviction and prior criminal record"); id., at 127 ("[J]udicial authority to impose exceptional sentences under the court's departure authority shows little evidence of disparity correlated with race"). The reduction is directly traceable to the constraining effects of the guidelines -- namely, its "presumptive range[s]" and limits on the imposition of "exceptional sentences" outside of those ranges. Id., at 128. For instance, sentencing judges still retain unreviewable discretion in first-time offender cases and in certain sex offender cases to impose alternative sentences that are far more lenient than those contemplated by the guidelines. To the extent that unjustifiable racial disparities have persisted in Washington, it has been in the imposition of such alternative sentences: "The lesson is powerful: racial disparity is correlated with unstructured and unreviewed discretion." Ibid.; see also Washington State Minority and Justice Commission, R. Crutchfield, J. Weis, R. Engen, & R. Gainey, Racial/Ethnic Disparities and Exceptional Sentences in Washington State, Final Report 51-53 (1993) ("[E]xceptional sentences are not a major source of racial disparities in sentencing").

[72] The majority does not, because it cannot, disagree that determinate sentencing schemes, like Washington's, serve important constitutional values. Ante, at 12. Thus, the majority says: "[t]his case is not about whether determinate sentencing is constitutional, only about how it can be implemented in a way that respects the Sixth Amendment." Ibid. But extension of Apprendi to the present context will impose significant costs on a legislature's determination that a particular fact, not historically an element, warrants a higher sentence. While not a constitutional prohibition on guidelines schemes, the majority's decision today exacts a substantial constitutional tax.

[73] The costs are substantial and real. Under the majority's approach, any fact that increases the upper bound on a judge's sentencing discretion is an element of the offense. Thus, facts that historically have been taken into account by sentencing judges to assess a sentence within a broad range --such as drug quantity, role in the offense, risk of bodily harm -- all must now be charged in an indictment and submitted to a jury, In re Winship, 397 U. S. 358 (1970), simply because it is the legislature, rather than the judge, that constrains the extent to which such facts may be used to impose a sentence within a pre-existing statutory range.

[74] While that alone is enough to threaten the continued use of sentencing guidelines schemes, there are additional costs. For example, a legislature might rightly think that some factors bearing on sentencing, such as prior bad acts or criminal history, should not be considered in a jury's determination of a defendant's guilt -- such "character evidence" has traditionally been off limits during the guilt phase of criminal proceedings because of its tendency to inflame the passions of the jury. See, e.g., Fed. Rule Evid. 404; 1 E. Imwinkelried, P. Giannelli, F. Gilligan, & F. Leaderer, Courtroom Criminal Evidence 285 (3d ed. 1998). If a legislature desires uniform consideration of such factors at sentencing, but does not want them to impact a jury's initial determination of guilt, the State may have to bear the additional expense of a separate, full-blown jury trial during the penalty phase proceeding.

[75] Some facts that bear on sentencing either will not be discovered, or are not discoverable, prior to trial. For instance, a legislature might desire that defendants who act in an obstructive manner during trial or post-trial proceedings receive a greater sentence than defendants who do not. See, e.g., United States Sentencing Commission, Guidelines Manual, §3C1.1 (Nov. 2003) (hereinafter USSG) (2-point increase in offense level for obstruction of justice). In such cases, the violation arises too late for the State to provide notice to the defendant or to argue the facts to the jury. A State wanting to make such facts relevant at sentencing must now either vest sufficient discretion in the judge to account for them or bring a separate criminal prosecution for obstruction of justice or perjury. And, the latter option is available only to the extent that a defendant's obstructive behavior is so severe as to constitute an already-existing separate offense, unless the legislature is willing to undertake the unlikely expense of criminalizing relatively minor obstructive behavior.

[76] Likewise, not all facts that historically have been relevant to sentencing always will be known prior to trial. For instance, trial or sentencing proceedings of a drug distribution defendant might reveal that he sold primarily to children. Under the majority's approach, a State wishing such a revelation to result in a higher sentence within a pre-existing statutory range either must vest judges with sufficient discretion to account for it (and trust that they exercise that discretion) or bring a separate criminal prosecution. Indeed, the latter choice might not be available -- a separate prosecution, if it is for an aggravated offense, likely would be barred altogether by the Double Jeopardy Clause. Blockburger v. United States, 284 U. S. 299 (1932) (cannot prosecute for separate offense unless the two offenses both have at least one element that the other does not).

[77] The majority may be correct that States and the Federal Government will be willing to bear some of these costs. Ante, at 13-14. But simple economics dictate that they will not, and cannot, bear them all. To the extent that they do not, there will be an inevitable increase in judicial discretion with all of its attendant failings.*fn14

[78] III.

[79] Washington's Sentencing Reform Act did not alter the statutory maximum sentence to which petitioner was exposed. See Wash. Rev. Code Ann. §9A.40.030 (2003) (second degree kidnaping class B felony since 1975); see also State v. Pawling, 23 Wash. App. 226, 228-229, 597 P. 2d 1367, 1369 (1979) (citing second degree kidnapping provision as existed in 1977). Petitioner was informed in the charging document, his plea agreement, and during his plea hearing that he faced a potential statutory maximum of 10 years in prison. App. 63, 66, 76. As discussed above, the guidelines served due process by providing notice to petitioner of the consequences of his acts; they vindicated his jury trial right by informing him of the stakes of risking trial; they served equal protection by ensuring petitioner that invidious characteristics such as race would not impact his sentence.

[80] Given these observations, it is difficult for me to discern what principle besides doctrinaire formalism actually motivates today's decision. The majority chides the Apprendi dissenters for preferring a nuanced interpretation of the Due Process Clause and Sixth Amendment jury trial guarantee that would generally defer to legislative labels while acknowledging the existence of constitutional constraints -- what the majority calls the "the law must not go too far" approach. Ante, at 11 (emphasis deleted). If indeed the choice is between adopting a balanced case-by-case approach that takes into consideration the values underlying the Bill of Rights, as well as the history of a particular sentencing reform law, and adopting a rigid rule that destroys everything in its path, I will choose the former. See Apprendi, 530 U. S., at 552-554 (O'Connor, J., dissenting) ("Because I do not believe that the Court's `increase in the maximum penalty' rule is required by the Constitution, I would evaluate New Jersey's sentence-enhancement statute by analyzing the factors we have examined in past cases" (citation omitted)).

[81] But even were one to accept formalism as a principle worth vindicating for its own sake, it would not explain Apprendi's, or today's, result. A rule of deferring to legislative labels has no less formal pedigree. It would be more consistent with our decisions leading up to Apprendi, see Almendarez-Torres v. United States, 523 U. S. 224 (1998) (fact of prior conviction not an element of aggravated recidivist offense); United States v. Watts, 519 U. S. 148 (1997) (per curiam) (acquittal of offense no bar to consideration of underlying conduct for purposes of guidelines enhancement); Witte v. United States, 515 U. S. 389 (1995) (no double jeopardy bar against consideration of uncharged conduct in imposition of guidelines enhancement); Walton v. Arizona, 497 U. S. 639 (1990) (aggravating factors need not be found by a jury in capital case); Mistretta v. United States, 488 U. S. 361 (1989) (Federal Sentencing Guidelines do not violate separation of powers); McMillan v. Pennsylvania, 477 U. S. 79 (1986) (facts increasing mandatory minimum sentence are not necessarily elements); and it would vest primary authority for defining crimes in the political branches, where it belongs. Apprendi, supra, at 523-554 (O'Connor, J., dissenting). It also would be easier to administer than the majority's rule, inasmuch as courts would not be forced to look behind statutes and regulations to determine whether a particular fact does or does not increase the penalty to which a defendant was exposed.

[82] The majority is correct that rigid adherence to such an approach could conceivably produce absurd results, ante, at 10; but, as today's decision demonstrates, rigid adherence to the majority's approach does and will continue to produce results that disserve the very principles the majority purports to vindicate. The pre-Apprendi rule of deference to the legislature retains a built-in political check to prevent lawmakers from shifting the prosecution for crimes to the penalty phase proceedings of lesser included and easier-to-prove offenses -- e.g., the majority's hypothesized prosecution of murder in the guise of a traffic offense sentencing proceeding. Ante, at 10. There is no similar check, however, on application of the majority's " `any fact that increases the upper bound of judicial discretion' " by courts.

[83] The majority claims the mantle of history and original intent. But as I have explained elsewhere, a handful of state decisions in the mid-19th century and a criminal procedure treatise have little if any persuasive value as evidence of what the Framers of the Federal Constitution intended in the late 18th century. See Apprendi, 530 U. S., at 525-528 (O'Connor, J., dissenting). Because broad judicial sentencing discretion was foreign to the Framers, id., at 478-479 (citing J. Archbold, Pleading and Evidence in Criminal Cases 44 (15th ed. 1862)), they were never faced with the constitutional choice between submitting every fact that increases a sentence to the jury or vesting the sentencing judge with broad discretionary authority to account for differences in offenses and offenders.

[84] IV.

[85] A.

[86] The consequences of today's decision will be as far reaching as they are disturbing. Washington's sentencing system is by no means unique. Numerous other States have enacted guidelines systems, as has the Federal Government. See, e.g., Alaska Stat. §12.55.155 (2003); Ark. Code Ann. §16-90-804 (Supp. 2003); Fla. Stat. §921.0016 (2003); Kan. Stat. Ann. §21-4701 et seq. (2003); Mich. Comp. Laws Ann. §769.34 (West Supp. 2004); Minn. Stat. §244.10 (2002); N. C. Gen. Stat. §15A-1340.16 (Lexis 2003); Ore. Admin. Rule §213-008-0001 (2003); 204 Pa. Code §303 et seq. (2004), reproduced following 42 Pa. Cons. Stat. Ann. §9721 (Purden Supp. 2004); 18 U. S. C. §3553; 28 U. S. C. §991 et seq. Today's decision casts constitutional doubt over them all and, in so doing, threatens an untold number of criminal judgments. Every sentence imposed under such guidelines in cases currently pending on direct appeal is in jeopardy. And, despite the fact that we hold in Schriro v. Summerlin, post, p. ___, that Ring (and a fortiori Apprendi) does not apply retroactively on habeas review, all criminal sentences imposed under the federal and state guidelines since Apprendi was decided in 2000 arguably remain open to collateral attack. See Teague v. Lane, 489 U. S. 288, 301 (1989) (plurality opinion) ("[A] case announces a new rule if the result was not dictated by precedent existing at the time the defendant's conviction became final").*fn15

[87] The practical consequences for trial courts, starting today, will be equally unsettling: How are courts to mete out guidelines sentences? Do courts apply the guidelines as to mitigating factors, but not as to aggravating factors? Do they jettison the guidelines altogether? The Court ignores the havoc it is about to wreak on trial courts across the country.

[88] B.

[89] It is no answer to say that today's opinion impacts only Washington's scheme and not others, such as, for example, the Federal Sentencing Guidelines. See ante, at 9, n. 9 ("The Federal Guidelines are not before us, and we express no opinion on them"); cf. Apprendi, supra, at 496-497 (claiming not to overrule Walton, supra, soon thereafter overruled in Ring); Apprendi, supra, at 497, n. 21 (reserving question of Federal Sentencing Guidelines). The fact that the Federal Sentencing Guidelines are promulgated by an administrative agency nominally located in the Judicial Branch is irrelevant to the majority's reasoning. The Guidelines have the force of law, see Stinson v. United States, 508 U. S. 36 (1993); and Congress has unfettered control to reject or accept any particular guideline, Mistretta, 488 U. S., at 393-394.

[90] The structure of the Federal Guidelines likewise does not, as the Government half-heartedly suggests, provide any grounds for distinction. Brief for United States as Amicus Curiae 27-29. Washington's scheme is almost identical to the upward departure regime established by 18 U. S. C. §3553(b) and implemented in USSG §5K2.0. If anything, the structural differences that do exist make the Federal Guidelines more vulnerable to attack. The provision struck down here provides for an increase in the upper bound of the presumptive sentencing range if the sentencing court finds, "considering the purpose of [the Act], that there are substantial and compelling reasons justifying an exceptional sentence." Wash. Rev. Code Ann. §9.94A.120 (2000). The Act elsewhere provides a nonexhaustive list of aggravating factors that satisfy the definition. §9.94A.390. The Court flatly rejects respondent's argument that such soft constraints, which still allow Washington judges to exercise a substantial amount of discretion, survive Apprendi. Ante, at 8-9. This suggests that the hard constraints found throughout chapters 2 and 3 of the Federal Sentencing Guidelines, which require an increase in the sentencing range upon specified factual findings, will meet the same fate. See, e.g., USSG §2K2.1 (increases in offense level for firearms offenses based on number of firearms involved, whether possession was in connection with another offense, whether the firearm was stolen); §2B1.1 (increase in offense level for financial crimes based on amount of money involved, number of victims, possession of weapon); §3C1.1 (general increase in offense level for obstruction of justice).

[91] Indeed, the "extraordinary sentence" provision struck down today is as inoffensive to the holding of Apprendi as a regime of guided discretion could possibly be. The list of facts that justify an increase in the range is nonexhaustive. The State's "real facts" doctrine precludes reliance by sentencing courts upon facts that would constitute the elements of a different or aggravated offense. See Wash. Rev. Code Ann. §9.94A.370(2) (2000) (codifying "real facts" doctrine). If the Washington scheme does not comport with the Constitution, it is hard to imagine a guidelines scheme that would.

[92] What I have feared most has now come to pass: Over 20 years of sentencing reform are all but lost, and tens of thousands of criminal judgments are in jeopardy. Apprendi, 530 U. S., at 549-559 (O'Connor, J., dissenting); Ring, 536 U. S., at 619-621 (O'Connor, J., dissenting). I respectfully dissent.

[93] Kennedy, J., dissenting

[94] Justice Kennedy, with whom Justice Breyer joins, dissenting.

[95] The majority opinion does considerable damage to our laws and to the administration of the criminal justice system for all the reasons well stated in Justice O'Connor's dissent, plus one more: The Court, in my respectful submission, disregards the fundamental principle under our constitutional system that different branches of government "converse with each other on matters of vital common interest." Mistretta v. United States, 488 U. S. 361, 408 (1989). As the Court in Mistretta explained, the Constitution establishes a system of government that presupposes, not just " `autonomy' " and " `separateness,' " but also " `interdependence' " and " `reciprocity.' " Id., at 381 (quoting Youngstown Sheet & Tube Co. v. Sawyer, 343 U. S. 579, 635 (1952) (Jackson, J., concurring)). Constant, constructive discourse between our courts and our legislatures is an integral and admirable part of the constitutional design. Case-by-case judicial determinations often yield intelligible patterns that can be refined by legislatures and codified into statutes or rules as general standards. As these legislative enactments are followed by incremental judicial interpretation, the legislatures may respond again, and the cycle repeats. This recurring dialogue, an essential source for the elaboration and the evolution of the law, is basic constitutional theory in action.

[96] Sentencing guidelines are a prime example of this collaborative process. Dissatisfied with the wide disparity in sentencing, participants in the criminal justice system, including judges, pressed for legislative reforms. In response, legislators drew from these participants' shared experiences and enacted measures to correct the problems, which, as Justice O'Connor explains, could sometimes rise to the level of a constitutional injury. As Mistretta recognized, this interchange among different actors in the constitutional scheme is consistent with the Constitution's structural protections.

[97] To be sure, this case concerns the work of a state legislature, and not of Congress. If anything, however, this distinction counsels even greater judicial caution. Unlike Mistretta, the case here implicates not just the collective wisdom of legislators on the other side of the continuing dialogue over fair sentencing, but also the interest of the States to serve as laboratories for innovation and experiment. See New State Ice Co. v. Liebmann, 285 U. S. 262, 311 (1932) (Brandeis, J., dissenting). With no apparent sense of irony that the effect of today's decision is the destruction of a sentencing scheme devised by democratically elected legislators, the majority shuts down alternative, non-judicial, sources of ideas and experience. It does so under a faintly disguised distrust of judges and their purported usurpation of the jury's function in criminal trials. It tells not only trial judges who have spent years studying the problem but also legislators who have devoted valuable time and resources "calling upon the accumulated wisdom and experience of the Judicial Branch ... on a matter uniquely within the ken of judges," Mistretta, supra, at 412, that their efforts and judgments were all for naught. Numerous States that have enacted sentencing guidelines similar to the one in Washington State are now commanded to scrap everything and start over.

[98] If the Constitution required this result, the majority's decision, while unfortunate, would at least be understandable and defensible. As Justice O'Connor's dissent demonstrates, however, this is simply not the case. For that reason, and because the Constitution does not prohibit the dynamic and fruitful dialogue between the judicial and legislative branches of government that has marked sentencing reform on both the state and the federal levels for more than 20 years, I dissent.

[99] Breyer, J., dissenting

[100] Justice Breyer, with whom Justice O'Connor joins, dissenting.

[101] The Court makes clear that it means what it said in Apprendi v. New Jersey, 530 U. S. 466 (2000). In its view, the Sixth Amendment says that " 'any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury.' " Ante, at 5 (quoting Apprendi, supra, at 490). " `[P]rescribed statutory maximum' " means the penalty that the relevant statute authorizes "solely on the basis of the facts reflected in the jury verdict." Ante, at 7 (emphasis deleted). Thus, a jury must find, not only the facts that make up the crime of which the offender is charged, but also all (punishment-increasing) facts about the way in which the offender carried out that crime.

[102] It is not difficult to understand the impulse that produced this holding. Imagine a classic example -- a statute (or mandatory sentencing guideline) that provides a 10-year sentence for ordinary bank robbery, but a 15-year sentence for bank robbery committed with a gun. One might ask why it should matter for jury trial purposes whether the statute (or guideline) labels the gun's presence (a) a sentencing fact about the way in which the offender carried out the lesser crime of ordinary bank robbery, or (b) a factual element of the greater crime of bank robbery with a gun? If the Sixth Amendment requires a jury finding about the gun in the latter circumstance, why should it not also require a jury to find the same fact in the former circumstance? The two sets of circumstances are functionally identical. In both instances, identical punishment follows from identical factual findings (related to, e.g., a bank, a taking, a thing-of-value, force or threat of force, and a gun). The only difference between the two circumstances concerns a legislative (or Sentencing Commission) decision about which label ("sentencing fact" or "element of a greater crime") to affix to one of the facts, namely, the presence of the gun, that will lead to the greater sentence. Given the identity of circumstances apart from the label, the jury's traditional factfinding role, and the law's insistence upon treating like cases alike, why should the legislature's labeling choice make an important Sixth Amendment difference?

[103] The Court in Apprendi, and now here, concludes that it should not make a difference. The Sixth Amendment's jury trial guarantee applies similarly to both. I agree with the majority's analysis, but not with its conclusion. That is to say, I agree that, classically speaking, the difference between a traditional sentencing factor and an element of a greater offense often comes down to a legislative choice about which label to affix. But I cannot jump from there to the conclusion that the Sixth Amendment always requires identical treatment of the two scenarios. That jump is fraught with consequences that threaten the fairness of our traditional criminal justice system; it distorts historical sentencing or criminal trial practices; and it upsets settled law on which legislatures have relied in designing punishment systems.

[104] The Justices who have dissented from Apprendi have written about many of these matters in other opinions. See 530 U. S., at 523-554 (O'Connor, J., dissenting); id., at 555-566 (Breyer, J., dissenting); Harris v. United States, 536 U. S. 545, 549-550, 556-569 (2002) (Kennedy, J.); id., at 569-572 (Breyer, J., concurring in part and concurring in judgment); Jones v. United States, 526 U. S. 227, 254, 264-272 (1999) (Kennedy, J., dissenting); Monge v. California, 524 U. S. 721, 728-729 (1998) (O'Connor, J.); McMillan v. Pennsylvania, 477 U. S. 79, 86-91 (1986) (Rehnquist, C. J.). At the risk of some repetition, I shall set forth several of the most important considerations here. They lead me to conclude that I must again dissent.

[105] I.

[106] The majority ignores the adverse consequences inherent in its conclusion. As a result of the majority's rule, sentencing must now take one of three forms, each of which risks either impracticality, unfairness, or harm to the jury trial right the majority purports to strengthen. This circumstance shows that the majority's Sixth Amendment interpretation cannot be right.

[107] A.

[108] A first option for legislators is to create a simple, pure or nearly pure "charge offense" or "determinate" sentencing system. See Breyer, The Federal Sentencing Guidelines and the Key Compromises upon Which They Rest, 17 Hofstra L. Rev. 1, 8-9 (1988). In such a system, an indictment would charge a few facts which, taken together, constitute a crime, such as robbery. Robbery would carry a single sentence, say, five years' imprisonment. And every person convicted of robbery would receive that sentence -- just as, centuries ago, everyone convicted of almost any serious crime was sentenced to death. See, e.g., Lillquist, The Puzzling Return of Jury Sentencing: Misgivings About Apprendi, 82 N. C. L. Rev. 621, 630 (2004).

[109] Such a system assures uniformity, but at intolerable costs. First, simple determinate sentencing systems impose identical punishments on people who committed their crimes in very different ways. When dramatically different conduct ends up being punished the same way, an injustice has taken place. Simple determinate sentencing has the virtue of treating like cases alike, but it simultaneously fails to treat different cases differently. Some commentators have leveled this charge at sentencing guideline systems themselves. See, e.g., Schulhofer, Assessing the Federal Sentencing Process: The Problem Is Uniformity, Not Disparity, 29 Am. Crim. L. Rev. 833, 847 (1992) (arguing that the "most important problem under the [Federal] Guidelines system is not too much disparity, but rather excessive uniformity" and arguing for adjustments, including elimination of mandatory minimums, to make the Guidelines system more responsive to relevant differences). The charge is doubly applicable to simple "pure charge" systems that permit no departures from the prescribed sentences, even in extraordinary cases.

[110] Second, in a world of statutorily fixed mandatory sentences for many crimes, determinate sentencing gives tremendous power to prosecutors to manipulate sentences through their choice of charges. Prosecutors can simply charge, or threaten to charge, defendants with crimes bearing higher mandatory sentences. Defendants, knowing that they will not have a chance to argue for a lower sentence in front of a judge, may plead to charges that they might otherwise contest. Considering that most criminal cases do not go to trial and resolution by plea bargaining is the norm, the rule of Apprendi, to the extent it results in a return to determinate sentencing, threatens serious unfairness. See Bibas, Judicial Fact-Finding and Sentence Enhancements in a World of Guilty Pleas, 110 Yale L. J. 1097, 1100-1101 (2001) (explaining that the rule of Apprendi hurts defendants by depriving them of sentencing hearings, "the only hearings they were likely to have"; forcing defendants to surrender sentencing issues like drug quantity when they agree to the plea; and transferring power to prosecutors).

[111] B.

[112] A second option for legislators is to return to a system of indeterminate sentencing, such as California had before the recent sentencing reform movement. See Payne v. Tennessee, 501 U. S. 808, 820 (1991) ("With the increasing importance of probation, as opposed to imprisonment, as a part of the penological process, some States such as California developed the `indeterminate sentence,' where the time of incarceration was left almost entirely to the penological authorities rather than to the courts"); Thompson, Navigating the Hidden Obstacles to Ex-Offender Re-entry, 45 Boston College L. Rev. 255, 267 (2004) ("In the late 1970s, California switched from an indeterminate criminal sentencing scheme to determinate sentencing" (footnote omitted)). Under indeterminate systems, the length of the sentence is entirely or almost entirely within the discretion of the judge or of the parole board, which typically has broad power to decide when to release a prisoner.

[113] When such systems were in vogue, they were criticized, and rightly so, for producing unfair disparities, including race-based disparities, in the punishment of similarly situated defendants. See, e.g., ante, at 2-3 (O'Connor, J., dissenting) (citing sources). The length of time a person spent in prison appeared to depend on "what the judge ate for breakfast" on the day of sentencing, on which judge you got, or on other factors that should not have made a difference to the length of the sentence. See Breyer, supra, at 4-5 (citing congressional and expert studies indicating that, before the United States Sentencing Com-mission Guidelines were promulgated, punishments for identical crimes in the Second Circuit ranged from 3 to 20 years' imprisonment and that sentences varied depending upon region, gender of the defendant, and race of the defendant). And under such a system, the judge could vary the sentence greatly based upon his findings about how the defendant had committed the crime -- findings that might not have been made by a "preponderance of the evidence," much less "beyond a reasonable doubt." See McMillan, 477 U. S., at 91 ("Sentencing courts have traditionally heard evidence and found facts without any prescribed burden of proof at all" (citing Williams v. New York, 337 U. S. 241 (1949))).

[114] Returning to such a system would diminish the " `reason' " the majority claims it is trying to uphold. Ante, at 5 (quoting 1 J. Bishop, Criminal Procedure §87, p. 55 (2d ed. 1872)). It also would do little to "ensur[e] [the] control" of what the majority calls "the peopl[e,]" i.e., the jury, "in the judiciary," ante, at 9, since "the peopl[e]" would only decide the defendant's guilt, a finding with no effect on the duration of the sentence. While "the judge's authority to sentence" would formally derive from the jury's verdict, the jury would exercise little or no control over the sentence itself. Ante, at 10. It is difficult to see how such an outcome protects the structural safeguards the majority claims to be defending.

[115] C.

[116] A third option is that which the Court seems to believe legislators will in fact take. That is the option of retaining structured schemes that attempt to punish similar conduct similarly and different conduct differently, but modifying them to conform to Apprendi's dictates. Judges would be able to depart downward from presumptive sentences upon finding that mitigating factors were present, but would not be able to depart upward unless the prosecutor charged the aggravating fact to a jury and proved it beyond a reasonable doubt. The majority argues, based on the single example of Kansas, that most legislatures will enact amendments along these lines in the face of the oncoming Apprendi train. See ante, at 13-14 (citing State v. Gould, 271 Kan. 394, 404-414, 23 P. 3d 801, 809-814 (2001); Act of May 29, 2002, ch. 170, 2002 Kan. Sess. Laws pp. 1018-1023 (codified at Kan. Stat. Ann. §21-4718 (2003 Cum. Supp.)); Brief for Kansas Appellate Defender Office as Amicus Curiae 3-7). It is therefore worth exploring how this option could work in practice, as well as the assumptions on which it depends.

[117] 1.

[118] This option can be implemented in one of two ways. The first way would be for legislatures to subdivide each crime into a list of complex crimes, each of which would be defined to include commonly found sentencing factors such as drug quantity, type of victim, presence of violence, degree of injury, use of gun, and so on. A legislature, for example, might enact a robbery statute, modeled on robbery sentencing guidelines, that increases punishment depending upon (1) the nature of the institution robbed, (2) the (a) presence of, (b) brandishing of, (c) other use of, a firearm, (3) making of a death threat, (4) presence of (a) ordinary, (b) serious, (c) permanent or life threatening, bodily injury, (5) abduction, (6) physical restraint, (7) taking of a firearm, (8) taking of drugs, (9) value of property loss, etc. Cf. United States Sentencing Commission, Guidelines Manual §2B3.1 (Nov. 2003) (hereinafter USSG).

[119] This possibility is, of course, merely a highly calibrated form of the "pure charge" system discussed in Part I-A, supra. And it suffers from some of the same defects. The prosecutor, through control of the precise charge, controls the punishment, thereby marching the sentencing system directly away from, not toward, one important guideline goal: rough uniformity of punishment for those who engage in roughly the same real criminal conduct. The artificial (and consequently unfair) nature of the resulting sentence is aggravated by the fact that prosecutors must charge all relevant facts about the way the crime was committed before a presentence investigation examines the criminal conduct, perhaps before the trial itself, i.e., before many of the facts relevant to punishment are known.

[120] This "complex charge offense" system also prejudices defendants who seek trial, for it can put them in the untenable position of contesting material aggravating facts in the guilt phases of their trials. Consider a defendant who is charged, not with mere possession of cocaine, but with the specific offense of possession of more than 500 grams of cocaine. Or consider a defendant charged, not with murder, but with the new crime of murder using a machete. Or consider a defendant whom the prosecution wants to claim was a "supervisor," rather than an ordinary gang member. How can a Constitution that guarantees due process put these defendants, as a matter of course, in the position of arguing, "I did not sell drugs, and if I did, I did not sell more than 500 grams" or, "I did not kill him, and if I did, I did not use a machete," or "I did not engage in gang activity, and certainly not as a supervisor" to a single jury? See Apprendi, 530 U. S., at 557-558 (Breyer, J., dissenting); Monge, 524 U. S., at 729. The system can tolerate this kind of problem up to a point (consider the defendant who wants to argue innocence, and, in the alternative, second-degree, not first-degree, murder). But a rereading of the many distinctions made in a typical robbery guideline, see supra, at 7, suggests that an effort to incorporate any real set of guidelines in a complex statute would reach well beyond that point.

[121] The majority announces that there really is no problem here because "States may continue to offer judicial factfinding as a matter of course to all defendants who plead guilty" and defendants may "stipulat[e] to the relevant facts or consen[t] to judicial factfinding." Ante, at 14. The problem, of course, concerns defendants who do not want to plead guilty to those elements that, until recently, were commonly thought of as sentencing factors. As to those defendants, the fairness problem arises because States may very well decide that they will not permit defendants to carve subsets of facts out of the new, Apprendi-required 17-element robbery crime, seeking a judicial determination as to some of those facts and a jury determination as to others. Instead, States may simply require defendants to plead guilty to all 17 elements or proceed with a (likely prejudicial) trial on all 17 elements.

[122] The majority does not deny that States may make this choice; it simply fails to understand why any State would want to exercise it. Ante, at 14, n. 12. The answer is, as I shall explain in a moment, that the alternative may prove too expensive and unwieldy for States to provide. States that offer defendants the option of judicial factfinding as to some facts (i.e., sentencing facts), say, because of fairness concerns, will also have to offer the defendant a second sentencing jury -- just as Kansas has done. I therefore turn to that alternative.

[123] 2.

[124] The second way to make sentencing guidelines Apprendi-compliant would be to require at least two juries for each defendant whenever aggravating facts are present: one jury to determine guilt of the crime charged, and an additional jury to try the disputed facts that, if found, would aggravate the sentence. Our experience with bifurcated trials in the capital punishment context suggests that requiring them for run-of-the-mill sentences would be costly, both in money and in judicial time and resources. Cf. Kozinski & Gallagher, Death: The Ultimate Run-On Sentence, 46 Case W. Res. L. Rev. 1, 13-15, and n. 64 (1995) (estimating the costs of each capital case at around $1 million more than each non-capital case); Tabak, How Empirical Studies Can Affect Positively the Politics of the Death Penalty, 83 Cornell L. Rev. 1431, 1439-1440 (1998) (attributing the greater cost of death penalty cases in part to bifurcated proceedings). In the context of non-capital crimes, the potential need for a second indictment alleging aggravating facts, the likely need for formal evidentiary rules to prevent prejudice, and the increased difficulty of obtaining relevant sentencing information, all will mean greater complexity, added cost, and further delay. See Part V, infra. Indeed, cost and delay could lead legislatures to revert to the complex charge offense system described in Part I-C-1, supra.

[125] The majority refers to an amicus curiae brief filed by the Kansas Appellate Defender Office, which suggests that a two-jury system has proved workable in Kansas. Ante, at 13-14. And that may be so. But in all likelihood, any such workability reflects an uncomfortable fact, a fact at which the majority hints, ante, at 14, but whose constitutional implications it does not seem to grasp. The uncomfortable fact that could make the system seem workable -- even desirable in the minds of some, including defense attorneys -- is called "plea bargaining." See Bibas, 110 Yale L. J., at 1150, and n. 330 (reporting that in 1996, fewer than 4% of adjudicated state felony defendants have jury trials, 5% have bench trials, and 91% plead guilty). See also ante, at 14 (making clear that plea bargaining applies). The Court can announce that the Constitution requires at least two jury trials for each criminal defendant -- one for guilt, another for sentencing -- but only because it knows full well that more than 90% of defendants will not go to trial even once, much less insist on two or more trials.

[126] What will be the consequences of the Court's holding for the 90% of defendants who do not go to trial? The truthful answer is that we do not know. Some defendants may receive bargaining advantages if the increased cost of the "double jury trial" guarantee makes prosecutors more willing to cede certain sentencing issues to the defense. Other defendants may be hurt if a "single-jury-decides-all" approach makes them more reluctant to risk a trial -- perhaps because they want to argue that they did not know what was in the cocaine bag, that it was a small amount regardless, that they were unaware a confederate had a gun, etc. See Bibas, 110 Yale L. J., at 1100 ("Because for many defendants going to trial is not a desirable option, they are left without any real hearings at all"); id., at 1151 ("The trial right does little good when most defendants do not go to trial").

[127] At the least, the greater expense attached to trials and their greater complexity, taken together in the context of an overworked criminal justice system, will likely mean, other things being equal, fewer trials and a greater reliance upon plea bargaining -- a system in which punishment is set not by judges or juries but by advocates acting under bargaining constraints. At the same time, the greater power of the prosecutor to control the punishment through the charge would likely weaken the relation between real conduct and real punishment as well. See, e.g., Schulhofer, 29 Am. Crim. L. Rev., at 845 (estimating that evasion of the proper sentence under the Federal Guidelines may now occur in 20%-35% of all guilty plea cases). Even if the Court's holding does not further embed plea-bargaining practices (as I fear it will), its success depends upon the existence of present practice. I do not understand how the Sixth Amendment could require a sentencing system that will work in practice only if no more than a handful of defendants exercise their right to a jury trial.

[128] The majority's only response is to state that "bargaining over elements . . . probably favors the defendant," ante, at 15, adding that many criminal defense lawyers favor its position, ante, at 16. But the basic problem is not one of "fairness" to defendants or, for that matter, "fairness" to prosecutors. Rather, it concerns the greater fairness of a sentencing system that a more uniform correspondence between real criminal conduct and real punishment helps to create. At a minimum, a two-jury system, by preventing a judge from taking account of an aggravating fact without the prosecutor's acquiescence, would undercut, if not nullify, legislative efforts to ensure through guidelines that punishments reflect a convicted offender's real criminal conduct, rather than that portion of the offender's conduct that a prosecutor decides to charge and prove.

[129] Efforts to tie real punishment to real conduct are not new. They are embodied in well-established pre-guidelines sentencing practices --practices under which a judge, looking at a presentence report, would seek to tailor the sentence in significant part to fit the criminal conduct in which the offender actually engaged. For more than a century, questions of punishment (not those of guilt or innocence) have reflected determinations made, not only by juries, but also by judges, probation officers, and executive parole boards. Such truth-seeking determinations have rested upon both adversarial and non-adversarial processes. The Court's holding undermines efforts to reform these processes, for it means that legislatures cannot both permit judges to base sentencing upon real conduct and seek, through guidelines, to make the results more uniform.

[130] In these and other ways, the two-jury system would work a radical change in pre-existing criminal law. It is not surprising that this Court has never previously suggested that the Constitution -- outside the unique context of the death penalty -- might require bifurcated jury-based sentencing. And it is the impediment the Court's holding poses to legislative efforts to achieve that greater systematic fairness that casts doubt on its constitutional validity.

[131] D.

[132] Is there a fourth option? Perhaps. Congress and state legislatures might, for example, rewrite their criminal codes, attaching astronomically high sentences to each crime, followed by long lists of mitigating facts, which, for the most part, would consist of the absence of aggravating facts. Apprendi, 530 U. S., at 541-542 (O'Connor, J., dissenting) (explaining how legislatures can evade the majority's rule by making yet another labeling choice). But political impediments to legislative action make such rewrites difficult to achieve; and it is difficult to see why the Sixth Amendment would require legislatures to undertake them.

[133] It may also prove possible to find combinations of, or variations upon, my first three options. But I am unaware of any variation that does not involve (a) the shift of power to the prosecutor (weakening the connection between real conduct and real punishment) inherent in any charge offense system, (b) the lack of uniformity inherent in any system of pure judicial discretion, or (c) the complexity, expense, and increased reliance on plea bargains involved in a "two-jury" system. The simple fact is that the design of any fair sentencing system must involve efforts to make practical compromises among competing goals. The majority's reading of the Sixth Amendment makes the effort to find those compromises -- already difficult -- virtually impossible.

[134] II.

[135] The majority rests its conclusion in significant part upon a claimed historical (and therefore constitutional) imperative. According to the majority, the rule it applies in this case is rooted in "longstanding tenets of common-law criminal jurisprudence," ante, at 5: that every accusation against a defendant must be proved to a jury and that " `an accusation which lacks any particular fact which the law makes essential to the punishment is . . . no accusation within the requirements of the common law, and it is no accusation in reason,' " ibid. (quoting Bishop, Criminal Procedure §87, at 55). The historical sources upon which the majority relies, however, do not compel the result it reaches. See ante, at 10 (O'Connor, J., dissenting); Apprendi, 530 U. S., at 525-528 (O'Connor, J., dissenting). The quotation from Bishop, to which the majority attributes great weight, stands for nothing more than the "unremarkable proposition" that where a legislature passes a statute setting forth heavier penalties than were available for committing a common-law offense and specifying those facts that triggered the statutory penalty, "a defendant could receive the greater statutory punishment only if the indictment expressly charged and the prosecutor proved the facts that made up the statutory offense, as opposed to simply those facts that made up the common-law offense." Id., at 526 (O'Connor, J., dissenting) (characterizing a similar statement of the law in J. Archbold, Pleading and Evidence in Criminal Cases 51, 188 (15th ed. 1862)).

[136] This is obvious when one considers the problem that Bishop was addressing. He provides as an example "statutes whereby, when [a common-law crime] is committed with a particular intent, or with a particular weapon, or the like, it is subjected to a particular corresponding punishment, heavier than that for" the simple common-law offense (though, of course, his concerns were not "limited to that example," ante, at 5-6, n. 5). Bishop, supra, §82, at 51-52 (discussing the example of common assault and enhanced-assault statutes, e.g., "assaults committed with the intent to rob"). That indictments historically had to charge all of the statutorily labeled elements of the offense is a proposition on which all can agree. See Apprendi, supra, at 526-527 (O'Connor, J., dissenting). See also J. Archbold, Pleading and Evidence in Criminal Cases 44 (11th ed. 1849) ("[E]very fact or circumstance which is a necessary ingredient in the offence must be set forth in the indictment" so that "there may be no doubt as to the judgment which should be given, if the defendant be convicted"); 1 T. Starkie, Criminal Pleading 68 (2d ed. 1822) (the indictment must state "the criminal nature and degree of the offence, which are conclusions of law from the facts; and also the particular facts and circumstances which render the defendant guilty of that offence").

[137] Neither Bishop nor any other historical treatise writer, however, disputes the proposition that judges historically had discretion to vary the sentence, within the range provided by the statute, based on facts not proved at the trial. See Bishop, supra, §85, at 54 ("[W]ithin the limits of any discretion as to the punishment which the law may have allowed, the judge, when he pronounces sentence, may suffer his discretion to be influenced by matter shown in aggravation or mitigation, not covered by the allegations of the indictment"); K. Stith & J. Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts 9 (1998). The modern history of pre-guidelines sentencing likewise indicates that judges had broad discretion to set sentences within a statutory range based on uncharged conduct. Usually, the judge based his or her sentencing decision on facts gleaned from a presentence report, which the defendant could dispute at a sentencing hearing. In the federal system, for example, Federal Rule of Criminal Procedure 32 provided that probation officers, who are employees of the Judicial Branch, prepared a presentence report for the judge, a copy of which was generally given to the prosecution and defense before the sentencing hearing. See Stith & Cabranes, supra, at 79-80, 221, note 5. See also ante, at 2 (O'Connor, J., dissenting) (describing the State of Washington's former indeterminate sentencing law).

[138] In this case, the statute provides that kidnaping may be punished by up to 10 years' imprisonment. Wash. Rev. Code Ann. §§9A.40.030(3), 9A.20.021(1)(b) (2000). Modern structured sentencing schemes like Washington's do not change the statutorily fixed maximum penalty, nor do they purport to establish new elements for the crime. Instead, they undertake to structure the previously unfettered discretion of the sentencing judge, channeling and limiting his or her discretion even within the statutory range. (Thus, contrary to the majority's arguments, ante, at 12-13, kidnapers in the State of Washington know that they risk up to 10 years' imprisonment, but they also have the benefit of additional information about how long -- within the 10-year maximum -- their sentences are likely to be, based on how the kidnaping was committed.)

[139] Historical treatises do not speak to such a practice because it was not done in the 19th century. Cf. Jones, 526 U. S., at 244 ("[T]he scholarship of which we are aware does not show that a question exactly like this one was ever raised and resolved in the period before the framing"). This makes sense when one considers that, prior to the 19th century, the prescribed penalty for felonies was often death, which the judge had limited, and sometimes no, power to vary. See Lillquist, 82 N. C. L. Rev., at 628-630. The 19th century saw a movement to a rehabilitative mode of punishment in which prison terms became a norm, shifting power to the judge to impose a longer or shorter term within the statutory maximum. See ibid. The ability of legislatures to guide the judge's discretion by designating presumptive ranges, while allowing the judge to impose a more or less severe penalty in unusual cases, was therefore never considered. To argue otherwise, the majority must ignore the significant differences between modern structured sentencing schemes and the history on which it relies to strike them down. And while the majority insists that the historical sources, particularly Bishop, should not be "limited" to the context in which they were written, ante, at 5-6, n. 5, it has never explained why the Court must transplant those discussions to the very different context of sentencing schemes designed to structure judges' discretion within a statutory sentencing range.

[140] Given history's silence on the question of laws that structure a judge's discretion within the range provided by the legislatively labeled maximum term, it is not surprising that our modern, pre-Apprendi cases made clear that legislatures could, within broad limits, distinguish between "sentencing facts" and "elements of crimes." See McMillan, 477 U. S., at 85-88. By their choice of label, legislatures could indicate whether a judge or a jury must make the relevant factual determination. History does not preclude legislatures from making this decision. And, as I argued in Part I, supra, allowing legislatures to structure sentencing in this way has the dual effect of enhancing and giving meaning to the Sixth Amendment's jury trial right as to core crimes, while affording additional due process to defendants in the form of sentencing hearings before judges -- hearings the majority's rule will eliminate for many.

[141] Is there a risk of unfairness involved in permitting Congress to make this labeling decision? Of course. As we have recognized, the "tail" of the sentencing fact might "wa[g] the dog of the substantive offense." McMillan, supra, at 88. Congress might permit a judge to sentence an individual for murder though convicted only of making an illegal lane change. See ante, at 10 (majority opinion). But that is the kind of problem that the Due Process Clause is well suited to cure. McMillan foresaw the possibility that judges would have to use their own judgment in dealing with such a problem; but that is what judges are there for. And, as Part I, supra, makes clear, the alternatives are worse -- not only practically, but, although the majority refuses to admit it, constitutionally as well.

[142] Historic practice, then, does not compel the result the majority reaches. And constitutional concerns counsel the opposite.

[143] III.

[144] The majority also overlooks important institutional considerations. Congress and the States relied upon what they believed was their constitutional power to decide, within broad limits, whether to make a particular fact (a) a sentencing factor or (b) an element in a greater crime. They relied upon McMillan as guaranteeing the constitutional validity of that proposition. They created sentencing reform, an effort to change the criminal justice system so that it reflects systematically not simply upon guilt or innocence but also upon what should be done about this now-guilty offender. Those efforts have spanned a generation. They have led to state sentencing guidelines and the Federal Sentencing Guideline system. E.g., ante, at 2-4 (O'Connor, J., dissenting) (describing sentencing reform in the State of Washington). These systems are imperfect and they yield far from perfect results, but I cannot believe the Constitution forbids the state legislatures and Congress to adopt such systems and to try to improve them over time. Nor can I believe that the Constitution hamstrings legislatures in the way that Justice O'Connor and I have discussed.

[145] IV.

[146] Now, let us return to the question I posed at the outset. Why does the Sixth Amendment permit a jury trial right (in respect to a particular fact) to depend upon a legislative labeling decision, namely, the legislative decision to label the fact a sentencing fact, instead of an element of the crime? The answer is that the fairness and effectiveness of a sentencing system, and the related fairness and effectiveness of the criminal justice system itself, depends upon the legislature's possessing the constitutional authority (within due process limits) to make that labeling decision. To restrict radically the legislature's power in this respect, as the majority interprets the Sixth Amendment to do, prevents the legislature from seeking sentencing systems that are consistent with, and indeed may help to advance, the Constitution's greater fairness goals.

[147] To say this is not simply to express concerns about fairness to defendants. It is also to express concerns about the serious practical (or impractical) changes that the Court's decision seems likely to impose upon the criminal process; about the tendency of the Court's decision to embed further plea bargaining processes that lack transparency and too often mean non-uniform, sometimes arbitrary, sentencing practices; about the obstacles the Court's decision poses to legislative efforts to bring about greater uniformity between real criminal conduct and real punishment; and ultimately about the limitations that the Court imposes upon legislatures' ability to make democratic legislative decisions. Whatever the faults of guidelines systems -- and there are many -- they are more likely to find their cure in legislation emerging from the experience of, and discussion among, all elements of the criminal justice community, than in a virtually unchangeable constitutional decision of this Court.

[148] V.

[149] Taken together these three sets of considerations, concerning consequences, concerning history, concerning institutional reliance, leave me where I was in Apprendi, i.e., convinced that the Court is wrong. Until now, I would have thought the Court might have limited Apprendi so that its underlying principle would not undo sentencing reform efforts. Today's case dispels that illusion. At a minimum, the case sets aside numerous state efforts in that direction. Perhaps the Court will distinguish the Federal Sentencing Guidelines, but I am uncertain how. As a result of today's decision, federal prosecutors, like state prosecutors, must decide what to do next, how to handle tomorrow's case.

[150] Consider some of the matters that federal prosecutors must know about, or guess about, when they prosecute their next case: (1) Does today's decision apply in full force to the Federal Sentencing Guidelines? (2) If so, must the initial indictment contain all sentencing factors, charged as "elements" of the crime? (3) What, then, are the evidentiary rules? Can the prosecution continue to use, say presentence reports, with their conclusions reflecting layers of hearsay? Cf. Crawford v. Washington, 541 U. S. __, __, __-__ (2004) (slip op., at 27, 32-33) (clarifying the Sixth Amendment's requirement of confrontation with respect to testimonial hearsay). Are the numerous cases of this Court holding that a sentencing judge may consider virtually any reliable information still good law when juries, not judges, are required to determine the matter? See, e.g., United States v. Watts, 519 U. S. 148, 153-157 (1997) (per curiam) (evidence of conduct of which the defendant has been acquitted may be considered at sentencing). Cf. Witte v. United States, 515 U. S. 389, 399-401 (1995) (evidence of uncharged criminal conduct used in determining sentence). (4) How are juries to deal with highly complex or open-ended Sentencing Guidelines obviously written for application by an experienced trial judge? See, e.g., USSG §3B1.1 (requiring a greater sentence when the defendant was a leader of a criminal activity that involved four or more participants or was "otherwise extensive" (emphasis added)); §§3D1.1-3D1.2 (highly complex "multiple count" rules); §1B1.3 (relevant conduct rules).

[151] Ordinarily, this Court simply waits for cases to arise in which it can answer such questions. But this case affects tens of thousands of criminal prosecutions, including federal prosecutions. Federal prosecutors will proceed with those prosecutions subject to the risk that all defendants in those cases will have to be sentenced, perhaps tried, anew. Given this consequence and the need for certainty, I would not proceed further piecemeal; rather, I would call for further argument on the ramifications of the concerns I have raised. But that is not the Court's view.

[152] For the reasons given, I dissent.


Opinion Footnotes


[153] *fn1 Parts of Washington's criminal code have been recodified and amended. We cite throughout the provisions in effect at the time of sentencing.

[154] *fn2 Petitioner further agreed to an additional charge of second-degree assault involving domestic violence, Wash. Rev. Code Ann. §§9A.36.021(1)(c), 10.99.020(3)(b) (2000). The 14-month sentence on that count ran concurrently and is not relevant here.

[155] *fn3 The domestic-violence stipulation subjected petitioner to such measures as a "no-contact" order, see §10.99.040, but did not increase the standard range of his sentence.

[156] *fn4 The judge found other aggravating factors, but the Court of Appeals questioned their validity under state law and their independent sufficiency to support the extent of the departure. See 111 Wash. App. 851, 868-870, and n. 3, 47 P. 3d 149, 158-159, and n. 3 (2002). It affirmed the sentence solely on the finding of domestic violence with deliberate cruelty. Ibid. We therefore focus only on that factor.

[157] *fn5 Justice Breyer cites Justice O'Connor's Apprendi dissent for the point that this Bishop quotation means only that indictments must charge facts that trigger statutory aggravation of a common-law offense. Post, at 14 (dissenting opinion). Of course, as he notes, Justice O'Connor was referring to an entirely different quotation, from Archbold's treatise. See 530 U. S., at 526 (citing J. Archbold, Pleading and Evidence in Criminal Cases 51, 188 (15th ed. 1862)). Justice Breyer claims the two are "similar," post, at 14, but they are as similar as chalk and cheese. Bishop was not "addressing" the "problem" of statutes that aggravate common-law offenses. Ibid. Rather, the entire chapter of his treatise is devoted to the point that "every fact which is legally essential to the punishment" must be charged in the indictment and proved to a jury. 1 J. Bishop, Criminal Procedure, ch. 6, pp. 50-56 (2d ed. 1872). As one "example" of this principle (appearing several pages before the language we quote in text above), he notes a statute aggravating common-law assault. Id., §82, at 51-52. But nowhere is there the slightest indication that his general principle was limited to that example. Even Justice Breyer's academic supporters do not make that claim. See Bibas, Judicial Fact-Finding and Sentence Enhancements in a World of Guilty Pleas, 110 Yale L. J. 1097, 1131-1132 (2001) (conceding that Bishop's treatise supports Apprendi, while criticizing its "natural-law theorizing").

[158] *fn6 As to Justice O'Connor's criticism of the quantity of historical support for the Apprendi rule, post, at 10 (dissenting opinion): It bears repeating that the issue between us is not whether the Constitution limits States' authority to reclassify elements as sentencing factors (we all agree that it does); it is only which line, ours or hers, the Constitution draws. Criticism of the quantity of evidence favoring our alternative would have some force if it were accompanied by any evidence favoring hers. Justice O'Connor does not even provide a coherent alternative meaning for the jury-trial guarantee, unless one considers "whatever the legislature chooses to leave to the jury, so long as it does not go too far" coherent. See infra, at 9-12.

[159] *fn7 The State does not contend that the domestic-violence stipulation alone supports the departure. That the statute lists domestic violence as grounds for departure only when combined with some other aggravating factor suggests it could not. See §§9.94A.390(2)(h)(i)-(iii).

[160] *fn8 Nor does it matter that the judge must, after finding aggravating facts, make a judgment that they present a compelling ground for departure. He cannot make that judgment without finding some facts to support it beyond the bare elements of the offense. Whether the judicially determined facts require a sentence enhancement or merely allow it, the verdict alone does not authorize the sentence.

[161] *fn9 The United States, as amicus curiae, urges us to affirm. It notes differences between Washington's sentencing regime and the Federal Sentencing Guidelines but questions whether those differences are constitutionally significant. See Brief for United States as Amicus Curiae 25-30. The Federal Guidelines are not before us, and we express no opinion on them.

[162] *fn10 Justice O'Connor believes that a "built-in political check" will prevent lawmakers from manipulating offense elements in this fashion. Post, at 10. But the many immediate practical advantages of judicial factfinding, see post, at 5-7, suggest that political forces would, if anything, pull in the opposite direction. In any case, the Framers' decision to entrench the jury-trial right in the Constitution shows that they did not trust government to make political decisions in this area.

[163] *fn11 Another example of conversion from separate crime to sentence enhancement that Justice O'Connor evidently does not consider going "too far" is the obstruction-of-justice enhancement, see post, at 6-7. Why perjury during trial should be grounds for a judicial sentence enhancement on the underlying offense, rather than an entirely separate offense to be found by a jury beyond a reasonable doubt (as it has been for centuries, see 4 W. Blackstone, Commentaries on the Laws of England 136-138 (1769)), is unclear.

[164] *fn12 Justice Breyer responds that States are not required to give defendants the option of waiving jury trial on some elements but not others. Post, at 8-9. True enough. But why would the States that he asserts we are coercing into hard-heartedness -- that is, States that want judge-pronounced determinate sentencing to be the norm but we won't let them -- want to prevent a defendant from choosing that regime? Justice Breyer claims this alternative may prove "too expensive and unwieldy for States to provide," post, at 9, but there is no obvious reason why forcing defendants to choose between contesting all elements of his hypothetical 17-element robbery crime and contesting none of them is less expensive than also giving them the third option of pleading guilty to some elements and submitting the rest to judicial factfinding. Justice Breyer's argument rests entirely on a speculative prediction about the number of defendants likely to choose the first (rather than the second) option if denied the third.

[165] *fn13 To be sure, Justice Breyer and the other dissenters would forbid those increases of sentence that violate the constitutional principle that tail shall not wag dog. The source of this principle is entirely unclear. Its precise effect, if precise effect it has, is presumably to require that the ratio of sentencing-factor add-on to basic criminal sentence be no greater than the ratio of caudal vertebrae to body in the breed of canine with the longest tail. Or perhaps no greater than the average such ratio for all breeds. Or perhaps the median. Regrettably, Apprendi has prevented full development of this line of jurisprudence.

[166] *fn14 The paucity of empirical evidence regarding the impact of extending Apprendi v. New Jersey, 530 U. S. 466 (2000), to guidelines schemes should come as no surprise to the majority. Ante, at 13. Prior to today, only one court had ever applied Apprendi to invalidate application of a guidelines scheme. Compare State v. Gould, 271 Kan. 394, 23 P. 3d 801 (2001), with, e.g., United States v. Goodine, 326 F. 3d 26 (CA1 2003); United States v. Luciano, 311 F. 3d 146 (CA2 2002); United States v. DeSumma, 272 F. 3d 176 (CA3 2001); United States v. Kinter, 235 F. 3d 192 (CA4 2000); United States v. Randle, 304 F. 3d 373 (CA5 2002); United States v. Helton, 349 F. 3d 295 (CA6 2003); United States v. Johnson, 335 F. 3d 589 (CA7 2003) (per curiam); United States v. Piggie, 316 F. 3d 789 (CA8 2003); United States v. Toliver, 351 F. 3d 423 (CA9 2003); United States v. Mendez-Zamora, 296 F. 3d 1013 (CA10 2002); United States v. Sanchez, 269 F. 3d 1250 (CA11 2001); United States v. Fields, 251 F. 3d 1041 (CADC 2001); State v. Dilts, 336 Ore. 158, 82 P. 3d 593 (2003); State v. Gore, 143 Wash. 2d 288, 21 P. 3d 262 (2001); State v. Lucas, 353 N. C. 568, 548 S. E. 2d 712 (2001); State v. Dean, No. C4-02-1225, 2003 WL 21321425 (Minn. Ct. App., June 10, 2003) (unpublished opinion). Thus, there is no map of the uncharted territory blazed by today's unprecedented holding.

[167] *fn15 The numbers available from the federal system alone are staggering. On March 31, 2004, there were 8,320 federal criminal appeals pending in which the defendant's sentence was at issue. Memorandum from Carl Schlesinger, Administrative Office of the United States Courts, to Supreme Court Library (June 1, 2004) (available in Clerk of the Court's case file). Between June 27, 2000, when Apprendi was decided, and March 31, 2004, there have been 272,191 defendants sentenced in federal court. Memorandum, supra. Given that nearly all federal sentences are governed by the Federal Sentencing Guidelines, the vast majority of these cases are Guidelines cases.