DICKERSON v. UNITED STATES

Chief Justice Rehnquist delivered the opinion of the Court.

In Miranda v. Arizona, 384 U.S. 436 (1966), we held that certain warnings must be given before a suspect's statement made during custodial interrogation could be admitted in evidence. In the wake of that decision, Congress enacted 18 U. S. C. § 3501, which in essence laid down a rule that the admissibility of such statements should turn only on whether or not they were voluntarily made. We hold that Miranda, being a constitutional decision of this Court, may not be in effect overruled by an Act of Congress, and we decline to overrule Miranda ourselves. We therefore hold that Miranda and its progeny in this Court govern the admissibility of statements made during custodial interrogation in both state and federal courts.

Petitioner Dickerson was indicted for bank robbery, conspiracy to commit bank robbery, and using a firearm in the course of committing a crime of violence, all in violation of the applicable provisions of Title 18 of the United States Code. Before trial, Dickerson moved to suppress a statement he had made at a Federal Bureau of Investigation field office, on the grounds that he had not received "Miranda warnings" before being interrogated. The District Court granted his motion to suppress, and the Government took an interlocutory appeal to the United States Court of Appeals for the Fourth Circuit. That court, by a divided vote, reversed the District Court's suppression order. It agreed with the District Court's conclusion that petitioner had not received Miranda warnings before making his statement. But it went on to hold that § 3501, which in effect makes the admissibility of statements such as Dickerson's turn solely on whether they were made voluntarily, was satisfied in this case. It then concluded that our decision in Miranda was not a constitutional holding, and that, therefore, Congress could by statute have the final say on the question of admissibility. 166 F.3d 667 (1999).

Because of the importance of the questions raised by the Court of Appeals' decision, we granted certiorari, 528 U.S. 1045 (1999), and now reverse.

We begin with a brief historical account of the law governing the admission of confessions. Prior to Miranda, we evaluated the admissibility of a suspect's confession under a voluntariness test. The roots of this test developed in the common law, as the courts of England and then the United States recognized that coerced confessions are inherently untrustworthy. See, e. g., King v. Rudd, 1 Leach 115, 117-118, 122-123, 168 Eng. Rep. 160, 161, 164 (K. B. 1783) (Lord Mansfield, C. J.) (stating that the English courts excluded confessions obtained by threats and promises); King v. Warickshall, 1 Leach 262, 263-264, 168 Eng. Rep. 234, 235 (K. B. 1783) ("A free and voluntary confession is deserving of the highest credit, because it is presumed to flow from the strongest sense of guilt . . . but a confession forced from the mind by the flattery of hope, or by the torture of fear, comes in so questionable a shape . . . that no credit ought to be given to it; and therefore it is rejected"); King v. Parratt, 4 Car. & P. 570, 172 Eng. Rep. 829 (N. P. 1831); Queen v. Garner, 1 Den. 329, 169 Eng. Rep. 267 (Ct. Crim. App. 1848); Queen v. Baldry, 2 Den. 430, 169 Eng. Rep. 568 (Ct. Crim. App. 1852); Hopt v. Territory of Utah, 110 U.S. 574 (1884); Pierce v. United States, 160 U.S. 355, 357 (1896). Over time, our cases recognized two constitutional bases for the requirement that a confession be voluntary to be admitted into evidence: the Fifth Amendment right against self-incrimination and the Due Process Clause of the Fourteenth Amendment. See, e. g., Bram v. United States, 168 U.S. 532, 542 (1897) (stating that the voluntariness test "is controlled by that portion of the Fifth Amendment . . . commanding that no person `shall be compelled in any criminal case to be a witness against himself' "); Brown v. Mississippi, 297 U.S. 278 (1936) (reversing a criminal conviction under the Due Process Clause because it was based on a confession obtained by physical coercion).

While Bram was decided before Brown and its progeny, for the middle third of the 20th century our cases based the rule against admitting coerced confessions primarily, if not exclusively, on notions of due process. We applied the due process voluntariness test in "some 30 different cases decided during the era that intervened between Brown and Escobedo v. Illinois, 378 U.S. 478 [(1964)]." Schneckloth v. Bustamonte, 412 U.S. 218, 223 (1973). See, e. g., Haynes v. Washington, 373 U.S. 503 (1963); Ashcraft v. Tennessee, 322 U.S. 143 (1944); Chambers v. Florida, 309 U.S. 227 (1940). Those cases refined the test into an inquiry that examines "whether a defendant's will was overborne" by the circumstances surrounding the giving of a confession. Schneckloth, 412 U. S., at 226. The due process test takes into consideration "the totality of all the surrounding circumstances—both the characteristics of the accused and the details of the interrogation." Ibid. See also Haynes, supra, at 513; Gallegos v. Colorado, 370 U.S. 49, 55 (1962); Reck v. Pate, 367 U.S. 433, 440 (1961) ("[A]ll the circumstances attendant upon the confession must be taken into account"); Malinski v. New York, 324 U.S. 401, 404 (1945) ("If all the attendant circumstances indicate that the confession was coerced or compelled, it may not be used to convict a defendant"). The determination "depend[s] upon a weighing of the circumstances of pressure against the power of resistance of the person confessing." Stein v. New York, 346 U.S. 156, 185 (1953).

We have never abandoned this due process jurisprudence, and thus continue to exclude confessions that were obtained involuntarily. But our decisions in Malloy v. Hogan, 378 U.S. 1 (1964), and Miranda changed the focus of much of the inquiry in determining the admissibility of suspects' incriminating statements. In Malloy, we held that the Fifth Amendment's Self-Incrimination Clause is incorporated in the Due Process Clause of the Fourteenth Amendment and thus applies to the States. 378 U. S., at 6-11. We decided Miranda on the heels of Malloy.

In Miranda, we noted that the advent of modern custodial police interrogation brought with it an increased concern about confessions obtained by coercion.1 384 U. S., at 445-458. Because custodial police interrogation, by its very nature, isolates and pressures the individual, we stated that "[e]ven without employing brutality, the `third degree' or [other] specific stratagems, . . . custodial interrogation exacts a heavy toll on individual liberty and trades on the weakness of individuals." Id., at 455. We concluded that the coercion inherent in custodial interrogation blurs the line between voluntary and involuntary statements, and thus heightens the risk that an individual will not be "accorded his privilege under the Fifth Amendment . . . not to be compelled to incriminate himself." Id., at 439. Accordingly, we laid down "concrete constitutional guidelines for law enforcement agencies and courts to follow." Id., at 442. Those guidelines established that the admissibility in evidence of any statement given during custodial interrogation of a suspect would depend on whether the police provided the suspect with four warnings. These warnings (which have come to be known colloquially as "Miranda rights") are: a suspect "has the right to remain silent, that anything he says can be used against him in a court of law, that he has the right to the presence of an attorney, and that if he cannot afford an attorney one will be appointed for him prior to any questioning if he so desires." Id., at 479.

Two years after Miranda was decided, Congress enacted § 3501. That section provides, in relevant part:

The law in this area is clear. This Court has supervisory authority over the federal courts, and we may use that authority to prescribe rules of evidence and procedure that are binding in those tribunals. Carlisle v. United States, 517 U.S. 416, 426 (1996). However, the power to judicially create and enforce nonconstitutional "rules of procedure and evidence for the federal courts exists only in the absence of a relevant Act of Congress." Palermo v. United States, 360 U.S. 343, 353, n. 11 (1959) (citing Funk v. United States, 290 U.S. 371, 382 (1933), and Gordon v. United States, 344 U.S. 414, 418 (1953)). Congress retains the ultimate authority to modify or set aside any judicially created rules of evidence and procedure that are not required by the Constitution. Palermo, supra, at 345-348; Carlisle, supra, at 426; Vance v. Terrazas, 444 U.S. 252, 265 (1980).

But Congress may not legislatively supersede our decisions interpreting and applying the Constitution. See, e. g., City of Boerne v. Flores, 521 U.S. 507, 517-521 (1997). This case therefore turns on whether the Miranda Court announced a constitutional rule or merely exercised its supervisory authority to regulate evidence in the absence of congressional direction. Recognizing this point, the Court of Appeals surveyed Miranda and its progeny to determine the constitutional status of the Miranda decision. 166 F. 3d, at 687-692. Relying on the fact that we have created several exceptions to Miranda `s warnings requirement and that we have repeatedly referred to the Miranda warnings as "prophylactic," New York v. Quarles, 467 U.S. 649, 653 (1984), and "not themselves rights protected by the Constitution," Michigan v. Tucker, 417 U.S. 433, 444 (1974),2 the Court of Appeals concluded that the protections announced in Miranda are not constitutionally required. 166 F. 3d, at 687-690.

We disagree with the Court of Appeals' conclusion, although we concede that there is language in some of our opinions that supports the view taken by that court. But first and foremost of the factors on the other side—that Miranda is a constitutional decision—is that both Miranda and two of its companion cases applied the rule to proceedings in state courts—to wit, Arizona, California, and New York. See 384 U. S., at 491-494, 497-499. Since that time, we have consistently applied Miranda `s rule to prosecutions arising in state courts. See, e. g., Stansbury v. California, 511 U.S. 318 (1994) (per curiam); Minnick v. Mississippi, 498 U.S. 146 (1990); Arizona v. Roberson, 486 U.S. 675 (1988); Edwards v. Arizona, 451 U.S. 477, 481-482 (1981). It is beyond dispute that we do not hold a supervisory power over the courts of the several States. Smith v. Phillips, 455 U.S. 209, 221 (1982) ("Federal courts hold no supervisory authority over state judicial proceedings and may intervene only to correct wrongs of constitutional dimension"); Cicenia v. Lagay, 357 U.S. 504, 508-509 (1958). With respect to proceedings in state courts, our "authority is limited to enforcing the commands of the United States Constitution." Mu'Min v. Virginia, 500 U.S. 415, 422 (1991). See also Harris v. Rivera, 454 U.S. 339, 344-345 (1981) (per curiam) (stating that "[f]ederal judges . . . may not require the observance of any special procedures" in state courts "except when necessary to assure compliance with the dictates of the Federal Constitution").3

The Miranda opinion itself begins by stating that the Court granted certiorari "to explore some facets of the problems . . . of applying the privilege against self-incrimination to in-custody interrogation, and to give concrete constitutional guidelines for law enforcement agencies and courts to follow. " 384 U. S., at 441-442 (emphasis added). In fact, the majority opinion is replete with statements indicating that the majority thought it was announcing a constitutional rule.4 Indeed, the Court's ultimate conclusion was that the unwarned confessions obtained in the four cases before the Court in Miranda "were obtained from the defendant under circumstances that did not meet constitutional standards for protection of the privilege."5 Id., at 491.

Additional support for our conclusion that Miranda is constitutionally based is found in the Miranda Court's invitation for legislative action to protect the constitutional right against coerced self-incrimination. After discussing the "compelling pressures" inherent in custodial police interrogation, the Miranda Court concluded that, "[i]n order to combat these pressures and to permit a full opportunity to exercise the privilege against self-incrimination, the accused must be adequately and effectively apprised of his rights and the exercise of those rights must be fully honored." Id., at 467. However, the Court emphasized that it could not foresee "the potential alternatives for protecting the privilege which might be devised by Congress or the States," and it accordingly opined that the Constitution would not preclude legislative solutions that differed from the prescribed Miranda warnings but which were "at least as effective in apprising accused persons of their right of silence and in assuring a continuous opportunity to exercise it."6 Ibid.

The Court of Appeals also relied on the fact that we have, after our Miranda decision, made exceptions from its rule in cases such as New York v. Quarles, 467 U.S. 649 (1984), and Harris v. New York, 401 U.S. 222 (1971). See 166 F. 3d, at 672, 689-691. But we have also broadened the application of the Miranda doctrine in cases such as Doyle v. Ohio, 426 U.S. 610 (1976), and Arizona v. Roberson, 486 U.S. 675 (1988). These decisions illustrate the principle—not that Miranda is not a constitutional rule—but that no constitutional rule is immutable. No court laying down a general rule can possibly foresee the various circumstances in which counsel will seek to apply it, and the sort of modifications represented by these cases are as much a normal part of constitutional law as the original decision.

The Court of Appeals also noted that in Oregon v. Elstad, 470 U.S. 298 (1985), we stated that "`[t]he Miranda exclusionary rule . . . serves the Fifth Amendment and sweeps more broadly than the Fifth Amendment itself.' " 166 F. 3d, at 690 (quoting Elstad, supra, at 306). Our decision in that case—refusing to apply the traditional "fruits" doctrine developed in Fourth Amendment cases—does not prove that Miranda is a nonconstitutional decision, but simply recognizes the fact that unreasonable searches under the Fourth Amendment are different from unwarned interrogation under the Fifth Amendment.

As an alternative argument for sustaining the Court of Appeals' decision, the court-invited amicus curiae7 contends that the section complies with the requirement that a legislative alternative to Miranda be equally as effective in preventing coerced confessions. See Brief for Paul G. Cassell as Amicus Curiae 28-39. We agree with the amicus ` contention that there are more remedies available for abusive police conduct than there were at the time Miranda was decided, see, e. g., Wilkins v. May, 872 F.2d 190, 194 (CA7 1989) (applying Bivens v. Six Unknown Fed. Narcotics Agents, 403 U.S. 388 (1971), to hold that a suspect may bring a federal cause of action under the Due Process Clause for police misconduct during custodial interrogation). But we do not agree that these additional measures supplement § 3501's protections sufficiently to meet the constitutional minimum. Miranda requires procedures that will warn a suspect in custody of his right to remain silent and which will assure the suspect that the exercise of that right will be honored. See, e. g., 384 U. S., at 467. As discussed above, § 3501 explicitly eschews a requirement of preinterrogation warnings in favor of an approach that looks to the administration of such warnings as only one factor in determining the voluntariness of a suspect's confession. The additional remedies cited by amicus do not, in our view, render them, together with § 3501, an adequate substitute for the warnings required by Miranda.

The dissent argues that it is judicial overreaching for this Court to hold § 3501 unconstitutional unless we hold that the Miranda warnings are required by the Constitution, in the sense that nothing else will suffice to satisfy constitutional requirements. Post, at 453-454, 465 (opinion of Scalia, J.). But we need not go further than Miranda to decide this case. In Miranda, the Court noted that reliance on the traditional totality-of-the-circumstances test raised a risk of overlooking an involuntary custodial confession, 384 U. S, at 457, a risk that the Court found unacceptably great when the confession is offered in the case in chief to prove guilt. The Court therefore concluded that something more than the totality test was necessary. See ibid.; see also id., at 467, 490-491. As discussed above, § 3501 reinstates the totality test as sufficient. Section 3501 therefore cannot be sustained if Miranda is to remain the law.

Whether or not we would agree with Miranda `s reasoning and its resulting rule, were we addressing the issue in the first instance, the principles of stare decisis weigh heavily against overruling it now. See, e. g., Rhode Island v. Innis, 446 U.S. 291, 304 (1980) (Burger, C. J., concurring in judgment) ("The meaning of Miranda has become reasonably clear and law enforcement practices have adjusted to its strictures; I would neither overrule Miranda, disparage it, nor extend it at this late date"). While "`stare decisis is not an inexorable command,' " State Oil Co. v. Khan, 522 U.S. 3, 20 (1997) (quoting Payne v. Tennessee, 501 U.S. 808, 828 (1991)), particularly when we are interpreting the Constitution, Agostini v. Felton, 521 U.S. 203, 235 (1997), "even in constitutional cases, the doctrine carries such persuasive force that we have always required a departure from precedent to be supported by some `special justification.' " United States v. International Business Machines Corp., 517 U.S. 843, 856 (1996) (quoting Payne, supra, at 842 (Souter, J., concurring), in turn quoting Arizona v. Rumsey, 467 U.S. 203, 212 (1984)).

We do not think there is such justification for overruling Miranda. Miranda has become embedded in routine police practice to the point where the warnings have become part of our national culture. See Mitchell v. United States, 526 U.S. 314, 331-332 (1999) (Scalia, J., dissenting) (stating that the fact that a rule has found "`wide acceptance in the legal culture' " is "adequate reason not to overrule" it). While we have overruled our precedents when subsequent cases have undermined their doctrinal underpinnings, see, e. g., Patterson v. McLean Credit Union, 491 U.S. 164, 173 (1989), we do not believe that this has happened to the Miranda decision. If anything, our subsequent cases have reduced the impact of the Miranda rule on legitimate law enforcement while reaffirming the decision's core ruling that unwarned statements may not be used as evidence in the prosecution's case in chief.

The disadvantage of the Miranda rule is that statements which may be by no means involuntary, made by a defendant who is aware of his "rights," may nonetheless be excluded and a guilty defendant go free as a result. But experience suggests that the totality-of-the-circumstances test which § 3501 seeks to revive is more difficult than Miranda for law enforcement officers to conform to, and for courts to apply in a consistent manner. See, e. g., Haynes v. Washington, 373 U. S., at 515 ("The line between proper and permissible police conduct and techniques and methods offensive to due process is, at best, a difficult one to draw"). The requirement that Miranda warnings be given does not, of course, dispense with the voluntariness inquiry. But as we said in Berkemer v. McCarty, 468 U.S. 420 (1984), "[c]ases in which a defendant can make a colorable argument that a self-incriminating statement was `compelled' despite the fact that the law enforcement authorities adhered to the dictates of Miranda are rare." Id., at 433, n. 20.

In sum, we conclude that Miranda announced a constitutional rule that Congress may not supersede legislatively. Following the rule of stare decisis, we decline to overrule Miranda ourselves.8 The judgment of the Court of Appeals is therefore

Those to whom judicial decisions are an unconnected series of judgments that produce either favored or disfavored results will doubtless greet today's decision as a paragon of moderation, since it declines to overrule Miranda v. Arizona, 384 U.S. 436 (1966). Those who understand the judicial process will appreciate that today's decision is not a reaffirmation of Miranda, but a radical revision of the most significant element of Miranda (as of all cases): the rationale that gives it a permanent place in our jurisprudence.

Marbury v. Madison, 1 Cranch 137 (1803), held that an Act of Congress will not be enforced by the courts if what it prescribes violates the Constitution of the United States. That was the basis on which Miranda was decided. One will search today's opinion in vain, however, for a statement (surely simple enough to make) that what 18 U. S. C. § 3501 prescribes—the use at trial of a voluntary confession, even when a Miranda warning or its equivalent has failed to be given—violates the Constitution. The reason the statement does not appear is not only (and perhaps not so much) that it would be absurd, inasmuch as § 3501 excludes from trial precisely what the Constitution excludes from trial, viz., compelled confessions; but also that Justices whose votes are needed to compose today's majority are on record as believing that a violation of Miranda is not a violation of the Constitution. See Davis v. United States, 512 U.S. 452, 457-458 (1994) (opinion of the Court, in which Kennedy, J., joined); Duckworth v. Eagan, 492 U.S. 195, 203 (1989) (opinion of the Court, in which Kennedy, J., joined); Oregon v. Elstad, 470 U.S. 298 (1985) (opinion of the Court by O'Connor, J.); New York v. Quarles, 467 U.S. 649 (1984) (opinion of the Court by Rehnquist, J.). And so, to justify today's agreed-upon result, the Court must adopt a significant new, if not entirely comprehensible, principle of constitutional law. As the Court chooses to describe that principle, statutes of Congress can be disregarded, not only when what they prescribe violates the Constitution, but when what they prescribe contradicts a decision of this Court that "announced a constitutional rule," ante, at 437. As I shall discuss in some detail, the only thing that can possibly mean in the context of this case is that this Court has the power, not merely to apply the Constitution but to expand it, imposing what it regards as useful "prophylactic" restrictions upon Congress and the States. That is an immense and frightening antidemocratic power, and it does not exist.

It takes only a small step to bring today's opinion out of the realm of power-judging and into the mainstream of legal reasoning: The Court need only go beyond its carefully couched iterations that "Miranda is a constitutional decision," ante, at 438, that "Miranda is constitutionally based," ante, at 440, that Miranda has "constitutional underpinnings," ante, at 440, n. 5, and come out and say quite clearly: "We reaffirm today that custodial interrogation that is not preceded by Miranda warnings or their equivalent violates the Constitution of the United States." It cannot say that, because a majority of the Court does not believe it. The Court therefore acts in plain violation of the Constitution when it denies effect to this Act of Congress.

I

Early in this Nation's history, this Court established the sound proposition that constitutional government in a system of separated powers requires judges to regard as inoperative any legislative Act, even of Congress itself, that is "repugnant to the Constitution."

It was once possible to characterize the so-called Miranda rule as resting (however implausibly) upon the proposition that what the statute here before us permits—the admission at trial of un-Mirandized confessions—violates the Constitution. That is the fairest reading of the Miranda case itself. The Court began by announcing that the Fifth Amendment privilege against self-incrimination applied in the context of extrajudicial custodial interrogation, see 384 U. S., at 460— 467—itself a doubtful proposition as a matter both of history and precedent, see id., at 510-511 (Harlan, J., dissenting) (characterizing the Court's conclusion that the Fifth Amendment privilege, rather than the Due Process Clause, governed station house confessions as a "trompe l'oeil "). Having extended the privilege into the confines of the station house, the Court liberally sprinkled throughout its sprawling 60-page opinion suggestions that, because of the compulsion inherent in custodial interrogation, the privilege was violated by any statement thus obtained that did not conform to the rules set forth in Miranda, or some functional equivalent. See id., at 458 ("Unless adequate protective devices are employed to dispel the compulsion inherent in custodial surroundings, no statement obtained from the defendant can truly be the product of his free choice" (emphases added)); id., at 461 ("An individual swept from familiar surroundings into police custody, surrounded by antagonistic forces, and subjected to the techniques of persuasion described above cannot be otherwise than under compulsion to speak"); id., at 467 ("We have concluded that without proper safeguards the process of in-custody interrogation . . . contains inherently compelling pressures which work to undermine the individual's will to resist and to compel him to speak where he would not otherwise do so freely"); id., at 457, n. 26 (noting the "absurdity of denying that a confession obtained under these circumstances is compelled").

The dissenters, for their part, also understood Miranda `s holding to be based on the "premise . . . that pressure on the suspect must be eliminated though it be only the subtle influence of the atmosphere and surroundings." Id., at 512 (Harlan, J., dissenting). See also id., at 535 (White, J., dissenting) ("[I]t has never been suggested, until today, that such questioning was so coercive and accused persons so lacking in hardihood that the very first response to the very first question following the commencement of custody must be conclusively presumed to be the product of an overborne will"). And at least one case decided shortly after Miranda explicitly confirmed the view. See Orozco v. Texas, 394 U.S. 324, 326 (1969) ("[T]he use of these admissions obtained in the absence of the required warnings was a flat violation of the Self-Incrimination Clause of the Fifth Amendment as construed in Miranda ").

So understood, Miranda was objectionable for innumerable reasons, not least the fact that cases spanning more than 70 years had rejected its core premise that, absent the warnings and an effective waiver of the right to remain silent and of the (thitherto unknown) right to have an attorney present, a statement obtained pursuant to custodial interrogation was necessarily the product of compulsion. See Crooker v. California, 357 U.S. 433 (1958) (confession not involuntary despite denial of access to counsel); Cicenia v. Lagay, 357 U.S. 504 (1958) (same); Powers v. United States, 223 U.S. 303 (1912) (lack of warnings and counsel did not render statement before United States Commissioner involuntary); Wilson v. United States, 162 U.S. 613 (1896) (same). Moreover, history and precedent aside, the decision in Miranda, if read as an explication of what the Constitution requires, is preposterous. There is, for example, simply no basis in reason for concluding that a response to the very first question asked, by a suspect who already knows all of the rights described in the Miranda warning, is anything other than a volitional act. See Miranda, supra, at 533-534 (White, J., dissenting). And even if one assumes that the elimination of compulsion absolutely requires informing even the most knowledgeable suspect of his right to remain silent, it cannot conceivably require the right to have counsel present. There is a world of difference, which the Court recognized under the traditional voluntariness test but ignored in Miranda, between compelling a suspect to incriminate himself and preventing him from foolishly doing so of his own accord. Only the latter (which is not required by the Constitution) could explain the Court's inclusion of a right to counsel and the requirement that it, too, be knowingly and intelligently waived. Counsel's presence is not required to tell the suspect that he need not speak; the interrogators can do that. The only good reason for having counsel there is that he can be counted on to advise the suspect that he should not speak. See Watts v. Indiana, 338 U.S. 49, 59 (1949) (Jackson, J., concurring in result in part and dissenting in part) ("[A]ny lawyer worth his salt will tell the suspect in no uncertain terms to make no statement to police under any circumstances").

Preventing foolish (rather than compelled) confessions is likewise the only conceivable basis for the rules (suggested in Miranda, see 384 U. S., at 444-445, 473-474), that courts must exclude any confession elicited by questioning conducted, without interruption, after the suspect has indicated a desire to stand on his right to remain silent, see Michigan v. Mosley, 423 U.S. 96, 105-106 (1975), or initiated by police after the suspect has expressed a desire to have counsel present, see Edwards v. Arizona, 451 U.S. 477, 484— 485 (1981). Nonthreatening attempts to persuade the suspect to reconsider that initial decision are not, without more, enough to render a change of heart the product of anything other than the suspect's free will. Thus, what is most remarkable about the Miranda decision—and what made it unacceptable as a matter of straightforward constitutional interpretation in the Marbury tradition—is its palpable hostility toward the act of confession per se, rather than toward what the Constitution abhors, compelled confession. See United States v. Washington, 431 U.S. 181, 187 (1977) ("[F]ar from being prohibited by the Constitution, admissions of guilt by wrongdoers, if not coerced, are inherently desirable"). The Constitution is not, unlike the Miranda majority, offended by a criminal's commendable qualm of conscience or fortunate fit of stupidity. Cf. Minnick v. Mississippi, 498 U.S. 146, 166-167 (1990) (Scalia, J., dissenting).

For these reasons, and others more than adequately developed in the Miranda dissents and in the subsequent works of the decision's many critics, any conclusion that a violation of the Miranda rules necessarily amounts to a violation of the privilege against compelled self-incrimination can claim no support in history, precedent, or common sense, and as a result would at least presumptively be worth reconsidering even at this late date. But that is unnecessary, since the Court has (thankfully) long since abandoned the notion that failure to comply with Miranda `s rules is itself a violation of the Constitution.

II

As the Court today acknowledges, since Miranda we have explicitly, and repeatedly, interpreted that decision as having announced, not the circumstances in which custodial interrogation runs afoul of the Fifth or Fourteenth Amendment, but rather only "prophylactic" rules that go beyond the right against compelled self-incrimination. Of course the seeds of this "prophylactic" interpretation of Miranda were present in the decision itself. See Miranda, 384 U. S., at 439 (discussing the "necessity for procedures which assure that the [suspect] is accorded his privilege"); id., at 447 ("[u]nless a proper limitation upon custodial interrogation is achieved— such as these decisions will advance—there can be no assurance that practices of this nature will be eradicated"); id., at 457 ("[i]n these cases, we might not find the defendants' statements to have been involuntary in traditional terms"); ibid. (noting "concern for adequate safeguards to protect precious Fifth Amendment rights" and the "potentiality for compulsion" in Ernesto Miranda's interrogation). In subsequent cases, the seeds have sprouted and borne fruit: The Court has squarely concluded that it is possible—indeed not uncommon—for the police to violate Miranda without also violating the Constitution.

Michigan v. Tucker, 417 U.S. 433 (1974), an opinion for the Court written by then-Justice Rehnquist, rejected the true-to-Marbury, failure-to-warn-as-constitutional-violation interpretation of Miranda. It held that exclusion of the "fruits" of a Miranda violation—the statement of a witness whose identity the defendant had revealed while in custody—was not required. The opinion explained that the question whether the "police conduct complained of directly infringed upon respondent's right against compulsory selfincrimination" was a "separate question" from "whether it instead violated only the prophylactic rules developed to protect that right." 417 U. S., at 439. The "procedural safeguards" adopted in Miranda, the Court said, "were not themselves rights protected by the Constitution but were instead measures to insure that the right against compulsory self-incrimination was protected," and to "provide practical reinforcement for the right," 417 U. S., at 444. Comparing the particular facts of the custodial interrogation with the "historical circumstances underlying the privilege," ibid., the Court concluded, unequivocally, that the defendant's statement could not be termed "involuntary as that term has been defined in the decisions of this Court," id., at 445, and thus that there had been no constitutional violation, notwithstanding the clear violation of the "procedural rules later established in Miranda, " ibid. Lest there be any confusion on the point, the Court reiterated that the "police conduct at issue here did not abridge respondent's constitutional privilege against compulsory self-incrimination, but departed only from the prophylactic standards later laid down by this Court in Miranda to safeguard that privilege." Id., at 446. It is clear from our cases, of course, that if the statement in Tucker had been obtained in violation of the Fifth Amendment, the statement and its fruits would have been excluded. See Nix v. Williams, 467 U.S. 431, 442 (1984).

The next year, in Oregon v. Hass, 420 U.S. 714 (1975), the Court held that a defendant's statement taken in violation of Miranda that was nonetheless voluntary could be used at trial for impeachment purposes. This holding turned upon the recognition that violation of Miranda is not unconstitutional compulsion, since statements obtained in actual violation of the privilege against compelled self-incrimination, "as opposed to . . . taken in violation of Miranda, " quite simply "may not be put to any testimonial use whatever against [the defendant] in a criminal trial," including as impeachment evidence. New Jersey v. Portash, 440 U.S. 450, 459 (1979). See also Mincey v. Arizona, 437 U.S. 385, 397— 398 (1978) (holding that while statements obtained in violation of Miranda may be used for impeachment if otherwise trustworthy, the Constitution prohibits "any criminal trial use against a defendant of his involuntary statement").

Nearly a decade later, in New York v. Quarles, 467 U.S. 649 (1984), the Court relied upon the fact that "[t]he prophylactic Miranda warnings . . . are `not themselves rights protected by the Constitution,' " id., at 654 (quoting Tucker, supra, at 444), to create a "public safety" exception. In that case, police apprehended, after a chase in a grocery store, a rape suspect known to be carrying a gun. After handcuffing and searching him (and finding no gun)—but before reading him his Miranda warnings—the police demanded to know where the gun was. The defendant nodded in the direction of some empty cartons and responded that "the gun is over there." The Court held that both the unwarned statement—"the gun is over there"—and the recovered weapon were admissible in the prosecution's case in chief under a "public safety exception" to the "prophylactic rules enunciated in Miranda. " 467 U. S., at 653. It explicitly acknowledged that if the Miranda warnings were an imperative of the Fifth Amendment itself, such an exigency exception would be impossible, since the Fifth Amendment's bar on compelled self-incrimination is absolute, and its "`strictures, unlike the Fourth's are not removed by showing reasonableness,' " 467 U. S., at 653, n. 3. (For the latter reason, the Court found it necessary to note that respondent did not "claim that [his] statements were actually compelled by police conduct which overcame his will to resist," id., at 654.)

The next year, the Court again declined to apply the "fruit of the poisonous tree" doctrine to a Miranda violation, this time allowing the admission of a suspect's properly warned statement even though it had been preceded (and, arguably, induced) by an earlier inculpatory statement taken in violation of Miranda. Oregon v. Elstad, 470 U.S. 298 (1985). As in Tucker, the Court distinguished the case from those holding that a confession obtained as a result of an unconstitutional search is inadmissible, on the ground that the violation of Miranda does not involve an "actual infringement of the suspect's constitutional rights," 470 U. S., at 308. Miranda, the Court explained, "sweeps more broadly than the Fifth Amendment itself," and "Miranda `s preventive medicine provides a remedy even to the defendant who has suffered no identifiable constitutional harm." 470 U. S., at 306-307. "[E]rrors [that] are made by law enforcement officers in administering the prophylactic Miranda procedures . . . should not breed the same irremediable consequences as police infringement of the Fifth Amendment itself." Id., at 308-309.

In light of these cases, and our statements to the same effect in others, see, e. g., Davis v. United States, 512 U. S., at 457-458; Withrow v. Williams, 507 U.S. 680, 690-691 (1993); Eagan, 492 U. S., at 203, it is simply no longer possible for the Court to conclude, even if it wanted to, that a violation of Miranda' s rules is a violation of the Constitution. But as I explained at the outset, that is what is required before the Court may disregard a law of Congress governing the admissibility of evidence in federal court. The Court today insists that the decision in Miranda is a "constitutional" one, ante, at 432, 438; that it has "constitutional underpinnings," ante, at 440, n. 5; a "constitutional basis" and a "constitutional origin," ante, at 439, n. 3; that it was "constitutionally based," ante, at 440; and that it announced a "constitutional rule," ante, at 437, 439, 441, 444. It is fine to play these word games; but what makes a decision "constitutional" in the only sense relevant here—in the sense that renders it impervious to supersession by congressional legislation such as § 3501— is the determination that the Constitution requires the result that the decision announces and the statute ignores. By disregarding congressional action that concededly does not violate the Constitution, the Court flagrantly offends fundamental principles of separation of powers, and arrogates to itself prerogatives reserved to the representatives of the people.

The Court seeks to avoid this conclusion in two ways: First, by misdescribing these post-Miranda cases as mere dicta. The Court concedes only "that there is language in some of our opinions that supports the view" that Miranda `s protections are not "constitutionally required." Ante, at 438. It is not a matter of language; it is a matter of holdings. The proposition that failure to comply with Miranda `s rules does not establish a constitutional violation was central to the holdings of Tucker, Hass, Quarles, and Elstad.

The second way the Court seeks to avoid the impact of these cases is simply to disclaim responsibility for reasoned decisionmaking. It says:

Finally, the Court asserts that Miranda must be a "constitutional decision" announcing a "constitutional rule," and thus immune to congressional modification, because we have since its inception applied it to the States. If this argument is meant as an invocation of stare decisis, it fails because, though it is true that our cases applying Miranda against the States must be reconsidered if Miranda is not required by the Constitution, it is likewise true that our cases (discussed above) based on the principle that Miranda is not required by the Constitution will have to be reconsidered if it is. So the stare decisis argument is a wash. If, on the other hand, the argument is meant as an appeal to logic rather than stare decisis, it is a classic example of begging the question: Congress's attempt to set aside Miranda, since it represents an assertion that violation of Miranda is not a violation of the Constitution, also represents an assertion that the Court has no power to impose Miranda on the States. To answer this assertion—not by showing why violation of Miranda is a violation of the Constitution—but by asserting that Miranda does apply against the States, is to assume precisely the point at issue. In my view, our continued application of the Miranda code to the States despite our consistent statements that running afoul of its dictates does not necessarily—or even usually—result in an actual constitutional violation, represents not the source of Miranda `s salvation but rather evidence of its ultimate illegitimacy. See generally J. Grano, Confessions, Truth, and the Law 173-198 (1993); Grano, Prophylactic Rules in Criminal Procedure: A Question of Article III Legitimacy, 80 Nw. U. L. Rev. 100 (1985). As Justice Stevens has elsewhere explained: "This Court's power to require state courts to exclude probative self-incriminatory statements rests entirely on the premise that the use of such evidence violates the Federal Constitution. . . . If the Court does not accept that premise, it must regard the holding in the Miranda case itself, as well as all of the federal jurisprudence that has evolved from that decision, as nothing more than an illegitimate exercise of raw judicial power." Elstad, 470 U. S., at 370 (dissenting opinion). Quite so.

III

There was available to the Court a means of reconciling the established proposition that a violation of Miranda does not itself offend the Fifth Amendment with the Court's assertion of a right to ignore the present statute. That means of reconciliation was argued strenuously by both petitioner and the United States, who were evidently more concerned than the Court is with maintaining the coherence of our jurisprudence. It is not mentioned in the Court's opinion because, I assume, a majority of the Justices intent on reversing believes that incoherence is the lesser evil. They may be right.

Petitioner and the United States contend that there is nothing at all exceptional, much less unconstitutional, about the Court's adopting prophylactic rules to buttress constitutional rights, and enforcing them against Congress and the States. Indeed, the United States argues that "[p]rophylactic rules are now and have been for many years a feature of this Court's constitutional adjudication." Brief for United States 47. That statement is not wholly inaccurate, if by "many years" one means since the mid-1960's. However, in their zeal to validate what is in my view a lawless practice, the United States and petitioner greatly overstate the frequency with which we have engaged in it. For instance, petitioner cites several cases in which the Court quite simply exercised its traditional judicial power to define the scope of constitutional protections and, relatedly, the circumstances in which they are violated. See Loretto v. Teleprompter Manhattan CATV Corp., 458 U.S. 419, 436-437 (1982) (holding that a permanent physical occupation constitutes a per se taking); Maine v. Moulton, 474 U.S. 159, 176 (1985) (holding that the Sixth Amendment right to the assistance of counsel is actually "violated when the State obtains incriminating statements by knowingly circumventing the accused's right to have counsel present in a confrontation between the accused and a state agent").

Similarly unsupportive of the supposed practice is Bruton v. United States, 391 U.S. 123 (1968), where we concluded that the Confrontation Clause of the Sixth Amendment forbids the admission of a nontestifying codefendant's facially incriminating confession in a joint trial, even where the jury has been given a limiting instruction. That decision was based, not upon the theory that this was desirable protection "beyond" what the Confrontation Clause technically required; but rather upon the self-evident proposition that the inability to cross-examine an available witness whose damaging out-of-court testimony is introduced violates the Confrontation Clause, combined with the conclusion that in these circumstances a mere jury instruction can never be relied upon to prevent the testimony from being damaging, see Richardson v. Marsh, 481 U.S. 200, 207-208 (1987).

The United States also relies on our cases involving the question whether a State's procedure for appointed counsel's withdrawal of representation on appeal satisfies the State's constitutional obligation to "`affor[d] adequate and effective appellate review to indigent defendants.' " Smith v. Robbins, 528 U.S. 259, 276 (2000) (quoting Griffin v. Illinois, 351 U.S. 12, 20 (1956)). In Anders v. California, 386 U.S. 738 (1967), we concluded that California's procedure governing withdrawal fell short of the constitutional minimum, and we outlined a procedure that would meet that standard. But as we made clear earlier this Term in Smith, which upheld a procedure different from the one Anders suggested, the benchmark of constitutionality is the constitutional requirement of adequate representation, and not some excrescence upon that requirement decreed, for safety's sake, by this Court.

In a footnote, the United States directs our attention to certain overprotective First Amendment rules that we have adopted to ensure "breathing space" for expression. See Gertz v. Robert Welch, Inc., 418 U.S. 323, 340, 342 (1974) (recognizing that in New York Times Co. v. Sullivan, 376 U.S. 254 (1964), we "extended a measure of strategic protection to defamatory falsehood" of public officials); Freedman v. Maryland, 380 U.S. 51, 58 (1965) (setting forth "procedural safeguards designed to obviate the dangers of a censorship system" with respect to motion picture obscenity). In these cases, and others involving the First Amendment, the Court has acknowledged that in order to guarantee that protected speech is not "chilled" and thus forgone, it is in some instances necessary to incorporate in our substantive rules a "measure of strategic protection." But that is because the Court has viewed the importation of "chill" as itself a violation of the First Amendment—not because the Court thought it could go beyond what the First Amendment demanded in order to provide some prophylaxis.

Petitioner and the United States are right on target, however, in characterizing the Court's actions in a case decided within a few years of Miranda, North Carolina v. Pearce, 395 U.S. 711 (1969). There, the Court concluded that due process would be offended were a judge vindictively to resentence with added severity a defendant who had successfully appealed his original conviction. Rather than simply announce that vindictive sentencing violates the Due Process Clause, the Court went on to hold that "[i]n order to assure the absence of such a [vindictive] motivation, . . . the reasons for [imposing the increased sentence] must affirmatively appear" and must "be based upon objective information concerning identifiable conduct on the part of the defendant occurring after the time of the original sentencing proceeding." Id., at 726. The Court later explicitly acknowledged Pearce `s prophylactic character, see Michigan v. Payne, 412 U.S. 47, 53 (1973). It is true, therefore, that the case exhibits the same fundamental flaw as does Miranda when deprived (as it has been) of its original (implausible) pretension to announcement of what the Constitution itself required. That is, although the Due Process Clause may well prohibit punishment based on judicial vindictiveness, the Constitution by no means vests in the courts "any general power to prescribe particular devices `in order to assure the absence of such a motivation,' " 395 U. S., at 741 (Black, J., dissenting). Justice Black surely had the right idea when he derided the Court's requirement as "pure legislation if there ever was legislation," ibid., although in truth Pearce `s rule pales as a legislative achievement when compared to the detailed code promulgated in Miranda.1

The foregoing demonstrates that, petitioner's and the United States' suggestions to the contrary notwithstanding, what the Court did in Miranda (assuming, as later cases hold, that Miranda went beyond what the Constitution actually requires) is in fact extraordinary. That the Court has, on rare and recent occasion, repeated the mistake does not transform error into truth, but illustrates the potential for future mischief that the error entails. Where the Constitution has wished to lodge in one of the branches of the Federal Government some limited power to supplement its guarantees, it has said so. See Amdt. 14, § 5 ("The Congress shall have power to enforce, by appropriate legislation, the provisions of this article"). The power with which the Court would endow itself under a "prophylactic" justification for Miranda goes far beyond what it has permitted Congress to do under authority of that text. Whereas we have insisted that congressional action under § 5 of the Fourteenth Amendment must be "congruent" with, and "proportional" to, a constitutional violation, see City of Boerne v. Flores, 521 U.S. 507, 520 (1997), the Miranda nontextual power to embellish confers authority to prescribe preventive measures against not only constitutionally prohibited compelled confessions, but also (as discussed earlier) foolhardy ones.

I applaud, therefore, the refusal of the Justices in the majority to enunciate this boundless doctrine of judicial empowerment as a means of rendering today's decision rational. In nonetheless joining the Court's judgment, however, they overlook two truisms: that actions speak louder than silence, and that (in judge-made law at least) logic will out. Since there is in fact no other principle that can reconcile today's judgment with the post-Miranda cases that the Court refuses to abandon, what today's decision will stand for, whether the Justices can bring themselves to say it or not, is the power of the Supreme Court to write a prophylactic, extra constitutional Constitution, binding on Congress and the States.

IV

Thus, while I agree with the Court that § 3501 cannot be upheld without also concluding that Miranda represents an illegitimate exercise of our authority to review state-court judgments, I do not share the Court's hesitation in reaching that conclusion. For while the Court is also correct that the doctrine of stare decisis demands some "special justification" for a departure from longstanding precedent—even precedent of the constitutional variety—that criterion is more than met here. To repeat Justice Stevens' cogent observation, it is "[o]bviou[s]" that "the Court's power to reverse Miranda's conviction rested entirely on the determination that a violation of the Federal Constitution had occurred." Elstad, 470 U. S., at 367, n. 9 (dissenting opinion) (emphasis added). Despite the Court's Orwellian assertion to the contrary, it is undeniable that later cases (discussed above) have "undermined [Miranda `s] doctrinal underpinnings," ante, at 443, denying constitutional violation and thus stripping the holding of its only constitutionally legitimate support. Miranda `s critics and supporters alike have long made this point. See Office of Legal Policy, U. S. Dept. of Justice, Report to Attorney General on Law of Pre-Trial Interrogation 97 (Feb. 12, 1986) ("The current Court has repudiated the premises on which Miranda was based, but has drawn back from recognizing the full implications of its decisions"); id., at 78 ("Michigan v. Tucker accordingly repudiated the doctrinal basis of the Miranda decision"); Sonenshein, Miranda and the Burger Court: Trends and Countertrends, 13 Loyola U. Chi. L. J. 405, 407-408 (1982) ("Although the Burger Court has not overruled Miranda, the Court has consistently undermined the rationales, assumptions, and values which gave Miranda life"); id., at 425-426 ("Seemingly, the Court [in Michigan v. Tucker ] utterly destroyed both Miranda `s rationale and its holding"); Stone, The Miranda Doctrine in the Burger Court, 1977 S. Ct. Rev. 99, 118 ("Mr. Justice Rehnquist's conclusion that there is a violation of the Self-Incrimination Clause only if a confession is involuntary . . . is an outright rejection of the core premises of Miranda ").

The Court cites Patterson v. McLean Credit Union, 491 U.S. 164, 173 (1989), as accurately reflecting our standard for overruling, see ante, at 443—which I am pleased to accept, even though Patterson was speaking of overruling statutory cases and the standard for constitutional decisions is somewhat more lenient. What is set forth there reads as though it was written precisely with the current status of Miranda in mind:

Moreover, it is not clear why the Court thinks that the "totality-of-the-circumstances test . . . is more difficult than Miranda for law enforcement officers to conform to, and for courts to apply in a consistent manner." Ante, at 444. Indeed, I find myself persuaded by Justice O'Connor's rejection of this same argument in her opinion in Williams, 507 U. S., at 711-712 (O'Connor, J., joined by Rehnquist, C. J., concurring in part and dissenting in part):

Finally, I am not convinced by petitioner's argument that Miranda should be preserved because the decision occupies a special place in the "public's consciousness." Brief for Petitioner 44. As far as I am aware, the public is not under the illusion that we are infallible. I see little harm in admitting that we made a mistake in taking away from the people the ability to decide for themselves what protections (beyond those required by the Constitution) are reasonably affordable in the criminal investigatory process. And I see much to be gained by reaffirming for the people the wonderful reality that they govern themselves—which means that "[t]he powers not delegated to the United States by the Constitution" that the people adopted, "nor prohibited . . . to the States" by that Constitution, "are reserved to the States respectively, or to the people," U. S. Const., Amdt. 10.2

* * *

Today's judgment converts Miranda from a milestone of judicial overreaching into the very Cheops' Pyramid (or perhaps the Sphinx would be a better analogue) of judicial arrogance. In imposing its Court-made code upon the States, the original opinion at least asserted that it was demanded by the Constitution. Today's decision does not pretend that it is—and yet still asserts the right to impose it against the will of the people's representatives in Congress. Far from believing that stare decisis compels this result, I believe we cannot allow to remain on the books even a celebrated decision—especially a celebrated decision—that has come to stand for the proposition that the Supreme Court has power to impose extra constitutional constraints upon Congress and the States. This is not the system that was established by the Framers, or that would be established by any sane supporter of government by the people.

I dissent from today's decision, and, until § 3501 is repealed, will continue to apply it in all cases where there has been a sustainable finding that the defendant's confession was voluntary.