492 U.S. 195
SCOTUS
1989

Lead Opinion

*197Chief Justice Rehnquist

delivered the opinion of the Court.

Respondent confessed to stabbing a woman nine times after she refused to have sexual relations with him, and he was convicted of attempted murder. Before confessing, respondent was given warnings by the police, which included the advice that a lawyer would be appointed “if and when you go to court.” The United States Court of Appeals for the Seventh Circuit held that such advice did not comply with the requirements of Miranda v. Arizona, 384 U. S. 436 (1966). We disagree and reverse.

Late on May 16, 1982, respondent contacted a Chicago police officer he knew to report that he had seen the naked body of a dead woman lying on a Lake Michigan beach. Respondent denied any involvement in criminal activity. He then took several Chicago police officers to the beach, where the woman was crying for help. When she saw respondent, the woman exclaimed: “Why did you stab me? Why did you stab me?” Respondent told the officers that he had been with .the woman earlier that night, but that they had been attacked by several men who abducted the woman in a van.

The next morning, after realizing that the crime had been committed in Indiana, the Chicago police turned the investigation over to the Hammond, Indiana, Police Department. Respondent repeated to the Hammond police officers his story that he had been attacked on the lakefront, and that the woman had been abducted by several men. After he filled out a battery complaint at a local police station, respondent agreed to go to the Hammond police headquarters for further questioning.

At about 11 a. m., the Hammond police questioned respondent. Before doing so, the police read to respondent a waiver form, entitled “Voluntary Appearance; Advice of Rights,” and they asked him to sign it. The form provided:

*198“Before we ask you any questions, you must understand your rights. You have the right to remain silent. Anything you say can be used against you in court. You have a right to talk to a lawyer for advice before we ask you any questions, and to have him with you during questioning. You have this right to the advice and presence of a lawyer even if you cannot afford to hire one. We have no way of giving you a lawyer, but one will be appointed for you, if you wish, if and when you go to court. If you wish to answer questions now without a lawyer present, you have the right to stop answering questions at any time. You also have the right to stop answering at any time until you’ve talked to a lawyer.” 843 F. 2d 1554, 1555-1556 (CA7 1988) (emphasis added).1

Respondent signed the form and repeated his exculpatory explanation for his activities of the previous evening.

Respondent was then placed in the “lockup” at the Hammond police headquarters. Some 29 hours later, at about 4 p.m. on May 18, the police again interviewed respondent. Before this questioning, one of the officers read the following waiver form to respondent:

“1. Before making this statement, I was advised that I have the right to remain silent and that anything I *199might say may or will be used against me in a court of law.
“2. That I have the right to consult with an attorney of my own choice before saying anything, and that an attorney may be present while I am making any statement or throughout the course of any conversation with any police officer if I so choose.
“3. That I can stop and request an attorney at any time during the course of the taking of any statement or during the course of any such conversation.
“4. That in the course of any conversation I can refuse to answer any further questions and remain silent, thereby terminating the conversation.
“5. That if I do not hire an attorney, one will be provided for me.” Id., at 1556.

Respondent read the form back to the officers and signed it. He proceeded to confess to stabbing the woman. The next morning, respondent led the officers to the Lake Michigan beach where they recovered the knife he had used in the stabbing and several items of clothing.

At trial, over respondent’s objection, the state court admitted his confession, his first statement denying any involvement in the crime, the knife, and the clothing. The jury found respondent guilty of attempted murder, but acquitted him of rape. He was sentenced to 35 years’ imprisonment. The conviction was upheld on appeal. Eagan v. State, 480 N. E. 2d 946 (Ind. 1985).

Respondent sought a writ of habeas corpus in the United States District Court for the Northern District of Indiana, claiming, inter alia, that his confession was inadmissible because the first waiver form did not comply with Miranda. The District Court denied the petition, holding that the record “clearly manifests adherence to Miranda . . . espe-*200dally as to the so-called second statement.” App. to Pet. for Cert. A52.

A divided United States Court of Appeals for the Seventh Circuit reversed. 843 F. 2d 1554 (1988). The majority held that the advice that counsel would be appointed “if and when you go to court,” which was included in the first warnings given to respondent, was “constitutionally defective because it denies an accused indigent a clear and unequivocal warning of the right to appointed counsel before any interrogation,” and “link[s] an indigent’s right to counsel before interrogation with a future event.” Id., at 1557. The majority relied on the Seventh Circuit’s decision in United States ex rel. Williams v. Twomey, 467 F. 2d 1248, 1250 (1972), which had condemned, as “misleading and confusing,” the inclusion of “if and when you go to court” language in Miranda warnings. Turning to the admissibility of respondent’s confession, the majority thought that “as a result of the first warning, [respondent] arguably believed that he could not secure a lawyer during interrogation” and that the second warning “did not explicitly correct this misinformation.” 843 F. 2d, at 1558. It therefore remanded the case for a determination whether respondent had knowingly and intelligently waived his right to an attorney during the second interview. The dissenting judge rejected the majority’s “formalistic, technical and unrealistic application of Miranda” and argued that the first warnings passed constitutional muster. Id., at 1562., In any case, he thought that remand was not necessary because the record indicated that this case was covered by Oregon v. Elstad, 470 U. S. 298 (1985). 843 F. 2d, at 1570-1571.

The Court of Appeals denied rehearing en banc, with four judges dissenting from that order. App. to Pet. for Cert. A1-A2. We then granted certiorari, 488 U. S. 888 (1988), to resolve a conflict among the lower courts as to whether informing a suspect that an attorney would be appointed for him “if and when you go to court” renders Miranda warn*201ings inádequate.2 We agree with the majority of the lower courts that it does not.3

In Miranda v. Arizona, 384 U. S. 436 (1966), the Court established certain procedural safeguards that require police to advise criminal suspects of their rights under the Fifth and Fourteenth Amendments before commencing custodial interrogation. In now-familiar words, the Court said that the *202suspect must be told that “he 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. The Court in Miranda “presumed that interrogation in certain custodial circumstances is inherently coercive and . . . that statements made under those circumstances are inadmissible unless the suspect is specifically warned of his Miranda rights and freely decides to forgo those rights.” New York v. Quarles, 467 U. S. 649, 654 (1984) (footnote omitted).

We have never insisted that Miranda warnings be given in the exact form described in that decision.4 In Miranda itself, the Court said that “[t]he warnings required and the waiver necessary in accordance with our opinion today are, in the absence of a fully effective equivalent, prerequisites to the admissibility of any statement made by a defendant.” 384 U. S., at 476 (emphasis added). See also Rhode Island v. Innis, 446 U. S. 291, 297 (1980) (referring to “the now familiar Miranda warnings ... or their equivalent”). In California v. Prysock, 453 U. S. 355 (1981) (per curiam), we stated that “the ‘rigidity’ of Miranda [does not] exten[d] to the precise formulation of the warnings given a criminal defendant,” and *203that “no talismanic incantation [is] required to satisfy its strictures.” Id., at 359.

Miranda has not been limited to station house questioning, see Rhode Island v. Innis, supra (police car), and the officer in the field may not always have access to printed Miranda warnings, or he may inadvertently depart from routine practice, particularly if a suspect requests an elaboration of the warnings. The prophylactic Miranda warnings are “not themselves rights protected by the Constitution but [are] instead measures to in.sure that the right against compulsory self-incrimination [is] protected.” Michigan v. Tucker, 417 U. S. 433, 444 (1974). Reviewing courts therefore need not examine Miranda warnings as if construing a will or defining the terms of an easement. The inquiry is simply whether the warnings reasonably “conve[y] to [a suspect] his rights as required by Miranda.” Pry sock, supra, at 361.

We think the initial warnings given to respondent touched all of the bases required by Miranda. The police told respondent that he had the right to remain silent, that anything he said could be used against him in court, that he had the right to speak to an attorney before and during questioning, that he had “this right to the advice and presence of a lawyer even if [he could] not afford to hire one,” and that he had the “right to stop answering at any time until [he] talked to a lawyer.” 843 F. 2d, at 1555-1556. As noted, the police also added that they could not provide respondent with a lawyer, but that one would be appointed “if and when you go to court.” The Court of Appeals thought this “if and when you go to court” language suggested that “only those accused who can afford an attorney have the right to have one present before answering any questions,” and “implie[d] that if the accused does not ‘go to court/ i. e.[,] the government does not file charges, the accused is not entitled to [counsel] at all.” Id., at 1557.

In our view, the Court of Appeals misapprehended the effect of the inclusion of “if and when you go to court” language *204in Miranda warnings. First, this instruction accurately described the procedure for the appointment of counsel in Indiana. Under Indiana law, counsel is appointed at the defendant’s initial appearance in court, Ind. Code § 35-33-7-6 (1988), and formal charges must be filed at or before that hearing, §35-33-7-3(a).5 We think it must be relatively commonplace for a suspect, after receiving Miranda warnings, to ask when he will obtain counsel. The “if and when you go to court” advice simply anticipates that question.6 Second, Miranda does not require that attorneys be producible on call, but only that the suspect be informed, as here, that he has the right to an attorney before and during questioning, and that an attorney would be appointed for him if he could not afford one.7 The Court in Miranda emphasized that it was not suggesting that “each police station must have a ‘station house lawyer’ present at all times to advise prisoners.” 384 U. S., at 474. If the police cannot provide appointed counsel, Miranda requires only that the police not question a suspect unless he waives his right to counsel. Ibid. Here, respondent did just that.

Respondent relies, Brief for Respondent 24-29, on language in California v. Prysock, where we suggested that Miranda warnings would not be sufficient “if the reference to the right to appointed counsel was linked [to a] future point in time after the police interrogation.” 453 U. S., at 360 (emphasis added). The Court of Appeals also referred to Prysock in finding deficient the initial warnings given to re*205spondent. 843 F. 2d, at 1557. But the vice referred to in Prysock was that such warnings would not apprise the accused of his right to have an attorney present if he chose to answer questions. The warnings in this case did not suffer from that defect. Of the eight sentences in the initial warnings, one described respondent’s right to counsel “before [the police] ask[ed] [him] questions,” while another stated his right to “stop answering at any time until [he] talk[ed] to a lawyer.” Id., at 1555-1556. We hold that the initial warnings given to respondent, in their totality, satisfied Miranda, and therefore that his first statement denying his involvement in the crime, as well as the knife and the clothing, was properly admitted into evidence.

The Court of Appeals thought it necessary to remand this case for consideration of whether respondent’s second statement was tainted by the first warnings. Id., at 1557-1558. In view of our disposition of this case, we need not reach that question.8 The judgment of the Court of Appeals is accordingly reversed, and the case is remanded for further proceedings consistent with our decision.

It is so ordered.

Notes

The remainder of the form signed by respondent provided:

“I, [Gary Eagan,] have come to the Detective Bureau of the Hammond, Indiana Police Department, of my own choice to talk with Officers ... In [sic] regard to an investigation they are conducting. I know that I am not under arrest and that I can leave this office if I wish to do so.

“Prior to any questioning, I was furnished with the above statement of my rights.... I have (read) (had read to me) this statement of my rights. I understand what my rights are. I am willing to answer questions and make a statement. I do not want a lawyer. I understand and know what I am doing. No promises or threats have been made to me and no pressure of any kind has been used against me.” 843 F. 2d, at 1560, n. 2.

The majority of federal and state courts to consider the issue have held that warnings that contained “if and when you go to court” language satisfied Miranda. See Wright v. North Carolina, 483 F. 2d 405, 406-407 (CA4 1973), cert. denied, 415 U. S. 936 (1974); Massimo v. United States, 463 F. 2d 1171, 1174 (CA2 1972), cert. denied, 409 U. S. 1117 (1973); United States v. Lacy, 446 F. 2d 511, 513 (CA5 1971); State v. Sterling, 377 So. 2d 58, 62-63 (La. 1979); Harrell v. State, 357 So. 2d 643, 645-646 (Miss. 1978); Rowbotham v. State, 542 P. 2d 610, 618-619 (Okla. Crim. App. 1975); Grennier v. State, 70 Wis. 2d 204, 213-215, 234 N. W. 2d 316, 321-322 (1975); Schade v. State, 512 P. 2d 907, 915-916 (Alaska 1973); State v. Mumbaugh, 107 Ariz. 589, 596-597, 491 P. 2d 443, 450-451 (1971); People v. Campbell, 26 Mich. App. 196, 201-202, 182 N. W. 2d 4, 6-7 (1970), cert. denied, 401 U. S. 945 (1971); People v. Swift, 32 App. Div. 2d 183, 186-187, 300 N. Y. S. 2d 639, 643-644 (1969), cert. denied, 396 U. S. 1018 (1970). Other courts, although not using the precise “if and when you go to court” language, have held Miranda was satisfied by a warning that an attorney could not be appointed for a suspect until he appeared in court. See United States v. Contreras, 667 F. 2d 976, 979 (CA11), cert. denied, 459 U. S. 849 (1982); Coyote v. United States, 380 F. 2d 305, 308 (CA10), cert. denied, 389 U. S. 992 (1967); State v. Maluia, 56 Haw. 428, 431-435, 539 P. 2d 1200, 1205-1207 (1975); Emler v. State, 259 Ind. 241, 243-244, 286 N. E. 2d 408, 410-411 (1972); Jones v. State, 69 Wis. 2d 337, 343-345, 230 N. W. 2d 677, 682-683 (1975).

On the other hand, a minority of federal and state courts, including the Seventh Circuit in this case, have held that “if and when you go to court” language did not satisfy Miranda. See United States ex rel. Williams v. Twomey, 467 F. 2d 1248, 1249-1250 (CA7 1972); Gilpin v. United States, 415 F. 2d 638, 641 (CA5 1969); State v. Dess, 184 Mont. 116, 120-122, 602 P. 2d 142, 144-145 (1979); Commonwealth v. Johnson, 484 Pa. 349, 352-357, 399 A. 2d 111, 112-114 (1979); Square v. State, 283 Ala. 548, 550, 219 So. 2d 377, 378-379 (1969).

Petitioner does not argue, and we therefore need not decide, whether Stone v. Powell, 428 U. S. 465 (1976), should be extended to bar relitigation on federal habeas of nonconstitutional claims under Miranda.

For example, the standard Miranda warnings used by the Federal Bureau of Investigation provide as follows:

“Before we ask you any questions, you must understand your rights.
“You have the right to remain silent.
“Anything you say can be used against you in court.
“You have the right to talk to a lawyer for advice before we ask you any questions and to have a lawyer with you during questioning.
“If you cannot afford a lawyer, one will be appointed for you before any questioning if you wish.
“If you decide to answer questions now without a lawyer present, you will still have the right to stop answering at any time. You also have the right to stop answering at any time until you talk to a lawyer.” Brief for United States as Amicus Curiae 1-2, n. 1.

In federal court, the defendant’s initial hearing, at which counsel is appointed, may occur before the filing of the indictment or information. Fed. Rules Crim. Proc. 5(a), (e).

At oral argument, the United States said that the federal law enforcement officials do not use this language in order to avoid “unnecessary litigation.” Tr. of Oral Arg. 16.

In Miranda, the Court stated that the FBI’s then-current practice of informing suspects “of a right to free counsel if they are unable to pay, and the availability of such counsel from the Judge,” 384 U. S., at 486, was “consistent with the procedure which we delineate today,” id., at 484.

Respondent argues that the second set of Miranda warnings he received were deficient. Brief for Respondent 38-40. These specific warnings have been upheld by the Seventh Circuit, Richardson v. Duckworth, 834 F. 2d 1366 (CA7 1987), and the Indiana Supreme Court, Robinson v. State, 272 Ind. 312, 397 N. E. 2d 956 (1979), and we think they plainly comply with Miranda.



Concurrence Opinion

Justice O’Connor,

with whom Justice Scalia joins, concurring.

I concur in The Chief Justice’s opinion for the Court. I write separately to address an alternative ground for decision in this case which was raised, but not relied upon, by the District Court. In my view, the rationale of our decision in Stone v. Powell, 428 U. S. 465 (1976), dictates that the suppression remedy be unavailable to respondent on federal habeas.

*206I — I

Over seven years ago respondent stabbed a woman nine times after she refused to have sexual relations with him. Claiming that he had innocently discovered the body, respondent led Chicago police to the woman, who, upon seeing respondent, immediately identified him as her assailant. Respondent was twice informed of his rights and questioned by detectives. The first time he gave an exculpatory statement indicating that he had been attacked by the same persons who had assaulted the victim. In the second interview, respondent confessed to the stabbing. He then led police to the knife he had used and to several items of his clothing which were found near the scene of the assault. Respondent sought suppression of both his statements and the knife and clothing on the ground that the warnings he was given were inadequate under Miranda v. Arizona, 384 U. S. 436 (1966). After an evidentiary hearing, the trial court denied the motion to suppress. The evidence was admitted at trial, and respondent was convicted of attempted murder and sentenced to 35 years’ imprisonment. On appeal, the Indiana Supreme Court rejected respondent’s claim that the warnings given him during his first encounter with the police were insufficient under Miranda. Eagan v. State, 480 N. E. 2d 946, 949-950 (1985). The Indiana Supreme Court also noted that there was no evidence that respondent’s two statements were the product of police coercion or overbearing. Id., at 950.

In 1986, respondent filed this petition for federal-habeas corpus under 28 U. S. C. §2254. He raised the same Mi-■ randa claim which had been fully litigated in, and rejected by, the state courts. The District Court noted the possibility that respondent’s claim might not be cognizable on federal habeas under our decision in Stone v. Powell, but indicated that “[f ]or present purposes that issue remains to be solved by the Supreme Court or this Circuit.” App. to Pet. for Cert. A-50. The District Court found no evidence of “coer*207cive conduct” on the part of the police in this case, and denied the petition. Id., at A-52 — A-53. A divided panel of the Court of Appeals for the Seventh Circuit reversed, finding that a technical violation of the Miranda rule had occurred, and remanding the case to the District Court for a further evidentiary hearing to determine whether respondent’s second statement was “tainted” by the allegedly inadequate warnings given in the first encounter. 843 F. 2d 1554, 1557 (1988). This Court now reverses. Eighteen state and federal judges have now given plenary consideration to respondent’s Miranda claims. None of these judges has intimated any doubt as to respondent’s guilt or the voluntariness and probative value of his confession. After seven years of litigation, the initial determination of the Miranda issue by the state trial judge and the Indiana Supreme Court has been found to be the correct one. In my view, the federal courts’ exercise of habeas jurisdiction in this case has served no one: no violation of the Fifth Amendment itself has ever been alleged; there is no doubt that respondent is guilty of the crime of which he was convicted and deserving of punishment; respondent had a full and fair opportunity to litigate his claim in state court; and the marginal possibility that police adherence to Miranda will be enhanced by suppression of highly probative evidence some seven years after the police conduct at issue in this case is far outweighed by the harm to society’s interest in punishing and incapacitating those who violate its criminal laws.

II

In Stone v. Powell this Court held that claims that probative evidence should have been excluded at trial because of police conduct alleged to have violated the Fourth Amendment would not be entertained in a federal habeas proceeding where a full and fair opportunity to litigate the claim had been made available in the state courts. The Stone Court noted that the exclusionary rule “ ‘is a judicially created remedy designed to safeguard Fourth Amendment rights gener*208ally through its deterrent effect.’” 428 U. S., at 486, quoting United States v. Calandra, 414 U. S. 338, 348 (1974). The costs of such a rule are high: highly probative and often conclusive evidence of a criminal defendant’s guilt is withheld from the trier of fact in the hope of “encouraging] those who formulate law enforcement policies, and the officers who implement them, to incorporate Fourth Amendment ideals into their value system.” Stone, supra, at 492. The exclusionary rule is a structural device designed to promote sensitivity to constitutional values through its deterrent effect. As such, the rule’s utility must, as this Court has long recognized, be weighed against other important values in its application. Where the rule’s deterrent effect is likely to be marginal, or where its application offends other values central to our system of constitutional governance or the judicial process, we have declined to extend the rule to that context. See, e. g., United States v. Leon, 468 U. S. 897, 920-921 (1984) (refusing to apply exclusionary rule where police rely in good faith on a warrant issued by a neutral magistrate); Calandra, supra, at 349 (refusing to extend the rule to grand jury proceedings because its application “would seriously impede the grand jury”); Walder v. United States, 347 U. S. 62, 65 (1954) (exclusionary rule does not create “a shield against contradiction of [the defendant’s] untruths” and evidence seized in violation of the Fourth Amendment may be used for impeachment purposes).

In Stone, we found that application of the exclusionary rule to Fourth Amendment violations on federal habeas was likely to have only marginal effectiveness in deterring police misconduct, while offending important principles of federalism and finality in the criminal law which have long informed the federal courts’ exercise of habeas jurisdiction. In my view, this same weighing process leads ineluctably to the conclusion that the suppression remedy should not be available on federal habeas where the state courts have accorded a petitioner a full and fair opportunity to litigate a claim that *209Miranda warnings were not given or were somehow deficient. Indeed, the scales appear to me to tip further toward finality and repose in this context than in Stone itself.

The Fifth Amendment guarantees that “[n]o person . . . shall be compelled in any criminal case to be a witness against himself.” The Amendment has its roots in the Framers’ belief that a system of justice in which the focus is on the extraction of proof of guilt from the criminal defendant himself is often an adjunct to tyranny and may lead to the conviction of innocent persons. Thus, a violation of the constitutional guarantee occurs when one is “compelled” by governmental coercion to bear witness against oneself in the criminal process. See Colorado v. Connelly, 479 U. S. 157, 163-164, and n. 1 (1986); Malloy v. Hogan, 378 U. S. 1, 6-8 (1964). The suppression remedy is quite possibly contained within the guarantee of the Fifth Amendment itself.

The Miranda rule is not, nor did it ever claim to be, a dictate of the Fifth Amendment itself. The Miranda Court implicitly acknowledged as much when it indicated that procedures other than the warnings dictated by the Court’s opinion might satisfy constitutional concerns, see Miranda, 384 U. S., at 444, and what was implicit in the Miranda opinion itself has been made explicit in our subsequent cases. See, e. g., Oregon v. Elstad, 470 U. S. 298, 306-310 (1985) (noting that the Miranda rule “sweeps more broadly than the Fifth Amendment itself” and “may be triggered even in the absence of a Fifth Amendment violation”); accord, New York v. Quarles, 467 U. S. 649 (1984); Michigan v. Tucker, 417 U. S. 433, 442-446 (1974). Like all prophylactic rules, the Miranda rule “overprotects” the value at stake. In the name of efficient judicial administration of the Fifth Amendment guarantee and the need to create institutional respect for Fifth Amendment values, it sacrifices society’s interest in uncovering evidence of crime and punishing those who violate its laws. While this balance of interests may be perfectly justified in the context of direct review of criminal convic*210tions, in my view the balance shifts when applied to a presumptively final criminal judgment which is collaterally attacked in a federal habeas corpus proceeding. As Justice Kennedy has recently noted:

“[Fjederal habeas review itself entails significant costs. It disturbs the State’s significant interest in repose for concluded litigation, denies society the right to punish some admitted offenders, and intrudes on state sovereignty to a degree matched by few exercises of federal judicial authority.” Harris v. Reed, 489 U. S. 255, 282 (1989) (dissenting opinion).

Indeed, within weeks after our decision in Miranda, we declined to apply that decision retroactively to state prisoners on federal habeas, noting that the Miranda rule was unrelated to the truth seeking function of the criminal trial, and that its application on federal habeas “would require the retrial or release of numerous prisoners found guilty by trustworthy evidence.” Johnson v. New Jersey, 384 U. S. 719, 730-731 (1966). As in the Fourth Amendment context addressed in Stone, we have consistently declined to extend the Miranda rule and the suppression remedy attached to it to situations where its deterrent effect is minimal and is outweighed by other compelling interests. See, e. g., Oregon v. Hass, 420 U. S. 714, 722-723 (1975) (statements taken in violation of Miranda may be used to impeach the defendant’s testimony at trial); Tucker, 417 U. S., at 448-449 (refusing to apply suppression remedy to third party testimony alleged to be the fruits of a Miranda violation); id., at 461 (White, J., concurring in judgment) (“The arguable benefits from excluding such testimony by way of possibly deterring police conduct that might compel admissions are, in my view, far outweighed by the advantages of having relevant and probative testimony, not obtained by actual coercion, available at criminal trials to aid in the pursuit of truth”).

*211In my view, these principles compel the conclusion that Miranda claims seeking suppression of probative evidence are not cognizable on federal habeas. Title 28 U. S. C. § 2243 requires a federal habeas court to “dispose of the matter as law and justice require,” and we have long recognized that “in some circumstances considerations of comity and concerns for the orderly administration of criminal justice require a federal court to forgo the exercise of its habeas corpus power.” Francis v. Henderson, 425 U. S. 536, 539 (1976). Relitigation of Miranda claims offers little or no additional structural incentive to the police to abide by the dictates of that decision. The awarding of habeas relief years after conviction will often strike like lightning, and it is absurd to think that this added possibility of exclusion years after the police conduct at issue will have any appreciable effect on police training or behavior. As Judge Friendly wrote: “The mere failure to administer Miranda warnings . . . creates little risk of unreliability, and the deterrent value of permitting collateral attack goes beyond the point of diminishing returns.” Friendly, Is Innocence Irrelevant?, Collateral Attack on Criminal Judgments, 38 U. Chi. L. Rev. 142, 163 (1970). On the other hand, the costs of suppression in the federal habeas setting are significantly magnified. As in this case, lower federal courts often sit in “review” of the judgments of the highest courts of a state judicial system. This situation has always been a flashpoint of tension in the delicate relationship of the federal and state courts, and this exercise of federal power should not be undertaken lightly where no significant federal values are at stake. Perhaps most troubling is the cost to society in the efficient enforcement of its criminal laws. Excluding probative evidence years after trial, when a new trial may be a practical impossibility, will often result in the release of an admittedly guilty individual who may pose a continuing threat to society. While federal courts must and do vindicate constitutional values outside the truth seeking function of a criminal trial, where those values are *212unlikely to be served by the suppression remedy, the result is positively perverse. Exclusion in such a situation teaches not respect for the law, but casts the criminal system as a game and sends the message that society is so unmoved by the violation of its own laws that it is willing to frustrate their enforcement for the smallest of returns. If Stone v. Powell bars relitigation of allegations of constitutional violations on federal habeas, it seems to me clear that its rationale is directly applicable to relitigation of nonconstitutional claims under Miranda.

Justice Marshall’s dissenting opinion accuses me of exhibiting “a profound distaste for Miranda,” post, at 224, in suggesting that the rationale of Stone v. Powell should be applied to Miranda claims on federal habeas review. It is not a sign of disrespect for a particular substantive rule to refuse to apply it in a situation where it does not serve the purposes for which it was designed. Our jurisprudence has long recognized a distinction between direct and collateral review, and I am not the first Justice of this Court to suggest that prophylactic rules should be treated differently in collateral proceedings than on direct review. See, e. g., Greer v. Miller, 483 U. S. 756, 767-769 (1987) (Stevens, J., concurring) (distinguishing between direct review and collateral proceedings for purposes of application of rule of Doyle v. Ohio, 426 U. S. 610 (1976), which forbids prosecutorial comment on postarrest silence); Brewer v. Williams, 430 U. S. 387, 420-429 (1977) (Burger, C. J., dissenting) (suggesting applicability of Stone v. Powell to Miranda claims on federal habeas); see also Rose v. Lundy, 455 U. S. 509, 543-544, and n. 8 (1982) (Stevens, J., dissenting); Vasquez v. Hillery, 474 U. S. 254, 272-273 (1986) (Powell, J., dissenting). Indeed, in United States v. Timmreck, 441 U. S. 780 (1979), a unanimous Court concluded that a purely formal violation of Federal Rule of Criminal Procedure 11 did not justify the granting of relief in collateral proceedings despite the fact that at the time of our decision in Timmreck such a violation *213was often considered grounds for automatic reversal on direct review. See McCarthy v. United States, 394 U. S. 459 (1969). The distinction did not lie in any “profound distaste” for the dictates of Rule 11, but rather upon considerations of finality which have special force in the context of a collateral proceeding challenging a final criminal judgment. Timmreck, supra, at 784.

The dissent’s charges of “judicial activism” and its assertion that “Congress has determined” that collateral review of claims like those at issue in this case outweighs any interests in bringing a final resolution to the criminal process, see post, at 222, 228, ring quite hollow indeed in the context of the federal habeas statute. The scope of federal habeas corpus jurisdiction has undergone a substantial judicial expansion, and a return to what “Congress intended” would reduce the scope of habeas jurisdiction far beyond the extension of Stone v. Powell to Miranda claims. See Kuhlmann v. Wilson, 477 U. S. 436, 445-446 (1986) (plurality opinion) (“Until the early years of this century, the substantive scope of the federal ha-beas corpus statutes was defined by reference to the scope of the writ at common law.... During this century, the Court gradually expanded the grounds on which habeas corpus relief was available”); see also Rose, supra, at 546-548 (Stevens, J., dissenting); Fay v. Noia, 372 U. S. 391, 445 (1963) (Clark, J., dissenting); id., at 448 (Harlan, J., dissenting). As noted above, the Court has long recognized that “habeas corpus has been traditionally regarded as governed by equitable principles,” id., at 438 (citation omitted), and thus has long defined the scope of the writ by reference to a balancing of state and federal interests which the dissent today condemns as “activism.”

While the State did not raise the applicability of Stone v. Powell to respondent’s Miranda claim below or in its petition for certiorari, there is language in Stone which suggests that the bar it raises to relitigation of certain claims on federal habeas is jurisdictional or quasi-jurisdictional in nature. *214Stone, 428 U. S., at 482, and n. 17. Other parts of the opinion appear to rest on the equitable nature of the writ of ha-beas corpus and the equity court’s power to withhold certain forms of relief. Id., at 494-495, n. 37. Since I do not read the Court’s opinion as foreclosing the analysis outlined above, I join the Court’s opinion and judgment.





Dissenting Opinion

Justice Marshall,

with whom Justice Brennan joins, and with whom Justice Blackmun and Justice Stevens join as to Part I, dissenting.

The majority holds today that a police warning advising a suspect that he is entitled to an appointed lawyer only “if and when he goes to court” satisfies the requirements of Miranda v. Arizona, 384 U. S. 436 (1966). The majority reaches this result by seriously mischaracterizing that decision. Under Miranda, a police warning must “clearly infor[mJ” a suspect taken into custody “that if he cannot afford an attorney one will be appointed for him prior to any questioning if he so desires.” Id., at 471, 479 (emphasis added). A warning qualified by an “if and when you go to court” caveat does nothing of the kind; instead, it leads the suspect to believe that a lawyer will not be provided until some indeterminate time in the future after questioning. I refuse to acquiesce in the continuing debasement of this historic precedent, see, e. g., Oregon v. Elstad, 470 U. S. 298 (1985); New York v. Quarles, 467 U. S. 649 (1984), and therefore dissent. I also write to express my disagreement with Justice O’Connor’s uninvited suggestion that the rationale of Stone v. Powell, 428 U. S. 465 (1976), should be extended to bar federal habeas review of Miranda claims.

I

In Miranda, the Court held that law enforcement officers who take a suspect into custody must inform the suspect of, among other things, his right to have counsel appointed to represent him before and during interrogation:

*215“In order fully to apprise a person interrogated of the extent of his rights , it is necessary to warn him not only that he has the right to consult with an attorney, but also that if he is indigent a lawyer will be appointed to represent him. Without this additional warning, the admonition of the right to consult with counsel would often be understood as meaning only that he can consult with a lawyer if he has one or has the funds to obtain one. The warning of a right to counsel would be hollow if not couched in terms that would convey to the indigent — the person most often subjected to interrogation — the knowledge that he too has a right to have counsel present. As with the warning of the right to remain silent and of the general right to counsel, only by effective and express explanation to the indigent of this right can there be assurance that he was truly in a position to exercise it.” 384 U. S., at 473 (footnotes omitted).

Miranda mandated no specific verbal formulation that police must use, but the Court, speaking through Chief Justice Warren, emphasized repeatedly that the offer of appointed counsel must be “effective and express.” Ibid.; see also id., at 476 (only a “fully effective equivalent” of the warnings described will pass muster); id., at 444 (requiring “other fully effective means”); id., at 467 (requiring alternative that is “at least as effective”); id., at 490 (stating that “Congress and the States are free to develop their own safeguards for the privilege [against self-incrimination], so long as they are fully, as effective as those described above”). A clear and unequivocal offer to provide appointed counsel prior to questioning is, in short, an “absolute prerequisite to interrogation.” Id., at 471.

In concluding that the first warning given to respondent Eagan, quoted ante, at 198, satisfies the dictates of Miranda, the majority makes a mockery of that decision. Eagan was initially advised that he had the right to the presence of counsel before and during questioning. But in the very next *216breath, the police informed Eagan that, if he could not afford a lawyer, one would be appointed to represent him only “if and when” he went to court. As the Court of Appeals found, Eagan could easily have concluded from the “if and when” caveat that only “those accused who can afford an attorney have the right to have one present before answering any questions; those who are not so fortunate must wait.” 843 F. 2d 1554, 1557 (CA7 1988); see also United States ex rel. Williams v. Twomey, 467 F. 2d 1248, 1250 (CA7 1972). Eagan was, after all, never told that questioning would be delayed until a lawyer was appointed “if and when” Eagan did, in fact, go to court. Thus, the “if and when” caveat may well have had the effect of negating the initial promise that counsel could be present. At best, a suspect like Eagan “would not know . . . whether or not he had a right to the services of a lawyer.” Emler v. State, 286 N. E. 2d 408, 412 (Ind. 1972) (DeBruler, J., dissenting).1

In lawyerlike fashion, The Chief Justice parses the initial warnings given Eagan and finds that the most plausible interpretation is that Eagan would not be questioned until a lawyer was appointed when he later appeared in court. What goes wholly overlooked in The Chief Justice’s analysis is that the recipients of police warnings are often frightened suspects unlettered in the law, not lawyers or judges or others schooled in interpreting legal or semantic nuance. Such suspects can hardly be expected to interpret, in as facile a manner as The Chief Justice, “the pretzel-like warnings here — intertwining, contradictory, and ambiguous as they *217are.” Commonwealth v. Johnson, 484 Pa. 349, 356, 399 A. 2d 111, 115 (1979) (citation omitted) (finding inadequate a similar “if and when” caveat). The majority thus refuses to recognize that “[t]he warning of a right to counsel would be hollow if not couched in terms that would convey to the indigent — the person most often subjected to interrogation — the knowledge that he too has the right to have counsel present.” Miranda, swpra, at 473 (footnote omitted).

Even if the typical suspect could draw the inference the majority does — that questioning will not commence until a lawyer is provided at a later court appearance — a warning qualified by an “if and when” caveat still fails to give a suspect any indication of when he will be taken to court. Upon hearing the warnings given in this case, a suspect would likely conclude that no lawyer would be provided until trial. In common parlance, “going to court” is synonymous with “going to trial.” Furthermore, the negative implication of the caveat is that, if the suspect is never taken to court, he “is not entitled to an attorney at all.” 843 F. 2d, at 1557. An unwitting suspect harboring uncertainty on this score is precisely the sort of person who may feel compelled to talk “voluntarily” to the police, without the presence of counsel, in an effort to extricate himself from his predicament:

“[The suspect] is effectively told that he can talk now or remain in custody — in an alien, friendless, harsh world— for an indeterminate length of time. To the average accused, still hoping at this stage to be home on time for dinner or to make it to work on time, the implication that his choice is to answer questions right away or remain in custody until that nebulous time ‘if and when’ he goes to court is a coerced choice of the most obvious kind.” Dickerson v. State, 276 N. E. 2d 845, 852 (Ind. 1972) (DeBruler, J., concurring in result) (finding inadequate a warning identical to the one in this case).

See also United States ex rel. Williams, supra, at 1250; Schade v. State, 512 P. 2d 907, 920 (Alaska 1973) (Boochever, *218J., concurring). That the warning given to Eagan “accurately described the procedure for the appointment of counsel in Indiana,” ante, at 204, does nothing to mitigate the possibility that he would feel coerced into talking to the police. Miranda, it is true, does not require the police to have a “station house lawyer” ready at all times to counsel suspects taken into custody. 384 U. S., at 474. But if a suspect does not understand that a lawyer will be made available within a reasonable period of time after he has been taken into custody and advised of his rights, the suspect may decide to talk to the police for that reason alone. The threat of an indefinite deferral of interrogation, in a system like Indiana’s, thus constitutes an effective means by which the police can pressure a suspect to speak without the presence of counsel. Sanctioning such police practices simply because the warnings given do not misrepresent state law does nothing more than let the state-law tail wag the federal constitutional dog.2

The majority’s misreading of Miranda — stating that police warnings need only “touc[h] all of the bases required by Miranda,” ante, at 203, that Miranda warnings need only be “reasonably ‘conve[yed]’ ” to a suspect, ibid, (citation omitted), and that Miranda warnings are to be measured not point by point but “in their totality,” ante, at 205 — is exacerbated by its interpretation of California v. Prysock, 453 U. S. 355 (1981) (per curiam), a decision that squarely supports Eagan’s claim in this case. The juvenile suspect in Prysock was initially told that he had the right to have a lawyer present before and during questioning. He then was told that he had the right to have his parents present as well. At this point the suspect was informed that a lawyer would be appointed to repre*219sent him at no cost if he could not afford one. The California Court of Appeal ruled these warnings insufficient because the suspect was not expressly told of his right to an appointed attorney before and during questioning. This Court reversed, finding that “nothing in the warnings given respondent suggested any limitation on the right to the presence of appointed counsel.” Id., at 360-361.

In reaching this result, the Prysock Court pointedly distinguished a series of lower court decisions that had found inadequate warnings in which “the reference to the right to appointed counsel was linked with some future point in time.” Id., at 360. In United States v. Garcia, 431 F. 2d 134 (CA9 1970) (per curiam), for example, the suspect had been informed on one occasion that she had the right to appointed counsel “‘when she answered any questions/” and on another occasion that she could “ ‘have an attorney appointed to represent [her] when [she] first appealed] before the U. S. Commissioner or the Court.’” Similarly, in People v. Bolinski, 260 Cal. App. 2d 705, 718, 67 Cal. Rptr. 347, 355 (1968), the suspect was advised that counsel would be appointed “‘if he was charged.’” These lower courts had correctly found these warnings defective, the Prysock Court explained, because “[i]n both instances the reference to appointed counsel was linked to a future point in time after police interrogation,”3 and therefore did not clearly advise the suspect of his right to appointed counsel before such in*220terrogation. 453 U. S., at 360. The initial, conditional warning given Eagan suffers from precisely the same fatal defect. It is highly disingenuous for the majority to ignore this fact, characterizing Prysock as involving only the question whether a particular warning “apprise[d] the accused of his right to have an attorney present if he chose to answer questions.” Ante, at 205.

It poses no great burden on law enforcement officers to eradicate the confusion stemming from the “if and when” caveat. Deleting the sentence containing the offending language is all that needs to be done. See United States v. Cassell, 452 F. 2d 533, 541, n. 8 (CA7 1971). Purged of this language, the warning tells the suspect in a straightforward fashion that he has the right to the presence of a lawyer before and during questioning, and that a lawyer will be appointed if he cannot afford one. The suspect is given no reason to believe that the appointment of an attorney may come after interrogation. To the extent one doubts that it is the “if and when” caveat that is the source of the confusion, compare the initial warning given Eagan, quoted ante, at 198, and the crystal-clear warning currently used by the FBI, quoted ante, at 202, n. 4. The majority’s claim that the two warnings are indistinguishable in the message conveyed to a suspect defies belief. I dissent.4

*221Not content with disemboweling Miranda directly, Justice O’Connor seeks to do so indirectly as well, urging that federal courts be barred from considering Miranda claims on habeas corpus review. In Stone v. Powell, 428 U. S. 465 (1976), the Court held that a state prisoner may not seek federal habeas corpus relief on the ground that evidence was obtained in violation of his Fourth Amendment rights if the state courts had provided a full and fair opportunity for litigation of that claim. I joined Justice Brennan’s dissenting opinion in that case, in which he warned that the majority’s rationale “portends substantial evisceration of federal habeas corpus jurisdiction.” Id., at 503. Justice Powell, writing for the Stone majority, dismissed as “misdirected” the “hyperbole of the dissenting opinion,” id., at 494, n. 37, insisting that his opinion was based on considerations unique to the exclusionary rule. Today, however, Justice O’Connor seeks to extend Stone beyond the Fourth Amendment even though this issue was not raised by petitioner Duckworth below or in his petition for certiorari. Her concurring opinion evinces such a palpable distaste for collateral review of state-court judgments that it can only be viewed as a harbinger of future assaults on federal habeas corpus.5

*222Stone was wrong when it was decided and it is wrong today. I have read and reread the federal habeas corpus statute, but I am unable to find any statement to the effect that certain federal claims are unworthy of collateral protection, or that certain federal claims are more worthy of collateral protection than others. Congress did not delineate “second class” claims when it created federal habeas jurisdiction. Stone, supra, at 515 (Brennan, J., dissenting). On the contrary, Congress deemed all federal claims worthy of collateral protection when it extended the writ to any person “in custody pursuant to the judgment of a State Court... in violation of the Constitution or laws or treaties of the United States.” 28 U. S. C. §2254(a). At a time when plain language is supposed to count for something, Justice O’Con-nor’s suggestion that the Court carve out an exception that has no rooting in the text of the habeas statute is difficult to justify.

Under Article III of the Constitution, Congress — not this Court — determines the scope of jurisdiction of the inferior federal courts. Congress is undoubtedly aware that federal habeas review of state criminal convictions might disserve interests of comity and finality and might make the enforcement of state criminal laws more difficult. Congress has determined, however, that the individual’s interest in vindicating his federal rights in a federal forum outweighs these concerns. Federal courts, not state courts, thus have the “last say.” Brown v. Allen, 344 U. S. 443, 508 (1953) (opinion of Frankfurter, J.). Regardless of whether we believe *223this congressional scheme accords too little respect to principles of federalism or other values, ours is not the choice to make. See Stone, supra, at 511 (Brennan, J., dissenting) (“[A]s between this Court on certiorari, and federal district courts on habeas, it is for Congress to decide what the most efficacious method is for . . . asserting the primacy of federal law”) (emphasis in original); see also Jackson v. Virginia, 443 U. S. 307, 323 (1979); cf. Mincey v. Arizona, 437 U. S. 385, 402-405 (1978) (Marshall, J., concurring).

That Justice O’Connor’s position is driven by general hostility toward collateral review of state court judgments is apparent. She writes:

“[L]ower federal courts often sit in ‘review’ of the judgments of the highest courts of a state judicial system. This situation has always been a flashpoint of tension in the delicate relationship of the federal and state courts, and this exercise of federal power should not be undertaken lightly where no significant federal values are at stake. Perhaps most troubling is the cost to society in the efficient enforcement of its criminal laws. Excluding probative evidence years after trial, when a new trial may be a practical impossibility, will often result in the release of an admittedly guilty individual who may pose a continuing threat to society.” Ante, at 211 (concurring opinion).

This logic sweeps within its broad compass claims far beyond those based on Miranda. Once the specter is raised that federal habeas review may lead to the release of guilty criminals, it is difficult to imagine any non-guilt-related claim that would be worthy of collateral protection. What Justice O’Connor ignores is that Congress believed that defendants have rights, often unrelated to guilt or innocence, that are worthy of collateral protection despite the apparent costs to society. Thus, in Rose v. Mitchell, 443 U. S. 545 (1979), we refused to extend Stone to preclude a federal ha-beas claim of racial discrimination in the selection of a state *224grand jury foreperson, even though the defendants’ culpability for the murders charged in that case was not disputed. Under Justice O’Connor’s view that federal habeas review should extend only to guilt-related claims, however, the claim raised in Rose, along with claims such as prosecutorial misconduct, double jeopardy, or the right to a speedy trial, could never be cognizable on federal habeas.

It is not only disapprobation for federal habeas review that pervades Justice O’Connor’s concurring opinion, but also a profound distaste for Miranda. How else to explain the remarkable statement that “no significant federal values are at stake” when Miranda claims are raised in federal habeas corpus proceedings? Ante, at 211 (concurring opinion) (emphasis added). But irrespective of one’s view of the merits of Miranda, the critical point is that Miranda is still good law. With few exceptions, prosecutors in state courts may not introduce statements taken from a criminal suspect in violation of his Miranda rights. If a state trial court permits the introduction of such statements, federal constitutional error has been committed. Unless the defendant’s conviction is reversed, he is indisputably being held “in violation of the Constitution... of the United States.” 28 U. S. C. § 2254(a). This is true whether the defendant challenges the introduction of the statements on direct appeal or on collateral review, for the federal violation does not “suddenly vanis[h] after the appellate process has been exhausted.” Stone, 428 U. S., at 511 (Brennan, J., dissenting); see also id., at 536-537 (White, J., dissenting).

Even assuming that Stone was correctly decided, and that the question is therefore whether the benefits of the suppression remedy for Miranda violations on federal habeas outweigh its costs, I would still reject Justice O’Connor’s conclusion that “the scales appear ... to tip further toward finality and repose in this context than in Stone itself. ” Ante, at 209 (concurring opinion). In Stone, Justice Powell did not rest his “cost” analysis solely on the fact that the exclu*225sionary rule operates, like the Miranda requirements, to prevent juries from considering highly probative evidence. Justice Powell’s analysis was far subtler than that, for he focused on evidence that was both probative and “typically reliable.” 428 U. S., at 490. The erroneous admission of this type of evidence, he explained, does not cast doubt upon the state trial court’s “truthfinding process.” Ibid.; see also id., at 497 (Burger, C. J., concurring). Like evidence that a ha-beas petitioner challenges under the exclusionary rule — “a pistol, a packet of heroin, counterfeit money, or the body of a murder victim,” ibid.— a self-incriminatory statement.that a habeas petitioner challenges under Miranda is ordinarily highly probative. But unlike physical evidence seized from a suspect in violation of his Fourth Amendment rights, a statement taken from a suspect in violation of his Miranda rights is presumptively unreliable. See New York v. Quarles, 467 U. S. 649, 664 (1984) (O’Connor, J., concurring in judgment in part and dissenting in part) (“When police ask custodial questions without administering the required warnings, Miranda quite clearly requires that the answers received be presumed compelled . . .”). Thus, when Miranda claims are raised on federal habeas, “the integrity of the factfinding process” of the state trial court is called into question. Brewer v. Williams, 430 U. S. 387, 414 (1977) (Powell, J., concurring). This is precisely the situation in which.collateral review is most appropriate.

Justice O’Connor’s extension of Stone overlooks another difference between claims based on the exclusionary rule and claims based on Miranda. According to the Stone majority, the primary justification for the exclusionary rule is the deterrence of police misconduct. 428 U. S., at 486; but see id., at 510, and n. 9 (Brennan, J., dissenting). By contrast, the rights secured by Miranda go to the heart of our accusatorial system — “a system in which the State must establish guilt by evidence independently and freely secured and may not by coercion prove its charge against an accused out of his own *226mouth.” Rogers v. Richmond, 365 U. S. 534, 541 (1961). Justice O’Connor recognizes as much, acknowledging that the privilege against self-incrimination reflects the longstanding belief that “the extraction of proof of guilt from the criminal defendant himself is often an adjunct to tyranny and may lead to the conviction of innocent persons.” Ante, at 209 (concurring opinion). Unlike the exclusionary rule, which purportedly exists solely for deterrence purposes, the Miranda requirements thus serve to protect “a criminal suspect’s exercise of [a] privilege which is one of the distinctive components of our criminal law.” White v. Finkbeiner, 687 F. 2d 885, 893 (CA7 1982) (declining to extend Stone to Miranda claims).

Justice O’Connor attempts to elide this distinction by advocating that only “nonconstitutional” Miranda claims be barred on federal habeas. Ante, at 212 (concurring opinion). By this she presumably means those claims that are based on so-called “voluntary statements.” Oregon v. Elstad, 470 U. S. 298, 307 (1985) (emphasis in original). I have never accepted the proposition that there is any such a thing as a “nonconstitutional” Miranda claim based on “voluntary” statements. The explicit premise of Miranda is that, unless a suspect taken into custody is properly advised of his rights, “no statement obtained from the [suspect] can truly be the product of his free choice” as a matter of federal constitutional law. 384 U. S., at 458; see also id., at 445. As Justice Douglas explained: “Miranda’s purpose was not promulgation of judicially preferred standards for police interrogation, a function we are quite powerless to perform; the decision enunciated ‘constitutional standards for protection of the privilege’ against self-incrimination.” Michigan v. Tucker, 417 U. S. 433, 465-466 (1974) (dissenting opinion), quoting Miranda, 384 U. S., at 491. Granted, Miranda “is an area of the law filled with technical rules, and the protections it affords defendants might at times be perceived as technicalities,” Jones v. Thomas, 491 U. S. 376, 387 (1989), *227but fundamental principles embodied in the Self-Incrimination Clause are at stake whenever a Miranda claim is raised. See Orozco v. Texas, 394 U. S. 324, 326 (1969) (“[T]he use of . . . admissions obtained in the absence of the required warnings [is] a flat violation of the Self-Incrimination Clause of the Fifth Amendment as construed in Miranda”).

Even if it were possible to identify a class of “nonconstitu-tional” Miranda claims, there will be little gained in attempting to extend Stone to these claims. It is simply not possible to know in advance which habeas petitioners raising Miranda claims will have their statements found “voluntary” and which will not. Federal habeas courts therefore will be obligated to inquire into the nature of each habeas petitioner’s Miranda claim before deciding whether Stone should apply. Moreover, many habeas petitioners will have coupled their Miranda claims with traditional involuntariness claims based on the Due Process Clause, thereby making such inquiries inevitable. See Cardwell v. Taylor, 461 U. S. 571, 573 (1983) (per curiam) (“[I]f the statements were involuntary, and therefore obtained in violation of the Fifth Amendment, . . . the federal courts [could] grant relief on collateral review”). Such claims require significant judicial attention because “[d]ifiiculties of proof and subtleties of interrogation technique [make] it impossible in most cases for the judiciary to decide with confidence whether the defendant had voluntarily confessed his guilt or whether his testimony had been unconstitutionally compelled.” Quarles, 467 U. S., at 683 (Marshall, J., dissenting). The purported “costs” of collateral review in the exclusionary rule context, such as preventing finality and overburdening the federal courts, see Stone, supra, at 491, n. 31, thus will still exist even if Stone is extended to “nonconstitutional” Miranda claims.

In any event, I vehemently oppose the suggestion that it is for the Court to decide, based on our own vague notions of comity, finality, and the intrinsic value of particular constitutional rights, which claims are worthy of collateral federal re*228view and which are not.6 Congress already engaged in that balancing process when it created habeas review and extended the federal courts’ jurisdiction to all claims based on a violation of federal law. The federal courts have been reviewing Miranda claims on federal habeas for 23 years, and Congress has never even remotely indicated that they have been remiss in doing so. To the extent Justice O’Connor is unhappy with Miranda, she should address that decision head on. But an end run through the habeas statute is judicial activism at its worst.

Numerous courts have found inadequate police warnings containing an “if and when” caveat or its equivalent. See ante, at 201, n. 2; see also, e. g., United States v. Cassell, 452 F. 2d 533 (CA7 1971); United States v. Garcia, 431 F. 2d 134 (CA9 1970); United States v. Oliver, 421 F. 2d 1034 (CA10 1970); Reed v. State, 255 Ark. 63, 498 S. W. 2d 877 (1973); Burns v. State, 486 S. W. 2d 310 (Tex. Crim. App. 1972); State v. Creach, 77 Wash. 2d 194, 461 P. 2d 329 (1969); State v. Robbins, 4 N. C. App. 463, 167 S. E. 2d 16 (1969); People v. Bolinski, 260 Cal. App. 2d 705, 67 Cal. Rptr. 347 (1968); Brooks v. State, 229 A. 2d 833 (Del. 1967).

Nothing in Miranda v. Arizona, 384 U. S. 436 (1966), nor any of our other cases for that matter, supports the notion that the police may indefinitely delay the point at which counsel is appointed. On the contrary, the Court indicated in Miranda that the police could detain a person without providing counsel for no more than “a reasonable period of time.” Id., at 474.

The Solicitor General, emphasizing the words “after police interrogation,” reasons that Prysock “does not condemn warnings that simply link the appointment of counsel to some future event.” Brief for United States as Amicus Cíiriae 18. This argument is spurious. Nothing in the warnings given in Garcia or Bolinski explicitly linked the appointment of counsel to a future event occurring after interrogation, yet the Prysock Court still cited those decisions with approval. Indeed, the basic problem with the warnings in those cases (and the warning in this ease) is that a suspect would etroneously believe that appointment of counsel would be delayed until after interrogation. See United States v. Contreras, 667 F. 2d 976, 979 (CA11), cert. denied, 459 U. S. 849 (1982).

“With no analysis whatsoever, the majority also holds that the second set of warnings read to Eagan and included in a waiver form that he signed prior to his second interrogation, quoted ante, at 198-199, “plainly comply with Miranda.” Ante, at 205, n. 8. This proposition is subject to dispute given the presence of the “of my own choice” language. See Sotelo v. State, 342 N. E. 2d 844, 851 (Ind. 1976) (DeBruler, J. concurring). But even assuming the second set of warnings complied with Miranda, it does not necessarily follow that Eagan’s subsequent waiver of rights was knowing and intelligent. Given “the misapprehension caused by the initial warning,” 843 F. 2d 1554, 1557 (CA7 1988), the issue is not whether the second warnings were adequate standing alone, but rather whether under the circumstances the mistaken impression Eagan was initially given was corrected. While various factors might inform this inquiry, such as the passage of *221time, the principal question must be whether the new warnings were sufficiently clear to correct the effect of the earlier, defective warning. As there is little in the record on “the factual circumstances surrounding these events because the state courts did not directly examine this issue,” id., at 1558; see also Brief for Respondent 34-38, I agree with the Court of Appeals that “remand for a determination of whether [Eagan] knowingly and intelligently waived his right to the presence of an attorney during the second interrogation” is the appropriate course. 843 F. 2d, at 1558.

Justice O’Connor attempts to justify raising this issue by claiming that Stone has a jurisdictional component. See ante, at 212 (concurring opinion). That is not so. Whatever faint allusions to jurisdiction Justice Powell may have made on page 482 of his Stone opinion, he made crystal clear later in the opinion that “[o]ur decision does not mean that the federal court lacks jurisdiction over ... a [Fourth Amendment] claim.” 428 U. S., at *222495, n. 37. Nor could a federal court lack jurisdiction after Stone, for it would then be powerless to consider even those Fourth Amendment claims that had not been fully and fairly litigated in the state courts. Furthermore, if Stone did in fact have a jurisdictional component, it is hard to understand why Justice Powell, in refusing in a subsequent case to consider whether Stone should be applied to Fifth and Sixth Amendment claims, explained that the “question has not been presented in the briefs or arguments.” Brewer v. Williams, 430 U. S. 387, 414 (1977) (concurring opinion).

To paraphrase Justice Brennan:

“[A]ll of the ‘costs’ of applying [Miranda] on habeas should already have been incutred at the trial or on direct review if the state court had not misapplied federal constitutional principles. As such, these ‘costs’ were evaluated and deemed to be outweighed when [the Miranda requirements were] fashioned. The only proper question on habeas is whether federal courts, acting under congressional directive to have the last say as to enforcement of federal constitutional principles, are to permit the States free enjoyment of the fruits of a conviction which by definition were only obtained through violations of the Constitution as interpreted in [Miranda].” Stone, 428 U. S., at 512, n. 10 (dissenting opinion) (emphasis in original).