SUPREME COURT OF THE UNITED STATES
24 FEBRUARY 2010
(delivered the opinion of the Court)
We consider whether a break in custody ends the presumption of involuntariness established in Edwards v Arizona, 451 U.S. 477 (1981).
In August 2003, a social worker assigned to the Child Advocacy Center in the Criminal Investigation Division of the Hagerstown Police Department referred to the department allegations that respondent Michael Shatzer, Sr., had sexually abused his 3-year-old son. At that time, Shatzer was incarcerated at the Maryland Correctional Institution-Hagerstown, serving a sentence for an unrelated child-sexual-abuse offense. Detective Shane Blankenship was assigned to the investigation and interviewed Shatzer at the correctional institution on August 7, 2003. Before asking any questions, Blankenship reviewed Shatzer’s Miranda rights with him, and obtained a written waiver of those rights. When Blankenship explained that he was there to question Shatzer about sexually abusing his son, Shatzer expressed confusion — he had thought Blankenship was an attorney there to discuss the prior crime for which he was incarcerated. Blankenship clarified the purpose of his visit, and Shatzer declined to speak without an attorney. Accordingly, Blankenship ended the interview, and Shatzer was released back into the general prison population. Shortly thereafter, Blankenship closed the investigation.
Two years and six months later, the same social worker referred more specific allegations to the department about the same incident involving Shatzer. Detective Paul Hoover, from the same division, was assigned to the investigation. He and the social worker interviewed the victim, then eight years old, who described the incident in more detail. With this new information in hand, on March 2, 2006, they went to the Roxbury Correctional Institute, to which Shatzer had since been transferred, and interviewed Shatzer in a maintenance room outfitted with a desk and three chairs. Hoover explained that he wanted to ask Shatzer about the alleged incident involving Shatzer’s son. Shatzer was surprised because he thought that the investigation had been closed, but Hoover explained they had opened a new file. Hoover then read Shatzer his Miranda rights and obtained a written waiver on a standard department form.
Hoover interrogated Shatzer about the incident for approximately 30 minutes. Shatzer denied ordering his son to perform fellatio on him, but admitted to masturbating in front of his son from a distance of less than three feet. Before the interview ended, Shatzer agreed to Hoover’s request that he submit to a polygraph examination. At no point during the interrogation did Shatzer request to speak with an attorney or refer to his prior refusal to answer questions without one.
Five days later, on March 7, 2006, Hoover and another detective met with Shatzer at the correctional facility to administer the polygraph examination. After reading Shatzer his Miranda rights and obtaining a written waiver, the other detective administered the test and concluded that Shatzer had failed. When the detectives then questioned Shatzer, he became upset, started to cry, and incriminated himself by saying, “‘I didn’t force him. I didn’t force him.’” 405 Md. 585, 590, 954 A. 2d 1118, 1121 (2008). After making this inculpatory statement, Shatzer requested an attorney, and Hoover promptly ended the interrogation.
The State’s Attorney for Washington County charged Shatzer with second-degree sexual offense, sexual child abuse, second-degree assault, and contributing to conditions rendering a child in need of assistance. Shatzer moved to suppress his March 2006 statements pursuant to Edwards. The trial court held a suppression hearing and later denied Shatzer’s motion. The Edwards protections did not apply, it reasoned, because Shatzer had experienced a break in custody for Miranda purposes between the 2003 and 2006 interrogations. No. 21–K–06–37799 (Cir. Ct. Washington Cty., Md., Sept. 14, 2006), App. 55. Shatzer pleaded not guilty, waived his right to a jury trial, and proceeded to a bench trial based on an agreed statement of facts. In accordance with the agreement, the State described the interview with the victim and Shatzer’s 2006 statements to the detectives. Based on the proffered testimony of the victim and the “admission of the defendant as to the act of masturbation,” the trial court found Shatzer guilty of sexual child abuse of his son. No. 21–K–06–37799 (Cir. Ct. Washington Cty., Md., Sept. 21, 2006), id., at 70, 79.
Over the dissent of two judges, the Court of Appeals of Maryland reversed and remanded. The court held that “the passage of time alone is insufficient to [end] the protections afforded by Edwards,” and that, assuming, arguendo, a break-in-custody exception to Edwards existed, Shatzer’s release back into the general prison population between interrogations did not constitute a break in custody. 405 Md., at 606–607, 954 A. 2d, at 1131. We granted certiorari, 555 U.S. ___ (2009).
The Fifth Amendment, which applies to the States by virtue of the Fourteenth Amendment, Malloy v Hogan, 378 U.S. 1, 6 (1964), provides that “[n]o person .... shall be compelled in any criminal case to be a witness against himself.” U.S. Const., Amdt. 5. In Miranda v Arizona, 384 U.S. 436 (1966), the Court adopted a set of prophylactic measures to protect a suspect’s Fifth Amendment right from the “inherently compelling pressures” of custodial interrogation. Id., at 467. The Court observed that “incommunicado interrogation” in an “unfamiliar,” “police-dominated atmosphere,” id. , at 456–457, involves psychological 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 467. Consequently, it reasoned, “[u]nless 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.” Id., at 458.
To counteract the coercive pressure, Miranda announced that police officers must warn a suspect prior to questioning that he has a right to remain silent, and a right to the presence of an attorney. Id., at 444. After the warnings are given, if the suspect indicates that he wishes to remain silent, the interrogation must cease. Id., at 473–474. Similarly, if the suspect states that he wants an attorney, the interrogation must cease until an attorney is present. Id., at 474. Critically, however, a suspect can waive these rights. Id., at 475. To establish a valid waiver, the State must show that the waiver was knowing, intelligent, and voluntary under the “high standar[d] of proof for the waiver of constitutional rights [set forth in] Johnson v Zerbst, 304 U.S. 458 (1938).” Id., at 475.
In Edwards, the Court determined that Zerbst’s traditional standard for waiver was not sufficient to protect a suspect’s right to have counsel present at a subsequent interrogation if he had previously requested counsel; “additional safeguards” were necessary. 451 U.S., at 484. The Court therefore superimposed a “second layer of prophylaxis,” McNeil v Wisconsin, 501 U.S. 171, 176 (1991). Edwards held 451 U.S., at 484–485:
|[W]hen an accused has invoked his right to have counsel present during custodial interrogation, a valid waiver of that right cannot be established by showing only that he responded to further police-initiated custodial interrogation even if he has been advised of his rights .... [He] is not subject to further interrogation by the authorities until counsel has been made available to him, unless the accused himself initiates further communication, exchanges, or conversations with the police.|
The rationale of Edwards is that once a suspect indicates that “he is not capable of undergoing [custodial] questioning without advice of counsel,” “any subsequent waiver that has come at the authorities’ behest, and not at the suspect’s own instigation, is itself the product of the ‘inherently compelling pressures’ and not the purely voluntary choice of the suspect.” Arizona v Roberson, 486 U.S. 675, 681 (1988). Under this rule, a voluntary Miranda waiver is sufficient at the time of an initial attempted interrogation to protect a suspect’s right to have counsel present, but it is not sufficient at the time of subsequent attempts if the suspect initially requested the presence of counsel. The implicit assumption, of course, is that the subsequent requests for interrogation pose a significantly greater risk of coercion. That increased risk results not only from the police’s persistence in trying to get the suspect to talk, but also from the continued pressure that begins when the individual is taken into custody as a suspect and sought to be interrogated — pressure likely to “increase as custody is prolonged,” Minnick v Mississippi, 498 U.S. 146, 153 (1990). The Edwards presumption of involuntariness ensures that police will not take advantage of the mounting coercive pressures of “prolonged police custody,” Roberson , 486 U.S., at 686, by repeatedly attempting to question a suspect who previously requested counsel until the suspect is “badgered into submission,” id., at 690 (Kennedy, J., dissenting).
We have frequently emphasized that the Edwards rule is not a constitutional mandate, but judicially prescribed prophylaxis. See, e.g., Montejo v Louisiana, 556 U.S. ___, ___ (2009) (slip op., at 7–8); Michigan v Harvey , 494 U.S. 344, 349 (1990) ; Solem v Stumes, 465 U.S. 638, n. 4 (1984). Because Edwards is “our rule, not a constitutional command,” “it is our obligation to justify its expansion.” Roberson, supra, at 688 (Kennedy, J., dissenting). Lower courts have uniformly held that a break in custody ends the Edwards presumption, see, e.g., People v Storm, 28 Cal. 4th 1007, 1023–1024, and n. 6, 52 P. 3d 52, 61–62, and n. 6 (2002) (collecting state and federal cases), but we have previously addressed the issue only in dicta, see McNeil, supra, at 177 (Edwards applies “assuming there has been no break in custody”).
A judicially crafted rule is “justified only by reference to its prophylactic purpose,” Davis v United States, 512 U.S. 452, 458 (1994) (internal quotation marks omitted), and applies only where its benefits outweigh its costs, Montejo, supra, at ___ (slip op., at 14). We begin with the benefits. Edwards’ presumption of involuntariness has the incidental effect of “conserv[ing] judicial resources which would otherwise be expended in making difficult determinations of voluntariness.” Minnick, supra, at 151. Its fundamental purpose, however, is to “[p]reserv[e] the integrity of an accused’s choice to communicate with police only through counsel,” Patterson v Illinois, 487 U.S. 285, 291 (1988), by “prevent[ing] police from badgering a defendant into waiving his previously asserted Miranda rights,” Harvey, supra, at 350. Thus, the benefits of the rule are measured by the number of coerced confessions it suppresses that otherwise would have been admitted. See Montejo, supra, at ___ (slip op., at 14).
It is easy to believe that a suspect may be coerced or badgered into abandoning his earlier refusal to be questioned without counsel in the paradigm Edwards case. That is a case in which the suspect has been arrested for a particular crime and is held in uninterrupted pretrial custody while that crime is being actively investigated. After the initial interrogation, and up to and including the second one, he remains cut off from his normal life and companions, “thrust into” and isolated in an “unfamiliar,” “police-dominated atmosphere,” Miranda, 384 U. S., at 456–457, where his captors “appear to control [his] fate,” Illinois v Perkins, 496 U.S. 292, 297 (1990). That was the situation confronted by the suspects in Edwards, Roberson, and Minnick, the three cases in which we have held the Edwards rule applicable. Edwards was arrested pursuant to a warrant and taken to a police station, where he was interrogated until he requested counsel. Edwards, 451 U.S., at 478–479. The officer ended the interrogation and took him to the county jail, but at 9:15 the next morning, two of the officer’s colleagues reinterrogated Edwards at the jail. Id., at 479. Roberson was arrested “at the scene of a just-completed burglary” and interrogated there until he requested a lawyer. Roberson, 486 U.S., at 678. A different officer interrogated him three days later while he “was still in custody pursuant to the arrest.” Ibid. Minnick was arrested by local police and taken to the San Diego jail, where two FBI agents interrogated him the next morning until he requested counsel. Minnick, 498 U.S., at 148–149. Two days later a Mississippi Deputy Sheriff reinterrogated him at the jail. Id., at 149. None of these suspects regained a sense of control or normalcy after they were initially taken into custody for the crime under investigation.
When, unlike what happened in these three cases, a suspect has been released from his pretrial custody and has returned to his normal life for some time before the later attempted interrogation, there is little reason to think that his change of heart regarding interrogation without counsel has been coerced. He has no longer been isolated. He has likely been able to seek advice from an attorney, family members, and friends. And he knows from his earlier experience that he need only demand counsel to bring the interrogation to a halt; and that investigative custody does not last indefinitely. In these circumstances, it is far fetched to think that a police officer’s asking the suspect whether he would like to waive his Miranda rights will any more “wear down the accused,” Smith v Illinois , 469 U.S. 91, 98 (1984) (per curiam), than did the first such request at the original attempted interrogation — which is of course not deemed coercive. His change of heart is less likely attributable to “badgering” than it is to the fact that further deliberation in familiar surroundings has caused him to believe (rightly or wrongly) that cooperating with the investigation is in his interest. Uncritical extension of Edwards to this situation would not significantly increase the number of genuinely coerced confessions excluded. The “justification for a conclusive presumption disappears when application of the presumption will not reach the correct result most of the time.” Coleman v Thompson, 501 U.S. 722, 737 (1991).
At the same time that extending the Edwards rule yields diminished benefits, extending the rule also increases its costs: the in-fact voluntary confessions it excludes from trial, and the voluntary confessions it deters law enforcement officers from even trying to obtain. Voluntary confessions are not merely “a proper element in law enforcement,” Miranda, supra, at 478, they are an “unmitigated good,” McNeil, 501 U.S., at 181, “‘essential to society’s compelling interest in finding, convicting, and punishing those who violate the law,’” ibid. (quoting Moran v Burbine , 475 U.S. 412, 426 (1986)).
The only logical endpoint of Edwards disability is termination of Miranda custody and any of its lingering effects. Without that limitation — and barring some purely arbitrary time-limit — every Edwards prohibition of custodial interrogation of a particular suspect would be eternal. The prohibition applies, of course, when the subsequent interrogation pertains to a different crime, Roberson, supra, when it is conducted by a different law enforcement authority, Minnick, 498 U.S. 146, and even when the suspect has met with an attorney after the first interrogation, ibid. And it not only prevents questioning ex ante; it would render invalid ex post, confessions invited and obtained from suspects who (unbeknownst to the interrogators) have acquired Edwards immunity previously in connection with any offense in any jurisdiction. In a country that harbors a large number of repeat offenders, this consequence is disastrous.
We conclude that such an extension of Edwards is not justified; we have opened its “protective umbrella,” Solem, 465 U.S., at 644, n. 4, far enough. The protections offered by Miranda, which we have deemed sufficient to ensure that the police respect the suspect’s desire to have an attorney present the first time police interrogate him, adequately ensure that result when a suspect who initially requested counsel is reinterrogated after a break in custody that is of sufficient duration to dissipate its coercive effects.
If Shatzer’s return to the general prison population qualified as a break in custody (a question we address in Part III, infra), there is no doubt that it lasted long enough (2 years) to meet that durational requirement. But what about a break that has lasted only one year? Or only one week? It is impractical to leave the answer to that question for clarification in future case-by-case adjudication; law enforcement officers need to know, with certainty and beforehand, when renewed interrogation is lawful. And while it is certainly unusual for this Court to set forth precise time limits governing police action, it is not unheard-of. In County of Riverside v McLaughlin, 500 U.S. 44 (1991), we specified 48 hours as the time within which the police must comply with the requirement of Gerstein v Pugh, 420 U.S. 103 (1975), that a person arrested without a warrant be brought before a magistrate to establish probable cause for continued detention.
Like McLaughlin, this is a case in which the requisite police action (there, presentation to a magistrate; here, abstention from further interrogation) has not been prescribed by statute but has been established by opinion of this Court. We think it appropriate to specify a period of time to avoid the consequence that continuation of the Edwards presumption “will not reach the correct result most of the time.” Coleman, supra, at 737. It seems to us that period is 14 days. That provides plenty of time for the suspect to get reacclimated to his normal life, to consult with friends and counsel, and to shake off any residual coercive effects of his prior custody.
The 14-day limitation meets Shatzer’s concern that a break-in-custody rule lends itself to police abuse. He envisions that once a suspect invokes his Miranda right to counsel, the police will release the suspect briefly (to end the Edwards presumption) and then promptly bring him back into custody for reinterrogation. But once the suspect has been out of custody long enough (14 days) to eliminate its coercive effect, there will be nothing to gain by such gamesmanship — nothing, that is, except the entirely appropriate gain of being able to interrogate a suspect who has made a valid waiver of his Miranda rights.
Shatzer argues that ending the Edwards protections at a break in custody will undermine Edwards’ purpose to conserve judicial resources. To be sure, we have said that “[t]he merit of the Edwards decision lies in the clarity of its command and the certainty of its application.” Minnick, 498 U.S., at 151. But clarity and certainty are not goals in themselves. They are valuable only when they reasonably further the achievement of some substantive end — here, the exclusion of compelled confessions. Confessions obtained after a 2-week break in custody and a waiver of Miranda rights are most unlikely to be compelled, and hence are unreasonably excluded. In any case, a break-in-custody exception will dim only marginally, if at all, the bright-line nature of Edwards. In every case involving Edwards, the courts must determine whether the suspect was in custody when he requested counsel and when he later made the statements he seeks to suppress. Now, in cases where there is an alleged break in custody, they simply have to repeat the inquiry for the time between the initial invocation and reinterrogation. In most cases that determination will be easy. And when it is determined that the defendant pleading Edwards has been out of custody for two weeks before the contested interrogation, the court is spared the fact-intensive inquiry into whether he ever, anywhere, asserted his Miranda right to counsel.
The facts of this case present an additional issue. No one questions that Shatzer was in custody for Miranda purposes during the interviews with Detective Blankenship in 2003 and Detective Hoover in 2006. Likewise, no one questions that Shatzer triggered the Edwards protections when, according to Detective Blankenship’s notes of the 2003 interview, he stated that “‘he would not talk about this case without having an attorney present,’” 405 Md., at 589, 954 A. 2d, at 1120. After the 2003 interview, Shatzer was released back into the general prison population where he was serving an unrelated sentence. The issue is whether that constitutes a break in Miranda custody.
We have never decided whether incarceration constitutes custody for Miranda purposes, and have indeed explicitly declined to address the issue. See Perkins, 496 U.S., at 299. See also Bradley v Ohio, 497 U.S. 1011, 1013 (1990) (Marshall, J., dissenting from denial of certiorari). Whether it does depends upon whether it exerts the coercive pressure that Miranda was designed to guard against — the “danger of coercion [that] results from the interaction of custody and official interrogation.” Perkins, supra, at 297 (emphasis added). To determine whether a suspect was in Miranda custody we have asked whether “there is a ‘formal arrest or restraint on freedom of movement’ of the degree associated with a formal arrest.” New York v Quarles, 467 U.S. 649, 655 (1984) ; see also Stansbury v California, 511 U.S. 318, 322 (1994) (per curiam). This test, no doubt, is satisfied by all forms of incarceration. Our cases make clear, however, that the freedom-of-movement test identifies only a necessary and not a sufficient condition for Miranda custody. We have declined to accord it “talismanic power,” because Miranda is to be enforced “only in those types of situations in which the concerns that powered the decision are implicated.” Berkemer v McCarty, 468 U.S. 420, 437 (1984). Thus, the temporary and relatively nonthreatening detention involved in a traffic stop or Terry stop, see Terry v Ohio, 392 U.S. 1 (1968), does not constitute Miranda custody. McCarty , supra, at 439–440. See also Perkins, supra, at 296.
Here, we are addressing the interim period during which a suspect was not interrogated, but was subject to a baseline set of restraints imposed pursuant to a prior conviction. Without minimizing the harsh realities of incarceration, we think lawful imprisonment imposed upon conviction of a crime does not create the coercive pressures identified in Miranda.
Interrogated suspects who have previously been convicted of crime live in prison. When they are released back into the general prison population, they return to their accustomed surroundings and daily routine — they regain the degree of control they had over their lives prior to the interrogation. Sentenced prisoners, in contrast to the Miranda paradigm, are not isolated with their accusers. They live among other inmates, guards, and workers, and often can receive visitors and communicate with people on the outside by mail or telephone.
Their detention, moreover, is relatively disconnected from their prior unwillingness to cooperate in an investigation. The former interrogator has no power to increase the duration of incarceration, which was determined at sentencing. And even where the possibility of parole exists, the former interrogator has no apparent power to decrease the time served. This is in stark contrast to the circumstances faced by the defendants in Edwards, Roberson , and Minnick, whose continued detention as suspects rested with those controlling their interrogation, and who confronted the uncertainties of what final charges they would face, whether they would be convicted, and what sentence they would receive.
Shatzer’s experience illustrates the vast differences between Miranda custody and incarceration pursuant to conviction. At the time of the 2003 attempted interrogation, Shatzer was already serving a sentence for a prior conviction. After that, he returned to the general prison population in the Maryland Correctional Institution-Hagerstown and was later transferred, for unrelated reasons, down the street to the Roxbury Correctional Institute. Both are medium-security state correctional facilities. See Maryland Div. of Correction Inmate Handbook 7 (2007), online at http://dpscs.md.gov/rehabservs/ doc/pdfs/2007_Inmate_Handbook.pdf (all Internet materials as visited Feb. 22, 2010, and available in Clerk of Court’s case file). Inmates in these facilities generally can visit the library each week, id., at 28; have regular exercise and recreation periods, id., at 17; can participate in basic adult education and occupational training, id., at 26, 7; are able to send and receive mail, id., at 21–22, 16; and are allowed to receive visitors twice a week, see http://dpscs.md.gov/locations/mcih.shtml; http://www. dpscs.state.md.us/locations/rci.shtml. His continued detention after the 2003 interrogation did not depend on what he said (or did not say) to Detective Blankenship, and he has not alleged that he was placed in a higher level of security or faced any continuing restraints as a result of the 2003 interrogation. The “inherently compelling pressures” of custodial interrogation ended when he returned to his normal life.
A few words in response to Justice Stevens’ concurrence: It claims we ignore that “[w]hen police tell an indigent suspect that he has the right to an attorney” and then “reinterrogate” him without providing a lawyer, “the suspect is likely to feel that the police lied to him and that he really does not have any right to a lawyer.” Post, at 2 (opinion concurring in judgment) (hereinafter concurrence). See also post, at 4, 7, n. 11, 11, n. 16. The fallacy here is that we are not talking about “reinterrogating” the suspect; we are talking about asking his permission to be interrogated. An officer has in no sense lied to a suspect when, after advising, as Miranda requires, “You have the right to remain silent, and if you choose to speak you have the right to the presence of an attorney,” he promptly ends the attempted interrogation because the suspect declines to speak without counsel present, and then, two weeks later, reapproaches the suspect and asks, “Are you now willing to speak without a lawyer present?”
The “concer[n] that motivated the Edwards line of cases,” post, at 2–3, n. 2, is that the suspect will be coerced into saying yes. That concern guides our decision today. Contrary to the concurrence’s conclusion, post, at 3, 5–6, there is no reason to believe a suspect will view confession as “‘the only way to end his interrogation’” when, before the interrogation begins, he is told that he can avoid it by simply requesting that he not be interrogated without counsel present — an option that worked before. If, as the concurrence argues will often be the case, post, at 5, a break in custody does not change the suspect’s mind, he need only say so.
The concurrence also accuses the Court of “ignor[ing] that when a suspect asks for counsel, until his request is answered, there are still the same ‘inherently compelling’ pressures of custodial interrogation on which the Miranda line of cases is based.” Post , at 4. We do not ignore these pressures; nor do we suggest that they disappear when custody is recommenced after a break, see post, at 5. But if those pressures are merely “the same” as before, then Miranda provides sufficient protection — as it did before. The Edwards presumption of involuntariness is justified only in circumstances where the coercive pressures have increased so much that suspects’ waivers of Miranda rights are likely to be involuntary most of the time. Contrary to the concurrence’s suggestion, post, at 3, it is only in those narrow circumstances — when custody is unbroken — that the Court has concluded a “fresh se[t] of Miranda warnings” is not sufficient. See Roberson, 486 U.S., at 686.
In the last analysis, it turns out that the concurrence accepts our principal points. It agrees that Edwards prophylaxis is not perpetual; it agrees that a break in custody reduces the inherently compelling pressure upon which Edwards was based; it agrees that Shatzer’s release back into the general prison population constituted a break in custody; and it agrees that in this case the break was long enough to render Edwards inapplicable. Post, at 10–12. We differ in two respects: Instead of terminating Edwards protection when the custodial pressures that were the basis for that protection dissipate, the concurrence would terminate it when the suspect would no longer “feel that he has ‘been denied the counsel he has clearly requested,’” post, at 11. This is entirely unrelated to the rationale of Edwards. If confidence in the police’s promise to provide counsel were the touchstone, Edwards would not have applied in Minnick, where the suspect in continuing custody actually met with appointed counsel. The concurrence’s rule is also entirely unrelated to the existence of a break in custody. While that may relieve the accumulated coercive pressures of custody that are the foundation for Edwards, it is hard to see how it bolsters the suspect’s confidence that if he asks for counsel he will get one.
And secondly, the concurrence differs from us in declining to say how long after a break in custody the termination of Edwards protection occurs. Two and one-half years, it says, is clearly enough — but it gives law enforcement authorities no further guidance. The concurrence criticizes our use of 14 days as arbitrary and unexplained, post, at 5, and n. 7. But in fact that rests upon the same basis as the concurrence’s own approval of a 2 1 2 -year break in custody: how much time will justify “treating the second interrogation as no more coercive than the first,” post, at 10. Failure to say where the line falls short of 2 1 2 years, and leaving that for future case-by-case determination, is certainly less helpful, but not at all less arbitrary.
* * *
Because Shatzer experienced a break in Miranda custody lasting more than two weeks between the first and second attempts at interrogation, Edwards does not mandate suppression of his March 2006 statements. Accordingly, we reverse the judgment of the Court of Appeals of Maryland, and remand the case for further proceedings not inconsistent with this opinion.
It is so ordered.
(concurring in part and concurring in the judgment)
I join Part III of the Court’s opinion, which holds that release into the general prison population constitutes a break in custody. I do not join the Court’s decision to extend the presumption of involuntariness established in Edwards v Arizona, 451 U.S. 477 (1981), for 14 days after custody ends.
It is not apparent to me that the presumption of involuntariness the Court recognized in Edwards is justifiable even in the custodial setting to which Edwards applies it. See, e.g., Minnick v Mississippi, 498 U.S. 146, 160 (1990) (Scalia, J., dissenting). Accordingly, I would not extend the Edwards rule “beyond the circumstances present in Edwards itself.” Id., at 162. But even if one believes that the Court is obliged to apply Edwards to any case involving continuing custody, the Court’s opinion today goes well beyond that. It extends the presumption of involuntariness Edwards applies in custodial settings to interrogations that occur after custody ends.
The Court concedes that this extension, like the Edwards presumption itself, is not constitutionally required. The Court nevertheless defends the extension as a judicially created prophylaxis against compelled confessions. Even if one accepts that such prophylaxis is both permissible generally and advisable for some period following a break in custody, the Court’s 14-day rule fails to satisfy the criteria our precedents establish for the judicial creation of such a safeguard.
Our precedents insist that judicially created prophylactic rules like those in Edwards and Miranda v Arizona, 384 U.S. 436 (1966), maintain “the closest possible fit” between the rule and the Fifth Amendment interests they seek to protect. United States v Patane, 542 U.S. 630, 640–641 (2004) (plurality opinion); see generally Montejo v Louisiana, 556 U.S. ___, ___ (2009) (slip op., at 18); Chavez v Martinez, 538 U.S. 760, 772 (2003) (plurality opinion). The Court’s 14-day rule does not satisfy this test. The Court relates its 14-day rule to the Fifth Amendment simply by asserting that 14 days between release and recapture should provide “plenty of time for the suspect .... to shake off any residual coercive effects of his prior custody,” ante, at 11.
This ipse dixit does not explain why extending the Edwards presumption for 14 days following a break in custody — as opposed to 0, 10, or 100 days — provides the “closest possible fit” with the Self-Incrimination Clause, Patane, supra, at 640–641; see ante, at 11 (merely stating that “[i]t seems to us that” the appropriate “period is 14 days”). Nor does it explain how the benefits of a prophylactic 14-day rule (either on its own terms or compared with other possible rules) “outweigh its costs” (which would include the loss of law enforcement information as well as the exclusion of confessions that are in fact voluntary). Ante, at 6 (citing Montejo, supra, at __ (slip op., at 14)).
To be sure, the Court’s rule has the benefit of providing a bright line. Ante, at 12. But bright-line rules are not necessary to prevent Fifth Amendment violations, as the Court has made clear when refusing to adopt such rules in cases involving other Miranda rights. See, e.g., Michigan v Mosley, 423 U.S. 96, 103–104 (1975). And an otherwise arbitrary rule is not justifiable merely because it gives clear instruction to law enforcement officers.
As the Court concedes, “clarity and certainty are not goals in themselves. They are valuable only when they reasonably further the achievement of some substantive end — here, the exclusion of compelled confessions” that the Fifth Amendment prohibits. Ante, at 12. The Court’s arbitrary 14-day rule fails this test, even under the relatively permissive criteria set forth in our precedents. Accordingly, I do not join that portion of the Court’s opinion.
(concurring in the judgment)
While I agree that the presumption from Edwards v Arizona, 451 U.S. 477 (1981), is not “eternal,” ante, at 9–10, and does not mandate suppression of Shatzer’s statement made after a 2-year break in custody, I do not agree with the Court’s newly announced rule: that Edwards always ceases to apply when there is a 14-day break in custody, ante, at 11.
In conducting its “cost-benefit” analysis, the Court demeans Edwards as a “‘second layer’” of “judicially prescribed prophylaxis,” ante, at 5, 6, 12, n. 7; see also ante, at 6 (describing Edwards as “‘our rule, not a constitutional command’” (quoting Arizona v Roberson, 486 U.S. 675, 688 (1988) ( Kennedy, J., dissenting))). The source of the holdings in the long line of cases that includes both Edwards and Miranda, however, is the Fifth Amendment’s protection against compelled self-incrimination applied to the “compulsion inherent in custodial” interrogation, Miranda v Arizona, 384 U.S. 436, 458 (1966), and the “significan[ce]” of “the assertion of the right to counsel,” Edwards, 451 U.S., at 485. The Court’s analysis today is insufficiently sensitive to the concerns that motivated the Edwards line of cases.
The most troubling aspect of the Court’s time-based rule is that it disregards the compulsion caused by a second (or third, or fourth) interrogation of an indigent suspect who was told that if he requests a lawyer, one will be provided for him. When police tell an indigent suspect that he has the right to an attorney, that he is not required to speak without an attorney present, and that an attorney will be provided to him at no cost before questioning, the police have made a significant promise. If they cease questioning and then reinterrogate the suspect 14 days later without providing him with a lawyer, the suspect is likely to feel that the police lied to him and that he really does not have any right to a lawyer.
When officers informed Shatzer of his rights during the first interrogation, they presumably informed him that if he requested an attorney, one would be appointed for him before he was asked any further questions. But if an indigent suspect requests a lawyer, “any further interrogation” (even 14 days later) “without counsel having been provided will surely exacerbate whatever compulsion to speak the suspect may be feeling.” Roberson, 486 U.S., at 686. When police have not honored an earlier commitment to provide a detainee with a lawyer, the detainee likely will “understan[d] his (expressed) wishes to have been ignored” and “may well see further objection as futile and confession (true or not) as the only way to end his interrogation.” Davis v United States, 512 U.S. 452, 472–473 (1994) (Souter, J., concurring in judgment). Cf. Cooper v Dupnik, 963 F. 2d 1220, 1225 (CA9 1992) (en banc) (describing an elaborate police task force plan to ignore a suspect’s requests for counsel, on the theory that such would induce hopelessness and thereby elicit an admission). Simply giving a “fresh se[t] of Miranda warnings” will not “‘reassure’ a suspect who has been denied the counsel he has clearly requested that his rights have remained untrammeled.” Roberson, 486 U.S., at 686.
The Court never explains why its rule cannot depend on, in addition to a break in custody and passage of time, a concrete event or state of affairs, such as the police having honored their commitment to provide counsel. Instead, the Court simply decides to create a time-based rule, and in so doing, disregards much of the analysis upon which Edwards and subsequent decisions were based. “[T]he assertion of the right to counsel” “[i]s a significant event.” Edwards, 451 U.S., at 485. As the Court today acknowledges, the right to counsel, like the right to remain silent, is one that police may “coerc[e] or badge[r],” ante, at 7, a suspect into abandoning. However, as discussed above, the Court ignores the effects not of badgering but of reinterrogating a suspect who took the police at their word that he need not answer questions without an attorney present. See Roberson, 486 U.S., at 686. The Court, moreover, ignores that when a suspect asks for counsel, until his request is answered, there are still the same “inherently compelling” pressures of custodial interrogation on which the Miranda line of cases is based, see 486 U.S., at 681, and that the concern about compulsion is especially serious for a detainee who has requested a lawyer, an act that signals his “inability to cope with the pressures of custodial interrogation,” id., at 686.
Instead of deferring to these well-settled understandings of the Edwards rule, the Court engages in its own speculation that a 14-day break in custody eliminates the compulsion that animated Edwards. But its opinion gives no strong basis for believing that this is the case. A 14-day break in custody does not eliminate the rationale for the initial Edwards rule: The detainee has been told that he may remain silent and speak only through a lawyer and that if he cannot afford an attorney, one will be provided for him. He has asked for a lawyer. He does not have one. He is in custody. And police are still questioning him. A 14-day break in custody does not change the fact that custodial interrogation is inherently compelling. It is unlikely to change the fact that a detainee “considers himself unable to deal with the pressures of custodial interrogation without legal assistance.” Roberson, 486 U.S., at 683. And in some instances, a 14-day break in custody may make matters worse  “[w]hen a suspect understands his (expressed) wishes to have been ignored” and thus “may well see further objection as futile and confession (true or not) as the only way to end his interrogation.” Davis, 512 U.S., at 472–473 (Souter, J., concurring in judgment).
The Court ignores these understandings from the Edwards line of cases and instead speculates that if a suspect is reinterrogated and eventually talks, it must be that “further deliberation in familiar surroundings has caused him to believe (rightly or wrongly) that cooperating with the investigation is in his interest.” Ante, at 9. But it is not apparent why that is the case. The answer, we are told, is that once a suspect has been out of Miranda custody for 14 days, “[h]e has likely been able to seek advice from an attorney, family members, and friends.” Ante, at 8. This speculation, however, is overconfident and only questionably relevant. As a factual matter, we do not know whether the defendant has been able to seek advice: First of all, suspects are told that if they cannot afford a lawyer, one will be provided for them. Yet under the majority’s rule, an indigent suspect who took the police at their word when he asked for a lawyer will nonetheless be assumed to have “been able to seek advice from an attorney.” Second, even suspects who are not indigent cannot necessarily access legal advice (or social advice as the Court presumes) within 14 days. Third, suspects may not realize that they need to seek advice from an attorney. Unless police warn suspects that the interrogation will resume in 14 days, why contact a lawyer? When a suspect is let go, he may assume that the police were satisfied. In any event, it is not apparent why interim advice matters. In Minnick v Mississippi, 498 U.S. 146, 153 (1990), we held that it is not sufficient that a detainee happened to speak at some point with a lawyer. See ibid. (noting that “consultation with an attorney” does not prevent “persistent attempts by officials to persuade [a suspect] to waive his rights” or shield against the “coercive pressures that accompany custody”). If the actual interim advice of an attorney is not sufficient, the hypothetical, interim advice of “an attorney, family members, and friends,” ante , at 8, is not enough.
The many problems with the Court’s new rule are exacerbated in the very situation in this case: a suspect who is in prison. Even if, as the Court assumes, a trip to one’s home significantly changes the Edwards calculus, a trip to one’s prison cell is not the same. A prisoner’s freedom is severely limited, and his entire life remains subject to government control. Such an environment is not conducive to “shak[ing] off any residual coercive effects of his prior custody.” Ante, at 11. Nor can a prisoner easily “seek advice from an attorney, family members, and friends,” ante, at 8, especially not within 14 days; prisoners are frequently subject to restrictions on communications. Nor, in most cases, can he live comfortably knowing that he cannot be badgered by police; prison is not like a normal situation in which a suspect “is in control, and need only shut his door or walk away to avoid police badgering.” Montejo v Louisiana, 556 U.S. ___, ___ (2009) (slip op., at 16). Indeed, for a person whose every move is controlled by the State, it is likely that “his sense of dependence on, and trust in, counsel as the guardian of his interests in dealing with government officials intensified.” United States v Green , 592 A. 2d 985, 989 (D. C. 1991); cf. Minnick, 498 U.S., at 153 (explaining that coercive pressures “may increase as custody is prolonged”). The Court ignores these realities of prison, and instead rests its argument on the supposition that a prisoner’s “detention .... is relatively disconnected from their prior unwillingness to cooperate in an investigation.” Ante, at 14. But that is not necessarily the case. Prisoners are uniquely vulnerable to the officials who control every aspect of their lives; prison guards may not look kindly upon a prisoner who refuses to cooperate with police. And cooperation frequently is relevant to whether the prisoner can obtain parole. See, e.g., Code of Md. Regs., tit. 12, §08.01.18(A)(3) (2008). Moreover, even if it is true as a factual matter that a prisoner’s fate is not controlled by the police who come to interrogate him, how is the prisoner supposed to know that? As the Court itself admits, compulsion is likely when a suspect’s “captors appear to control [his] fate,” ante, at 7 (internal quotation marks omitted). But when a guard informs a suspect that he must go speak with police, it will “appear” to the prisoner that the guard and police are not independent. “Questioning by captors, who appear to control the suspect’s fate, may create mutually reinforcing pressures that the Court has assumed will weaken the suspect’s will.” Illinois v Perkins , 496 U.S. 292, 297 (1990) (emphasis added).
Because, at the very least, we do not know whether Shatzer could obtain a lawyer, and thus would have felt that police had lied about providing one, I cannot join the Court’s opinion. I concur in today’s judgment, however, on another ground: Even if Shatzer could not consult a lawyer and the police never provided him one, the 2-year break in custody is a basis for treating the second interrogation as no more coercive than the first. Neither a break in custody nor the passage of time has an inherent, curative power. But certain things change over time. An indigent suspect who took police at their word that they would provide an attorney probably will feel that he has “been denied the counsel he has clearly requested,” Roberson, 486 U.S., at 686, when police begin to question him, without a lawyer, only 14 days later. But, when a suspect has been left alone for a significant period of time, he is not as likely to draw such conclusions when the police interrogate him again. It is concededly “impossible to determine with precision” where to draw such a line. Barker v Wingo, 407 U.S. 514, 521 (1972). In the case before us, however, the suspect was returned to the general prison population for 2 years. I am convinced that this period of time is sufficient. I therefore concur in the judgment.
 The State filed a nolle prosequi to the second-degree sexual offense charge, and consented to dismissal of the misdemeanor charges as barred by the statute of limitations.
 Jail is a “local government’s detention center where persons awaiting trial or those convicted of misdemeanors are confined.” Black’s Law Dictionary 910 (9th ed. 2009). Prison, by contrast, is a “state or federal facility of confinement for convicted criminals, esp. felons.” Id., at 1314.
 Justice Stevens points out, post, at 7 (opinion concurring in judgment), that in Minnick, actual pre-reinterrogation consultation with an attorney during continued custody did not suffice to avoid application of Edwards. That does not mean that the ability to consult freely with attorneys and others does not reduce the level of coercion at all, or that it is “only questionably relevant,” post, at 7, to whether termination of custody reduces the coercive pressure that is the basis for Edwards’ super-prophylactic rule.
 The State’s alternative argument in the present case is that the substantial lapse in time between the 2003 and 2006 attempts at interrogation independently ended the Edwards presumption. Our disposition makes it unnecessary to address that argument.
 This assumes that Roberson’s extension of Edwards to subsequent interrogation for a different crime, and Minnick’s extension of Edwards to subsequent interrogation by a different law enforcement agency would apply even when the place of custody and the identity of the custodial agency are not the same (as they were in Roberson and Minnick) as those of the original interrogation. That assumption would seem reasonable if the Edwards-suspending effect of a termination of custody is rejected. Reinterrogation in different custody or by a different interrogating agency would seem, if anything, less likely than termination of custody to reduce coercive pressures. At the original site, and with respect to the original interrogating agency, the suspect has already experienced cessation of interrogation when he demands counsel — which he may have no reason to expect elsewhere.
 According to a recent study, 67.5% of prisoners released from 15 States in 1994 were rearrested within three years. See Dept. of Justice, Bureau of Justice Statistics, Special Report, Recidivism of Prisoners Released in 1994 (NCJ 193427, 2002).
 A defendant who experiences a 14-day break in custody after invoking the Miranda right to counsel is not left without protection. Edwards establishes a presumption that a suspect’s waiver of Miranda rights is involuntary. See Arizona v. Roberson, 486 U.S. 675, 681 (1988). Even without this “second layer of prophylaxis,” McNeil v. Wisconsin, 501 U.S. 171, 176 (1991), a defendant is still free to claim the prophylactic protection of Miranda — arguing that his waiver of Miranda rights was in fact involuntary under Johnson v. Zerbst, 304 U.S. 458 (1938). See Miranda, 384 U.S., at 475.
 We distinguish the duration of incarceration from the duration of what might be termed interrogative custody. When a prisoner is removed from the general prison population and taken to a separate location for questioning, the duration of that separation is assuredly dependent upon his interrogators. For which reason once he has asserted a refusal to speak without assistance of counsel Edwards prevents any efforts to get him to change his mind during that interrogative custody.
 At a minimum the latter proposition is questionable. I concede that some police officers might badger a suspect during a subsequent interrogation after a break in custody, or might use catch-and-release tactics to suggest they will not take no for an answer. But if a suspect reenters custody after being questioned and released, he need only invoke his right to counsel to ensure Edwards’ protection for the duration of the subsequent detention. And, if law enforcement officers repeatedly release and recapture a suspect to wear down his will — such that his participation in a subsequent interrogation is no longer truly voluntary — the “high standar[d] of proof for the waiver of constitutional rights [set forth in] Johnson v. Zerbst, 304 U.S. 458 (1938),” will protect against the admission of the suspect’s statements in court. Miranda v. Arizona, 384 U.S. 436, 475 (1966). The Zerbst inquiry takes into account the totality of the circumstances surrounding the waiver — including any improper pressures by police. See id., at 464; cf. ante, at 11–12, n. 6 (stating that “[e]ven without [Edwards’] second layer of prophylaxis, a defendant is still free to claim the prophylactic protection of Miranda — arguing that his waiver of Miranda rights was in fact involuntary under Johnson v. Zerbst” (internal quotation marks and citation omitted)).
 Though the Court asserts that its 14-day rule will tell “law enforcement officers .... with certainty and beforehand, when renewed interrogation is lawful,” ante, at 10, that is not so clear. Determining whether a suspect was previously in custody, and when the suspect was released, may be difficult without questioning the suspect, especially if state and federal authorities are conducting simultaneous investigations.
 See Dickerson v. United States, 530 U.S. 428, 438 (2000) (holding that “the protections announced in Miranda” are “constitutionally required”); Shea v. Louisiana, 470 U.S. 51, 52 (1985) (“In Edwards ..., this Court ruled that a criminal defendant’s rights under the Fifth and Fourteenth Amendment s were violated by the use of his confession obtained by police-instigated interrogation — without counsel present — after he requested an attorney”); Oregon v. Bradshaw, 462 U.S. 1039, 1043 (1983) (plurality opinion) (“[The] subsequent incriminating statements made without [an] attorney present violated the rights secured to the defendant by the Fifth and Fourteenth Amendment s to the United States Constitution”); Miranda, 384 U.S., at 458 (examining the “history and precedent underlying the Self-Incrimination Clause to determine its applicability in this situation”).
 The Court states that this argument rests on a “fallacy” because “we are not talking about ‘reinterrogating’ the suspect; we are talking about asking his permission to be interrogated.” Ante, at 16 (emphasis deleted). Because, however, a suspect always has the right to remain silent, this is a distinction without a difference: Any time that the police interrogate or reinterrogate, and read a suspect his Miranda rights, the suspect may decline to speak. And if this is a “fallacy,” it is the same “fallacy” upon which this Court has relied in the Edwards line of cases that held that police may not continue to interrogate a suspect who has requested a lawyer: Police may not continue to ask such a suspect whether they may interrogate him until that suspect has a lawyer present. The Court’s apparent belief that this is a “fallacy” only underscores my concern that its analysis is insufficiently sensitive to the concerns that motivated the Edwards line of cases.
 Indeed, a lawyer has a “unique ability to protect the Fifth Amendment rights of a client undergoing custodial interrogation.” Fare v. Michael C., 442 U.S. 707, 719 (1979). Counsel can curb an officer’s overbearing conduct, advise a suspect of his rights, and ensure that there is an accurate record of any interrogation. “Because of this special ability of the lawyer to help the client preserve his Fifth Amendment rights once the client becomes enmeshed in the adversary process, the Court found that the right to have counsel present at the interrogation is indispensible to the protection of the Fifth Amendment privilege.” Arizona v. Roberson, 486 U.S. 675, n. 4 (1988) (internal quotation marks omitted). Thus, “once the accused has requested counsel,” courts must be especially wary of “coercive form[s] of custodial interrogation.” Bradshaw, 462 U.S., at 1051 (Powell, J., concurring in judgment).
 See Michigan v. Harvey, 494 U.S. 344, 350 (1990) (subsequent confession suggests the police “badger[ed] a defendant into waiving his previously asserted Miranda rights”).
 See Minnick v. Mississippi, 498 U.S. 146, 155 (1990) (“[N]either admissions nor waivers are effective unless there are both particular and systemic assurances that the coercive pressures of custody were not the inducing cause”); cf. Smith v. Illinois, 469 U.S. 91, 98 (1984) (per curiam) (“[T]he authorities through ‘badger[ing]’ or ‘overreaching’ — explicit or subtle, deliberate or unintentional — might otherwise wear down the accused and persuade him to incriminate himself notwithstanding his earlier request for counsel’s assistance”).
 See Roberson, 486 U.S., at 681 (“[I]f a suspect believes that he is not capable of undergoing such questioning without advice of counsel, then it is presumed that any subsequent waiver that has come at the authorities’ behest, and not at the suspect’s own instigation, is itself the product of the ‘inherently compelling pressures’”); Michigan v. Mosley, 423 U.S. 96, 110, n. 2 (1975) (White, J., concurring in result) (“[T]he accused having expressed his own view that he is not competent to deal with the authorities without legal advice, a later decision at the authorities’ insistence to make a statement without counsel’s presence may properly be viewed with skepticism”).
 Today’s decision, moreover, offers no reason for its 14-day time period. To be sure, it may be difficult to marshal conclusive evidence when setting an arbitrary time period. But in light of the basis for Edwards, we should tread carefully. Instead, the only reason for choosing a 14-day time period, the Court tells us, is that “[i]t seems to us that period is 14 days.” Ante, at 11. That time period is “plenty of time for the suspect to get reacclimated to his normal life, to consult with friends and counsel, and to shake off any residual coercive effects of his prior custody.” ibid. But the Court gives no reason for that speculation, which may well prove inaccurate in many circumstances.
 In Roberson, for example, we observed that once a suspect has asserted his right to an attorney, courts must presume he does “not feel sufficiently comfortable with the pressures of custodial interrogation to answer questions without an attorney. This discomfort is precisely the state of mind that Edwards presumes to persist ....” 486 U.S., at 684. We held in Roberson that just because different police come to speak about a different investigation, that presumption does not change: “[T]here is no reason to assume that a suspect’s state of mind is in any way investigation-specific.” ibid. Nor is there any reason to believe that it is arrest specific.
 The compulsion is heightened by the fact that “[t]he uncertainty of fate that being released from custody and then reapprehended entails is, in some circumstances, more coercive than continual custody.” Strauss, Reinterrogation, 22 Hastings Const. L. Q. 359, 390 (1995).
 Not only is this a likely effect of reinterrogation, but police may use this effect to their advantage. Indeed, the Court’s rule creates a strange incentive to delay formal proceedings, in order to gain additional information by way of interrogation after the time limit lapses. The justification for Fifth Amendment rules “must be consistent with .... practical realities,” Roberson, 486 U.S., at 688 (Kennedy, J., dissenting), and the reality is that police may operate within the confines of the Fifth Amendment in order to extract as many confessions as possible, see Leo & White, Adapting to Miranda: Modern Interrogators’ Strategies for Dealing with the Obstacles Posed by Miranda, 84 Minn. L. Rev. 397 (1999). With a time limit as short as 14 days, police who hope that they can eventually extract a confession may feel comfortable releasing a suspect for a short period of time. The resulting delay will only increase the compelling pressures on the suspect.
 It is important to distinguish this from the point that I make above about indigent suspects. If the police promise to provide a lawyer and never do so, it sends a message to the suspect that the police have lied and that the rights read to him are hollow. But the mere fact that a suspect consulted a lawyer does not itself reduce the compulsion when police reinterrogate him.
 Cf. Orozco v. Texas, 394 U.S. 324, 326 (1969) (holding that a suspect was in custody while being held in own home, despite his comfort and familiarity with the surroundings); Mathis v. United States, 391 U.S. 1, 5 (1968) (holding that a person serving a prison sentence for one crime was in custody when he was interrogated in prison about another, unrelated crime); Miranda v. Arizona, 384 U.S. 436, 478 (1966) (“[W]hen an individual is .... deprived of his freedom by the authorities in any significant way and is subjected to questioning, the privilege against self-incrimination is jeopardized”).
 Prison also presents a troubling set of incentives for police. First, because investigators know that their suspect is also a prisoner, there is no need formally to place him under arrest. Thus, police generally can interview prisoners even without probable cause to hold them. This means that police can interrogate suspects with little or no evidence of guilt, and police can do so time after time, without fear of being sued for wrongful arrest. Second, because police know that their suspect is otherwise detained, there is no need necessarily to resolve the case quickly. Police can comfortably bide their time, interrogating and reinterrogating their suspect until he slips up. Third, because police need not hold their suspect, they do not need to arraign him or otherwise initiate formal legal proceedings that would trigger various protections.
 The Court attempts to distinguish detention in prison from the “paradigm Edwards case,” ante, at 7, but it is not clear why that is so. The difference cannot be simply that convicted prisoners’ “detention .... is relatively disconnected from their prior unwillingness to cooperate in an investigation,” ante, at 14, because in many instances of pretrial custody, the custody will continue regardless of whether a detainee answers questions. Take Roberson for example. Roberson was arrested and being held for one crime when, days later, a different officer interrogated him about a different crime. 486 U.S., at 678. Regardless of whether he cooperated with the second investigation, he was still being held for the first crime. Yet under the Court’s analysis, had Roberson been held long enough that he had become “accustomed” to the detention facility, ante, at 14, there would have been a break in custody between each interrogation. Thus, despite the fact that coercive pressures “may increase as custody is prolonged,” Minnick, 498 U.S., at 153, the real problem in Roberson may have been that the police did not leave him sitting in jail for long enough. This problem of pretrial custody also highlights a tension with the Court’s decision last Term in Montejo v. Louisiana, 556 U.S. ___ (2009). In Montejo, the Court overturned Michigan v. Jackson, 475 U.S. 625, 636 (1986), which had protected an accused’s Sixth Amendment right to counsel by “forbidding police to initiate interrogation of a criminal defendant once he has requested counsel at an arraignment or similar proceeding.” 556 U.S., at ___ (slip op., at 1). In so doing, the Court emphasized that because the Edwards “regime suffices to protect the integrity of ‘a suspect’s voluntary choice not to speak outside his lawyer’s presence,’ before his arraignment, it is hard to see why it would not also suffice to protect that same choice after arraignment.” 556 U.S., at ___ (slip op., at 15) (quoting Texas v. Cobb, 532 U.S. 162, 175 (2001) (Kennedy, J., concurring); citation omitted). But typically, after arraignment, defendants are released on bail or placed in detention facilities, both of which, according to the majority’s logic, sometimes constitute breaks in custody. How then, under the Court’s decision today, will Edwards serve the role that the Court placed on it in Montejo?
 The Court responds that “[i]f confidence in the police’s promise to provide counsel were the touchstone, Edwards would not have applied in Minnick, where the suspect in continuing custody actually met with appointed counsel.” Ante, at 17–18. But my view is not that “confidence in the police’s promise to provide counsel” is “the touchtone.” Ante, at 17. Rather, my view is that although an appropriate break in custody will mitigate many of the reasons that custodial reinterrogation of a suspect who requested counsel is inherently compelling, it will not mitigate the effect of an indigent detainee believing that he has “been denied the counsel he has clearly requested,” Roberson, 486 U.S., at 686. If police tell an indigent suspect that he is not required to speak without an attorney, and that they will provide him with an attorney, and that suspect asserts his right to an attorney, but police nonetheless do not provide an attorney and reinterrogate him (even if there was a break in custody between the interrogations), the indigent suspect is likely to feel that the police lied to him or are ignoring his rights. This view is not in tension with Minnick. Minnick holds only that consultation with an attorney between interrogations is not sufficient to end the Edwards presumption and therefore that when there has been no break in custody, “counsel’s presence at interrogation,” 498 U.S., at 152, is necessary to address the compulsion with which the Edwards line of cases is concerned.
 I do not doubt that some of the compulsion caused by reinterrogating an indigent suspect without providing a lawyer may survive even a break in custody and a very long passage of time. The relevant point here is more limited: A long break in time, far longer than 14 days, diminishes, rather than eliminates, that compulsion.
all rights reserved