Remaining Silent after Salinas

Brandon L. Garrett

80 U Chi L Rev Dialogue 116 [Essay PDF]


     In Salinas v Texas,1 the Supreme Court eroded what little remained of the Fifth Amendment jurisprudence protecting against coercive police questioning of suspects. The result encourages precisely the types of informal, undocumented questioning that can cause false confessions and wrongful convictions. A majority of the Justices held that a person’s silence in response to a question asked before any arrest or placement in custody can be used as evidence of guilt at trial.2 A suspect cannot simply remain silent, but must in an unexplained manner (only a plurality discussed the point) explicitly invoke a Fifth Amendment right to remain silent, without ever having been informed of one’s rights following the Court’s landmark ruling in Miranda v Arizona.3
     The Miranda ruling—requiring the police to give a suspect the well-known warnings that “have become part of our national culture,” prior to a custodial interrogation, or risk suppression of any confession statements—had already been badly eroded, although the Court affirmed the constitutional stature of the ruling in Dickerson v United States4 in 2000.5 Over the past four decades, the Court limited Miranda’s reach in a death-by-a-thousand-cuts accretion of rulings. These rulings include: finding a public safety exception,6 an exception for questioning by covert or undercover officers,7 and an exception for questioning during routine booking;8 holding there is no civil recourse for a violation;9 permitting police to renew questioning if the suspect initiates further discussion or after the passage of time;10 requiring that a suspect use precise language to invoke the right;11 permitting police to use the “fruits” of non-Mirandized statements;12 and permitting impeachment use of statements obtained in violation of Miranda.13 Most recently, in its 2010 ruling in Berghuis v Thompkins,14 the Court held that prolonged silence after being given the Miranda warnings was not sufficient to invoke their protection.15
     Of what importance is the new Salinas exception, added to that litany of rulings that sharply narrowed if not “stealth overruled” Miranda?16 Indeed, where the vast majority of suspects readily waive Miranda rights, and where police are trained to provide Miranda warnings in a manner that encourages ready waiver, even perfect compliance with Miranda may not prevent coercion during an interrogation or contamination of false confessions.17 Moreover, the Court has separately done much to erode the limited protections of its highly deferential voluntariness jurisprudence.18
     Even in that badly eroded landscape, the Salinas decision has practical importance and troubling corrosive power. The Court’s rulings have long emphasized “coercion inherent in custodial interrogation” which can “blur[ ] the line between voluntary and involuntary statements.”19 The Salinas decision further encourages police to blur the line between a custodial and noncustodial interrogation. Police interrogate adult suspects somewhat infrequently, chiefly in serious criminal cases like homicides in which confession evidence is critical to closing the case. But police interact and question people on the beat in all sorts of varied everyday contexts. To be sure, the Court had already narrowed the interpretation of the “custody” requirement of Miranda in the circumstances in which a suspect was not under a “formal arrest,” even when questioned at a police station.20 As a result of the Salinas ruling, those routine police interactions can be used to secure admissible evidence if a person does not respond to a solitary question.
     Still more troubling, the Court’s ruling encourages police to question suspects in informal settings that not only lack clear rules, but are not documented and therefore prone to the dangers of confession contamination.21 I will develop why the very type of questioning at issue in Salinas poses special risks of eliciting false information from innocent suspects, and why police and policymakers should continue to vigilantly resist unsound interrogation practices that the Court appears to embrace.

I.  The Salinas Case

     Two brothers were shot and killed at their home in Houston, Texas, on the morning of December 18, 1992.22 Although a neighbor saw a person run out of the house and drive off in a dark-colored car after hearing shots fired, there were no eyewitnesses to the murder itself. The police did find six shotgun shell casings at the scene.23 The police learned that a man named Genovevo Salinas had been at a party at the victims’ house the night before the murders.24 As part of their investigation, police then visited Salinas’ home.25 There they found a dark-colored car parked outside and Salinas himself, who readily spoke to the police and cooperated, including by giving the police his father’s shotgun.26 The police then asked Salinas to accompany them to the station for further questioning, and he agreed. They questioned him at the station for about one hour.27 During this time, the police did not arrest Salinas, and as a result, they did not read him the well-known Miranda warnings.28
     Although Salinas was questioned at a police station, all sides in the litigation apparently agreed that questioning consisted in a noncustodial interview. Was that correct? Was this a noncustodial interrogation? The Court in Miranda defined a custodial interrogation as “questioning initiated by law enforcement officers after a person has been taken into custody or otherwise deprived of his freedom of action in any significant way.”29 However, the Court subsequently limited its definition of custody to a “formal arrest” or something akin to that.30 Further, an “interrogation” requires more than just placing a person in custody, but also some additional compulsion through “express questioning,” or its “functional equivalent” which includes “any words or actions . . . that the police should know are reasonably likely to elicit an incriminating response from the suspect.”31 The Court in Rhode Island v Innis32 had previously found that questioning in a police car using a “few offhand remarks” was not the kind of “lengthy harangue” that counts as an interrogation.33
     The critical moment during the questioning at the police station related to the six shotgun casings found at the crime scene. The police asked Salinas “whether his shotgun ‘would match the shells recovered at the scene of the murder.’”34 Obviously, if he had said “yes,” the answer would have been highly incriminating. However, according to the testimony of the police at his trial, Salinas did not answer the question. He stopped talking.35 The officer described his body language after being asked that question as follows: Salinas “[l]ooked down at the floor, shuffled his feet, bit his bottom lip, cl[e]nched his hands in his lap, [and] began to tighten up.”36
     Salinas had not been told that he had a right to remain silent. But he did remain silent. And after a “few moments of silence,” police changed the subject and asked other questions that Salinas did answer.37 Police then released him. Prosecutors concluded they lacked sufficient evidence to charge him with the murders, and did not file charges until a few days later, after the police obtained a statement from a man who said he overheard Salinas confessing.38 The murder case went to trial years later. (Salinas had gone into hiding.39)
     At his criminal trial, Salinas did not testify. However, the jury heard police and prosecutors describe how he had been silent, accompanied by reportedly uncomfortable body language, after he was asked about whether his shotgun would match the shells recovered. There was no video or record of the police questioning. The officer at trial said, “it’s been a long time ago and there’s a lot of details about this case and many other cases in between that,” making it hard to remember everything that happened.40 And although police, like this officer, can place special emphasis on nonverbal cues by suspects, research shows that such supposed clues do not in fact provide reliable evidence of truthfulness or guilt.41 Nevertheless, jurors may tend to believe that there are telltale signs of a liar. And the prosecutor told the jury, for example, that “‘[a]n innocent person’ would have said, ‘What are you talking about? I didn’t do that. I wasn’t there.’”42 Salinas argued this violated his Fifth Amendment rights.43

II.  The Salinas Decision

     The Salinas decision lacked a majority ruling on what words must be used to invoke one’s Fifth Amendment rights during noncustodial questioning by the police. What is definitive about the Salinas ruling, and what did garner five votes, is that silence during noncustodial questioning may draw an adverse inference at a trial. Justice Samuel Alito wrote the plurality decision, explaining that since Salinas was “free to leave at any time during the interview,” he would have had to verbally assert his right to remain silent.44 While Salinas did remain silent when answering the question at issue, in Justice Alito’s view, he should have “expressly invoke[d]” his Fifth Amendment rights, rather than “standing mute.”45 Justice Alito highlighted that nothing deprived Salinas of the opportunity to voluntarily invoke those rights. Of course, he had not been informed of his rights by the police. However, Justice Alito concluded that “it would have been a simple matter for him to say that he was not answering the officer’s question on Fifth Amendment grounds.”46 It was not clear whether the plurality would require a suspect to invoke the Fifth Amendment itself, or what language would be required to do so short of referring to the specific constitutional right. Although the plurality noted the concern raised that there could be ongoing litigation over “what a suspect must say to invoke the privilege,” the plurality nevertheless declined to provide further guidance.47
     Justice Clarence Thomas and Justice Antonin Scalia wrote separately, and they did not shed any light on how a suspect should invoke Fifth Amendment rights. Instead, they briefly stated their position that Griffin v California48 was wrongly decided, and that the Fifth Amendment does not “prohibit[ ] a prosecutor or judge from commenting on a defendant’s failure to testify.”49
     In Berghuis v Thompkins, the Court had previously addressed a custodial interrogation in which a suspect received the Miranda warnings, and then remained silent during two and three-quarter hours of questioning.50 The Berghuis Court held that remaining silent in that fashion during a custodial interrogation was insufficient to invoke the protections of the Fifth Amendment.51 As Justice Breyer pointed out in his Salinas dissent, that case dealt with the later speech of the suspect, and in that case “[t]he Court said nothing at all about a prosecutor’s right to comment on his preceding silence and no prosecutor sought to do so.”52 Justice Breyer emphasized, “how could a prosecutor lawfully have tried to do so” in the Berghuis case, “given this Court’s statement in Miranda itself that a prosecutor cannot comment on the fact that, after receiving Miranda warnings, the suspect ‘stood mute’?”53
     To be sure, the Court had previously required a suspect to use clear language to invoke the right,54 although Miranda had warned against a rule that “would discriminate against . . . the very defendant who most needs counsel.”55 In contrast, waiver of the right could be inferred from the circumstances (as in Berghuis, in which waiver was inferred from maintaining silence during two and three-quarters hours of police questioning) and did not require any explicit statement by the suspect.56
     Although the Salinas plurality stated that its result flowed from Berghuis,57 as noted, that case involved the situation in which the suspect was in custody, and in which there was at least arguably a waiver of the Miranda right. Salinas involved a suspect who concededly was not in custody and voluntarily accompanied police to the police station. (Whether such a person would really feel free to leave is another question.) But Salinas did not waive any rights, because Salinas was not given them.
     Salinas was also a remarkable extension of the Court’s prior ruling in Jenkins v Anderson58 that pre-arrest conduct could be used for impeachment at trial.59 While the plurality in Salinas purported to rely on Jenkins, that case involved someone who was never questioned by police, who therefore had no occasion to decide whether he was invoking privilege; rather, in that case, the prosecutor commented on the defendant’s failure to turn himself in.60 In contrast, Salinas was questioned at a police station. After Salinas, people must invoke the right to remain silent even when they are not formal suspects, being formally questioned, and when they have not heard the Miranda warnings. This is not remotely realistic. We should not be reassured by the plurality assurance that it would have been a “simple matter” to assert Fifth Amendment rights under such circumstances.61
     In dissent, Justice Stephen Breyer argued, “it was obvious that the new question sought to ferret out whether [Mr.] Salinas was guilty of murder.”62 Moreover, the Salinas rule places a suspect in an “impossible predicament.”63 The suspect must either answer the question and potentially incriminate herself, or remain silent, which may be used negatively at trial.
     Perhaps sophisticated and well-advised suspects might know to assert their Fifth Amendment rights outside of custody. But as Professor Tracey Maclin put it in the context of waiver of Fourth Amendment rights, “very few persons will have the moxie to assert their [ ] rights in the face of police authority.”64 Or as Justice William Brennan put it in his dissent in Oregon v Elstad,65 this is “marble-palace psychoanalysis.”66 For those who are more sophisticated, “after Salinas, potential targets of white collar investigations should be especially sensitive to the risks of cooperating with any government inquiry without counsel.”67 For those less sophisticated, Salinas may similarly make members of the community more fearful of cooperating with the police, lest an uncomfortable look or gesture or silence be interpreted as a guilty gesture or an incriminating silence.

III.  Informal Questioning and False Confessions

     Justice Breyer’s opinion brought up another danger: if the suspect is not silent, but “answers the question, he may well reveal, for example, prejudicial facts, disreputable associates, or suspicious circumstances—even if he is innocent.”68 The Salinas ruling, as Justice Breyer indicates, creates a catch-22 situation, posing special dangers for the innocent suspect. Informal or noncustodial questioning is often not carefully documented by police, just as in the Salinas case itself. As a result, such informal questioning poses special dangers that false confessions may result, even unintentionally, and may prove very difficult to uncover after the fact.
     The danger of confession contamination is particular great during undocumented police questioning. Confession contamination refers to the situation in which police, even inadvertently, disclose key facts during questioning, so that not only may an innocent person falsely confess, but that innocent person may parrot back details that make a false confession seem uncannily corroborated and accurate.69 I do not mean to express here any view on whether or not Salinas was guilty or not of the two murders. However, the cases of people who we do know are innocent, with the benefit of postconviction DNA testing, provide powerful concrete examples that illustrate just how confession contamination can occur and how insidious such contamination can become. I wrote a book examining the first 250 DNA exonerees, and 40 of those innocent people had falsely confessed.70 (There have now been 59 false cases of confessions in the first 300 DNA exonerations.71) Some had supposedly admitted their guilt by making statements to the police before there was any custodial interrogation.
     A case in point is that of Nicholas Yarris, who spent twenty years in prison, after having been convicted and sentenced to death in Pennsylvania. A woman had been found raped, beaten, and stabbed a half mile from her car, a Chrysler Cordoba. Police said that when they initially questioned Yarris, during an undocumented session that was not considered custodial, he had volunteered two key pieces of information. The first was that he knew the victim had been raped. The second was still more detailed: that the victim’s Chrysler had a brown “landau” roof (or a vinyl faux-convertible-type roof).72 The police said that they had kept all such details out of the press, following standard investigative practice. As a detective testified at Yarris’s trial: “This is one of the things we decided to keep confidential in the investigation from the press.”73 As is typical during such a noncustodial interview, there was no recording of the questioning in which those details were allegedly volunteered (although the detectives proceeded to record their subsequent interrogation of Yarris). The detectives did not provide any notes or written statements taken to document that noncustodial interview either. However, those two details provided crucial evidence at trial that Yarris knew the kind of inside information that only the killer could have known.74
     In 2003, DNA tests on several pieces of biological evidence from the crime cleared Yarris, and he was freed and exonerated.75 We now know that Yarris’s confession was likely contaminated, since as an innocent man not privy to information kept confidential during this murder investigation, he could not have known about the two key details.
     Other troubling exonerations involved the same question-first tactics and undocumented interrogations that supposedly produced telling details volunteered by the suspects. The Bruce Godschalk case involved a taped interrogation, but the detective claimed that Godschalk had volunteered a series of nonpublic facts in an initial conversation before the recording was made.76 The Marty Tankleff case, a non-DNA exoneration, presents a slightly different example in which courts found permissible resuming an interrogation after an initial questioning that was non-Mirandized.77 One of the defendants in the well-known Central Park Jogger case claimed never to have confessed at all during initial interviewing that was not videotaped or documented, but a detective claimed that his notes reflected an inculpatory statement.78 Of course, even if Salinas had come out the other way, police might still take advantage of supposedly unsolicited admissions; in a series of troubling cases innocent convicts were also said to have volunteered incriminating information.79
     The crucial reform to prevent contamination of confessions is to ensure that the entire interrogation is videotaped. A clear record of who said what and when during interrogations can help to prevent wrongful convictions. But rules requiring detectives to produce such a record of interrogations are subverted if police who are required to videotape custodial interrogations are encouraged to question first in noncustodial and undocumented settings.


     Salinas is like the Terry v Ohio80 of the law of confessions, giving police such broad discretion in their less-formal work that criminal procedure protections ostensibly applicable during more formal actions become all the more irrelevant.81 One purpose of the Miranda ruling was “to give concrete constitutional guidelines for law enforcement agencies and courts to follow.”82 The Salinas ruling does the opposite, like the litany of rulings dating back to the 1970s that have gradually whittled away at Miranda. Perhaps scholars and amici could have better alerted the Justices to the implications of this ruling. The media did not react with any particular degree of concern to the Salinas ruling.83 Then again, perhaps scholars and amici assumed there was little hope that the Court would do anything to breathe life into what remains of Miranda. In the years ahead, however, Salinas may play a particularly troubling role in interactions between police and citizens, and not because of the way it weakened Miranda, but rather because of the way the Court tacitly encouraged police to question first outside of custody.
     Police may have broad power to stop and frisk citizens under Terry, but they may also restrain use of that discretion for fear of community outrage, including over racial profiling. Similar restraint may come into play in the context of police questioning, if communities increasingly fear cooperating with the police. Then again, people may assume that those who speak to police and say something inculpatory are guilty. False confessions come to light years later only in rare cases where postconviction evidence of innocence, like DNA tests, can be obtained. But are those who remain uncomfortably silent during informal questioning necessarily guilty?
     The Salinas ruling may have a range of consequences that undermine criminal investigations as well as the rights of the accused. After the Supreme Court’s decision in Salinas, lawyers may counsel suspects not to speak to the police at all. In contrast, for the unsophisticated suspect, any conversation with the police may be a highly fraught encounter. Returning to the catch-22 that the Salinas decision creates, an “impossible predicament,”84 as Justice Breyer put it, suppose Salinas had answered the question posed about the shotgun casings by exclaiming, “I tell you, I am innocent!” Could police testify at trial that his exclamation rang false or was unconvincing? Suppose Salinas had simply answered the question, “No, they couldn’t match my father’s shotgun.” Could police nevertheless assert that he paused or fidgeted or looked uncomfortable while making an exculpatory comment? As noted, although police often place particular emphasis on nonverbal cues when questioning suspects, social science research shows that such nonverbal clues are not diagnostic of truthfulness or guilt.85 Moreover, in the Salinas case, there was no evidence like a video that documented that Salinas in fact gave some appearance of being uncomfortable when asked about the shotgun.
     The Fifth Amendment protections against self-incrimination are obviously most critical when police interrogate suspects, given the inherent coercive power of the police when using modern psychological questioning tactics. Police conduct interrogations in the most serious and high-profile criminal cases, often in homicides, in which there are no eyewitnesses or highly probative forensic evidence. The pressure to secure a confession in such cases can be intense. However, police also question suspects far more routinely without conducting formal interrogations. As in Terry, the Justices may have been uncomfortable limiting police discretion to questioning informally, particularly in the range of situations that may arise outside the stationhouse.
     However, it is in precisely that kind of undocumented questioning that we should worry about the problem of contamination. As a result of the Supreme Court’s tolerance of a questions-first, rights-later approach, police have more incentives to informally question suspects with an eye to a confession. The result encourages police to question suspects without the protections that more and more departments have adopted precisely to prevent false and contaminated confessions.  More departments now videotape entire interrogations, an inexpensive reform that both carefully documents sound, professionally conducted interrogations, and can help to identify situations in which a confession was contaminated. Without a complete record of an entire interrogation, the confession cannot be easily evaluated for its reliability. Unless it is a rare case in which DNA testing can be done, or other powerful independent evidence is uncovered, there may be no way to know whether the confession was true or false.
     Hopefully, the mounting numbers of states and local jurisdictions that have already reformed their interrogation practices, despite decades of steadily weakened Miranda and voluntariness jurisprudence, will continue to improve the way they take confessions.86 The Supreme Court’s stealth overruling of Miranda may render the Court’s thinning jurisprudence increasingly irrelevant. A bare majority of the Justices may be indifferent to the problem of false confessions, but criminal justice actors on the ground do not have that luxury. They have increasingly learned the hard way that wrongful convictions result from tolerating unreliable interrogation practices. While the Supreme Court’s Fifth Amendment jurisprudence continues to erode, the real safeguards of the integrity of confession evidence are being built up on the ground.

          Roy L. and Rosamond Woodruff Morgan Professor of Law, University of Virginia School of Law. For invaluable comments on an earlier draft, I thank Yale Kamisar and Saul Kassin. I previously wrote about the Salinas decision in an op-ed in Slate, for which I thank Emily Bazelon for helpful edits and comments. Brandon L. Garrett, You Don’t Have the Right to Remain Silent, Slate (June 19, 2013), online at (visited Sept 22, 2013).

      1     133 S Ct 2174 (2013).

      2     Id at 2180 (plurality), 2184 (Thomas concurring).

      3     384 US 436 (1966); Salinas, 133 S Ct at 2180.

      4     530 US 428 (2000).

      5     Id at 443. For scholarship discussing the “erosion” of Miranda, see, for example, Leslie A. Lunney, The Erosion of Miranda: Stare Decisis Consequences, 48 Cath U L Rev 727, 745–86 (1999); Barry Friedman, The Wages of Stealth Overruling (with Particular Attention to Miranda v. Arizona), 99 Georgetown L J 1, 24 (2010) (“Miranda has effectively been overruled.”); Charles D. Weisselberg, Mourning Miranda, 96 Cal L Rev 1519, 1521 (2008) (“Miranda is largely dead. It is time to ‘pronounce the body,’ as they say on television, and move on.”); Yale Kamisar, The Rise, Decline, and Fall (?) of Miranda, 87 Wash L Rev 965, 984 (2012) (identifying the “piece-by-piece ‘overruling’ of Miranda”).

      6     See New York v Quarles, 467 US 649, 653 (1984).

      7     See Illinois v Perkins, 496 US 292, 294 (1990).

      8     See Pennsylvania v Muniz, 496 US 582, 601–02 (1990).

      9     See Chavez v Martinez, 538 US 760, 766–73 (2003).

      10    See Edwards v Arizona, 451 US 477, 485 (1981); Michigan v Mosley, 423 US 96, 107 (1975); Oregon v Bradshaw, 462 US 1039, 1043–44 (1983); Oregon v Elstad, 470 US 298, 310 (1985); Missouri v Seibert, 542 US 600, 611 (2004) (plurality).

      11    See Connecticut v Barrett, 479 US 523, 527–30 (1987).

      12    United States v Patane, 542 US 630, 636–37 (2004).

      13    Harris v New York, 401 US 222, 226 (1971); Oregon v Hass, 420 US 714, 722–23 (1975).

      14    130 S Ct 2250 (2010).

      15    See id at 2259–60.

      16    Friedman, 99 Georgetown L J at 16 (cited in note 5).

      17    See Saul M. Kassin, et al, Police-Induced Confessions: Risk Factors and Recommendations, 34 L & Human Beh 3, 23 (2010) (“[I]t is clear that Miranda warnings may not adequately protect the citizens who need it most—those accused of crimes they did not commit.”); Richard J. Ofshe and Richard A. Leo, The Decision to Confess Falsely: Rational Choice and Irrational Action, 74 Denver U L Rev 979, 1119 (1997). Innocent people may be particularly likely to waive Miranda rights. See Saul M. Kassin, On the Psychology of Confessions: Does Innocence Put Innocents at Risk?, 60 Am Psych 215, 223–24 (2005).

      18    See Colorado v Connelly, 479 US 157, 168 (1986). See also Brandon L. Garrett, The Substance of False Confessions, 62 Stan L Rev 1051, 1109–12 (2010).

      19    Dickerson, 530 US at 435.

      20    See Beckwith v United States, 425 US 341, 347 (1976); Oregon v Mathiason, 429 US 492, 495 (1977); California v Beheler, 463 US 1121, 1124–25 (1983).

      21    Brandon L. Garrett, Convicting the Innocent: Where Criminal Prosecutions Go Wrong 14–44 (Harvard 2011).

      22    Salinas, 133 S Ct at 2178.

      23    Id.

      24    Id.

      25    Id.

      26    Salinas, 133 S Ct at 2178.

      27    Id.

      28    Id.

      29    Miranda, 384 US at 444.

      30    Beheler, 463 US at 1125 (limiting the definition of “custody” to “‘formal arrest or restraint on freedom of movement’ of the degree associated with a formal arrest”), quoting Mathiason, 429 US at 495.

      31    Rhode Island v Innis, 446 US 291, 301 (1980).

      32    446 US 291 (1980).

      33    Id at 303.

      34    Salinas, 133 S Ct at 2178, quoting Joint Appendix, Salinas v Texas, No 12-246, *17 (US filed Feb 20, 2013) (available on Westlaw at 2013 WL 634511) (“Joint Appendix”).

      35    Salinas, 133 S Ct at 2178.

      36    Id (brackets in original), quoting Joint Appendix at *18.

      37    Salinas, 133 S Ct at 2178.

      38    Id.

      39    Id.

      40    Joint Appendix at *10.

      41    See Saul M. Kassin, Human Judges of Truth, Deception, and Credibility: Confident but Erroneous, 23 Cardozo L Rev 809, 809–17 (2002); Aldert Vrij, Detecting Lies and Deceit: Pitfalls and Opportunities 142–51 (Wiley 2d ed 2008).

      42    Salinas, 133 S Ct at 2185 (Breyer dissenting) (brackets in original), quoting Salinas v Texas, 368 SW3d 550, 556 (Tex App 2011).

      43    Salinas 133 S Ct at 2178 (majority).

      44    Id at 2180, quoting Brief for Petitioner, Salinas v Texas, No 12-246, *2–3 (US filed Jan 20, 2013) (available on Westlaw at 2013 WL 633595).

      45    Salinas, 133 S Ct at 2179, 2181.

      46    Id at 2180.

      47    Id at 2183.

      48    380 US 609 (1965).

      49    Salinas, 133 S Ct at 2184–85 (Thomas concurring).

      50    Berghuis, 130 S Ct at 2259.

      51    See id at 2259–60.

      52    Salinas, 133 S Ct at 2189 (Breyer dissenting).

      53    Id (Breyer dissenting), quoting Miranda, 384 US at 468 n 37.

      54    Connecticut v Barrett, 479 US 523, 527–30 (1987).

      55    Miranda, 384 US at 471, quoting People v Dorado, 398 P2d 361, 369–70 (Cal 1965).

      56    Berghuis, 130 S Ct at 2262.

      57    Salinas, 133 S Ct at 2182.

      58    447 US 231 (1980).

      59    Id at 240.

      60    Id at 233–34. Justice Breyer in his dissent describes this well, as well as why other cases relied upon by the Salinas plurality were inapposite. See Salinas, 133 S Ct at 2187–88 (Breyer dissenting).

      61    Salinas, 133 S Ct at 2180 (majority).

      62    Id at 2189 (Breyer dissenting).

      63    Id at 2186 (Breyer dissenting).

      64    Tracey Maclin, The Decline of the Right of Locomotion: The Fourth Amendment on the Streets, 75 Cornell L Rev 1258, 1306 (1990).

      65    470 US 298 (1985).

      66    Id at 324 (Brennan dissenting).

      67    Kirk Ogrosky, Murad Hussain, and Charles B. Weinograd, Silence As Evidence: U.S. Supreme Court Holds That the Fifth Amendment Does Not Bar Using a Suspect’s Silence as Evidence of Guilt *4 (Arnold & Porter LLP June 21, 2013), online at (visited Sept 22, 2013).

      68    Salinas, 133 S Ct at 2186. See also United States v Hale, 422 US 171, 181 (1975) (Burger concurring) (“It is no more accurate than to say, for example, that the innocent, rather than the guilty, are the first to protest their innocence. There is simply no basis for declaring a generalized probability one way or the other.”).

      69    Laura H. Nirider, Joshua A. Tepfer, and Steven A. Drizin, Book Review, Combating Contamination in Confession Cases, 79 U Chi L Rev 837, 847 (2012).

      70    Garrett, Convicting the Innocent at 18–19 (cited in note 21).

      71    This data is on file with the author, and is in the process of being updated with detailed information about each false confession. Perhaps it should be surprising that there would be so many false confessions, nineteen of the fifty cases, in the more recent DNA exonerations, but the increasingly confession-heavy makeup of DNA exonerations may be explained by a series of cases in which exonerations were substantially delayed where prosecutors were highly reluctant to concede innocence in the face of DNA tests, due to the fact that the defendant had confessed. See, for example, Andrew Martin, The Prosecution’s Case Against DNA, NY Times Magazine 44, 46 (Nov 27, 2011) (describing how the three-page confession statement of Juan Rivera led to his conviction at three different trials, including at a third trial that occurred after DNA tests had excluded him).

      72    See Garrett, 62 Stan L Rev at 1078 (cited in note 18); Brandon L. Garrett, You Don’t Have the Right to Remain Silent, Slate (June 19, 2013), online at (visited Sept 22, 2013); Pennsylvania v Yarris, 549 A2d 513, 520 (Pa 1988).

      73    Trial Transcript, Commonwealth v Yarris, No 690-82 (Pa Ct Common Pleas June 29, 1982) (on file with author).

      74    For the trial testimony concerning the confession, see Brandon Garrett, False Confessions: Transcripts and Testimony, Library (University of Virginia School of Law), online at (visited Oct 9, 2013).

      75    See Nicholas Yarris, The National Registry of Exonerations (University of Michigan Law School), online at
casedetail.aspx?caseid=3771 (visited Oct 9, 2013).

      76    Garrett, 62 Stan L Rev at 1081 (cited in note 18).

      77    Tankleff v Senkowski, 135 F3d 235, 242–46 (2d Cir 1998).

      78    Garrett, 62 Stan L Rev at 1084 (cited in note 18).

      79    Id at 1106–07.

      80    392 US 1 (1968).

      81    See id at 20–22. For criticism of the Court’s balancing approach in Terry, see, for example, Carol S. Steiker, Second Thoughts about First Principles, 107 Harv L Rev 820, 855 (1994); Scott E. Sundby, A Return to Fourth Amendment Basics: Undoing the Mischief of Camara and Terry, 72 Minn L Rev 383, 401–04, 418–20 (1988).

      82    Arizona v Roberson, 486 US 675, 680 (1988), quoting Miranda, 384 US at 441–42.

      83    See Garrett, You Don’t Have the Right to Remain Silent (cited in note 72).

      84    Salinas, 133 S Ct at 2186 (Breyer dissenting).

      85    See Kassin, 23 Cardozo L Rev at 809–17 (cited in note 41); Aldert Vrij, Detecting Lies and Deceit at 142–51 (cited in note 41).

      86    See Garrett, Convicting the Innocent at 247–48 (cited in note 21) (describing state and local interrogation reforms).