Confession and Confrontation

The constitutional law of confessions has a critical blind spot. In theory, the law serves two interests. First, it protects the autonomy of suspects by stipulating that they can be questioned while in custody only with their consent. Second, it restrains official misconduct by forbidding interrogation methods that overbear a suspect’s will. Even if the law adequately safeguards those interests, something is missing: reliability. As false confessions emerge as a major source of wrongful convictions and as social scientists expose how standard interrogation tactics prompt innocent people to confess, the Supreme Court and conventional wisdom insist that the reliability of confessions is not a constitutional concern.

The Supreme Court and conventional wisdom are wrong. Inattention to reliability is a jurisprudential oversight, not a feature of constitutional design. The problem is that the Court and commentators have neglected the part of the Constitution that unabashedly curates prosecutorial evidence: the Confrontation Clause. For much of our constitutional history, that omission could be excused, as the Confrontation Clause was a sleepy corner of the Sixth Amendment. Twenty years ago, in Crawford v. Washington, the Supreme Court enlivened it.

Table of Contents Show

Introduction

The constitutional law of confessions is broken. False confessions, once written off as “scarcely conceivable,”[1] are now impossible to ignore.[2] In the past few decades, courts have exonerated scores of people who falsely confessed to crimes following manipulative interrogations.[3] Meanwhile, psychological experiments have revealed that it takes surprisingly little to coax innocent people to “admit” wrongdoing. The tactics that American law enforcement officers routinely use, it turns out, are enough.[4]

And yet, our constitutional law does little about false confessions. Courts regulate confessions pursuant to the Self-Incrimination Clause of the Fifth Amendment (the source of Miranda warnings) and the Due Process Clauses of the Fifth and Fourteenth Amendments (which require that confessions be “voluntary”).[5] But generations of scholars have identified weaknesses in the Miranda and voluntariness doctrines. Miranda’s regulatory potential was undone by its capacious waiver rule;[6] voluntariness, meanwhile, is only tangentially concerned with the reliability of confessions.[7] Voluntariness may vindicate other values, but it does not resist unreliable confessions induced by psychologically manipulative interrogation tactics.

To fix our confessions law, we should look beyond Miranda and voluntariness. The Constitution has a provision addressed to the reliability of prosecutorial evidence: the Sixth Amendment’s Confrontation Clause.[8] Courts and commentators have overlooked the Confrontation Clause as a source of constitutional regulation of confessions. This omission was defensible prior to 2004, when the Court decided Crawford v. Washington.[9] It no longer is.

This Article shows that on a novel but straightforward reading of Crawford, the Confrontation Clause could be a bulwark against manipulative interrogation tactics that yield false confessions. Here’s the short version of the argument. The Confrontation Clause guarantees defendants the right to “be confronted with” (i.e., to cross-examine) the “witnesses against” them.[10] Crawford revitalized the confrontation doctrine by holding that the “witnesses against” a defendant include anyone who made a “testimonial” out-of-court statement that the government seeks to use at trial, a category that includes people who accused the defendant during a formal police interview.[11] Those “witnesses” must testify at trial.[12] If death, privilege, or absence makes them unavailable, then, with limited exceptions, their unconfronted out-of-court accusations are inadmissible.[13]

Once we understand that a person can become a Sixth Amendment “witness” against a defendant on account of what they said in a police interrogation room, there is no good reason why the defendant themself cannot be a witness, too.[14] After all, when the government uses a defendant’s confession at trial, the defendant becomes not just a witness against themself, but (ordinarily) the prosecution’s best witness.[15] That’s where things get interesting. In their capacity as a prosecution witness, a defendant’s testimonial hearsay should be admissible only if they testify. By testifying, they’d “confront” their own out-of-court accusation just as they’d confront the out-of-court accusation of any witness who recanted before trial. They would try to persuade the jury that what they said was coerced, taken out of context, or otherwise not true.

There is one major difference between defendants who confessed and other out-of-court accusers. The Confrontation Clause requires that prosecutors procure the live testimony of Sixth Amendment witnesses.[16] But defendants have a right not to testify,[17] and they cannot be forced to choose between their right to confrontation and their privilege against self-incrimination.[18] That means that the government cannot procure a defendant’s testimony without their consent. So, if a defendant chooses not to testify, their testimonial confession will remain unconfronted and thus (unless an exception applies) inadmissible.

The final piece of the puzzle concerns whether the Confrontation Clause is amenable to an exception that would make a defendant’s confession admissible without confrontation. The Court uses a historical approach to identify exceptions to the confrontation right.[19] The Sixth Amendment allows testimonial hearsay into evidence without confrontation, the Court says, only if the hearsay would have been admissible without confrontation in common law courts circa 1791.[20] Crawford observed that at common law, a “suspect’s confession” could be offered “against himself,”[21] so the door to an exception is already half open. The Court’s historical point, however, is subject to an important caveat. Under a rule well known to legal historians but long forgotten by American courts, confessions were admissible at common law (as of 1791) only if they were not “forced from the mind by the flattery of hope, or by the torture of fear.”[22] Only confessions obtained from suspects without “hope” or “fear,” then, can be carved out of the confrontation right.[23]

Putting it all together, we come to a relatively simple rule: Testimonial confessions induced by hope or fear should be inadmissible unless the defendant testifies. Precisely which modern interrogation methods implicate hope or fear will have to be worked out, but tactics that minimize a crime’s moral or legal gravity—which are legion in police interrogation manuals—surely count.[24] Confessions unblemished by hope or fear, meanwhile, should be free from Confrontation Clause challenge.[25]

The full case unfolds in three parts. Part I makes the doctrinal argument, just previewed, for regulating confessions under the Confrontation Clause. Part II then responds to four categories of likely objections.

First, an objector might contend that Crawford was concerned only with hearsay, and confessions are not hearsay by virtue of Federal Rule of Evidence 801(d)(2)(A).[26] Several courts have adopted this view, but it is unserious. That the drafters of the Federal Rules of Evidence decided to label party-opponent statements “exempt” from the hearsay rule, rather than admit them as hearsay “exceptions,” is a distinction without constitutional significance.[27]

Second, an objector might maintain that confessions are already adequately governed by Miranda and the voluntariness requirement, such that there is no need to saddle them with further scrutiny under the Sixth Amendment. This is not so, for reasons I have already alluded to and will elaborate below.[28]

Third, an objector might argue that a defendant’s opportunity to confront their confession, by testifying in their own case, suffices. It does not, because the Confrontation Clause imposes a duty on prosecutors, not defendants, to secure the trial testimony of government witnesses.[29]

Fourth, originalist objectors might argue that (1) the framers of the Confrontation Clause would never have thought to condition the admissibility of a confession on a defendant testifying, and/or (2) the original public meaning of the confrontation right didn’t include defendants’ confessions.[30] It’s true that the admissibility of confessions at common law didn’t hinge on whether defendants testified. Indeed, defendants were disqualified from testifying.[31] But this Article’s proposal requires confrontation of confessions that were categorically inadmissible at common law, so historical antecedents take us only so far. There was, moreover, no ascertainable “confrontation right” in 1791. Instead, there were loosely connected common law rules and practices to which we attach the “confrontation” label. Whether the label covers defendants’ confessions can’t be resolved by historical inquiry alone.

Part III steps back from the main argument to consider connections between Crawford and the Court’s even more famous decision in Miranda v. Arizona.[32] The Court’s goal in Miranda was to curb manipulative tactics in custodial interrogation.[33] It failed, in large part because Miranda and its progeny permit interrogators to use their psychological tools so long as they first advise a suspect of their rights and persuade them to waive those rights. This “warn-and-waive” protocol was a consequence of the Court’s reliance on the Self-Incrimination Clause.[34] A ruling based on the Confrontation Clause would not have been so vulnerable because the confrontation right attaches to trials, not investigations.[35] In the interrogation room, suspects can no more waive their right to confrontation at trial than they can waive their rights to a unanimous jury or compulsory process. It would be unfair to fault Miranda for not anticipating Crawford. But knowing what we know now—about the Confrontation Clause and false confessions—we could repair our constitutional law of confessions.

The Conclusion considers one last objection—that the proposal would obstruct law enforcement by depriving police of interrogation tools that sometimes produce true confessions. That is a possible outcome, but not an inevitable one. Non-manipulative interrogation methods exist and are widely used outside the United States. This Article’s proposal would, if adopted, be a powerful inducement for American law enforcement officials to innovate their interrogation practices. It could bring about an interrogation regime that seeks true confessions and eschews false ones.

Before we start, a word about methodology. Except in Part III, this Article treats the decisions of the Supreme Court as fixed points. In other words, I take the cases (or at least their holdings) as I find them.[36] As every law student knows, Supreme Court rulings are not actually fixed points.[37] Treating them as if they were is thus artificial, but for a good reason. It allows me to show that mending the constitutional law of confessions doesn’t require the strong medicine of overruling Supreme Court decisions. Rather, it can be done merely by extending the logic of cases already on the Court’s books. No precedents were harmed in the making of this argument.

I. Why (and When) the Confrontation Clause Applies to Confessions Offered Against Their Makers

Crawford’s framework for analyzing the Confrontation Clause is familiar ground.[38] Because my argument depends on its details, I must nonetheless trace it. Readers already versed in Crawford may wish to skip to Part I.A.

The Confrontation Clause guarantees criminal defendants the right to cross-examine the “witnesses against” them.[39] Crawford is the current answer to an enduring question: Does a person become a “witness against” a defendant when the government uses that person’s hearsay at trial?[40] They do, Crawford decided, if the hearsay was “testimonial.”[41] So, unless they testify and submit to cross-examination, their hearsay is inadmissible.[42]

Justice Scalia, writing for the Court, reached this conclusion via a two-step process.[43] He first identified the principal historical evil that the Confrontation Clause was meant to remedy: the prosecutorial use of out-of-court statements as a substitute for live testimony.[44] That evil was routine, Scalia explained, in the English courts of the sixteenth and seventeenth centuries, not only in high-profile trials like Walter Raleigh’s, but also in everyday cases where judges allowed unconfronted examinations conducted by justices of the peace to be read to the jury.[45] But by the time the Sixth Amendment was framed and ratified, common law courts had seen the error of their ways. Judges allowed examinations in evidence only if the examinee was unavailable to testify at trial and the defendant had a prior opportunity to cross-examine them.[46]

Having found (or purported to find) the Confrontation Clause’s historical roots, Justice Scalia turned to the second step in the process: updating the confrontation right for the modern world.[47] There are, after all, few roving justices of the peace examining suspects and witnesses these days. Not all hearsay, Scalia reasoned, implicates the Confrontation Clause.[48] He drew a line—albeit a fuzzy one—between “testimonial” and “nontestimonial” hearsay.[49] The Confrontation Clause does not reach nontestimonial hearsay, so states and the federal courts may regulate it as they see fit.[50] Testimonial hearsay, on the other hand, triggers confrontation.[51] Its makers become “witnesses against” the defendant when the government seeks to use their statements at trial.[52] The government must therefore produce them to testify.[53] If it cannot because the declarant is unavailable, the government can introduce the hearsay if the defendant had a prior opportunity to cross-examine the declarant.[54] Failing that, the hearsay is inadmissible, subject to very limited exceptions. Crawford identified only two explicitly: (1) forfeiture by wrongdoing[55] and (2) possibly dying declarations, though the Court declined to rule definitively on it.[56] The Court also noted that “a suspect’s confession could be admitted . . . against himself,” which was, we will see, in effect a third exception.[57]

Obviously, a lot turns on what it means for hearsay to be (or not to be) testimonial. Crawford did not offer a “comprehensive” definition,[58] and the issue has proven thorny across several domains.[59] But one type of statement is almost always testimonial: statements to a police officer during a formal interrogation.[60] Those, Justice Scalia noted, “bear a striking resemblance to examinations by justices of the peace in England.”[61]

Before Crawford, the Court had tethered the confrontation right to modern hearsay law, including many of its exceptions.[62] If a statement was admissible under a hearsay exception, it was a good bet that the Confrontation Clause wouldn’t preclude its admission either.[63] Crawford recognized that by detaching confrontation from hearsay law, it was depriving juries of evidence bearing a “judicial determination of reliability.”[64] The Court saw this as a feature of its new doctrine.[65] The Sixth Amendment does not “command” that “evidence be reliable,” the Court explained, “but that reliability be assessed in a particular manner: by testing in the crucible of cross-examination.”[66] And the Sixth Amendment insists on that mechanism even when a statement’s reliability is beyond reproach. “Dispensing with confrontation because testimony is obviously reliable,” Justice Scalia quipped, “is akin to dispensing with jury trial because a defendant is obviously guilty.”[67] “That is not what the Sixth Amendment prescribes,” he added for good measure.[68]

Having surveyed Crawford’s approach to confrontation, we can now get to work. This Part analyzes Crawford’s implications for confessions offered against the defendants who made them.[69] The analysis involves three distinct questions: (1) Can a defendant be a “witness against” themself, to whom the Confrontation Clause applies? (2) If so, can a defendant simultaneously invoke both the Fifth Amendment right not to testify and the Sixth Amendment confrontation right? (3) If so, does Crawford permit a “confession exception” to the Confrontation Clause? To preview: (1) yes; (2) yes; and (3) yes, but it’s limited.

A.      Defendants Can Be “Witnesses Against” Themselves

The first question is whether making a confession can transform a criminal defendant into a “witness against” themself.[70] Crawford’s framework makes this one easy.

As we just saw, when a prosecutor seeks to use testimonial hearsay in a criminal trial, the declarant becomes a “witness against” the defendant.[71] A defendant therefore becomes a “witness against” themself the moment the government seeks to use their testimonial confession at trial.[72] From that point forward, they are not just a defendant; they are also a government witness. For purposes of the Confrontation Clause, they are little different from any other out-of-court accuser who recanted before trial. A person need not be unfriendly to a defendant to be a “witness against” them. Rather, they just need to have created out-of-court “testimony” that the prosecutor wants to use.

Still, one might wonder, what does it mean for a defendant to confront themself? Is “self-confrontation” even coherent? Simply put, a defendant confronts themself by telling the jury that it should not believe their out-of-court statement.[73] Depending on the case, that could mean explaining that they were confused or beleaguered during the police interrogation or that they said something they did not mean. The defendant is the only witness who can give the jury a first-hand explanation of how the interrogation affected them.[74] While the jury may be skeptical of their explanation,[75] futility is not a reason to bypass confrontation.[76]

Recognizing that defendants can be Confrontation Clause witnesses is consistent with (even dictated by) Crawford, and it makes sense conceptually. It has the additional benefit of reconciling the term “witness” in the Confrontation Clause with the same term in other corners of constitutional criminal procedure. That a defendant can be understood as a “witness against himself” is palpable in the Fifth Amendment’s Self-Incrimination Clause.[77] Closer to the Confrontation Clause’s home, the Sixth Amendment’s Compulsory Process Clause empowers defendants to secure the testimony of “witnesses in his favor.”[78] One of those witnesses, the Supreme Court tells us, is the defendant.[79] Confrontation and compulsory process are “conceptual twins.”[80] If a defendant who provides exculpatory testimony is a “witness in his [own] favor,” a defendant who confesses can be a “witness against him[self].”[81]

B.      Defendants Cannot Be Compelled to Choose Between Confrontation and the Privilege Against Self-Incrimination

Enough big picture. How would self-confrontation work as a matter of trial practice? Imagine that during the government’s case-in-chief, the prosecutor calls a police officer to testify that Defendant Daniels confessed during a custodial interview at the police station. Daniels’s lawyer objects. The Confrontation Clause, their lawyer explains, requires the government to produce testimonial declarants for live testimony,[82] and it hasn’t produced Daniels. “That’s fine,” the prosecutor replies, “the state calls Daniels.” Trial can thereafter proceed in three ways:

Scenario 1: Daniels consents to testify during the government’s case-in-chief and their lawyer cross-examines them. Because the hearsay declarant has testified and was subject to cross-examination, the Confrontation Clause is satisfied. (It goes without saying that few, if any, defendants would choose this scenario.)

Scenario 2: Daniels invokes their Fifth Amendment right not to testify during the government’s case but later testifies during their own. Because Daniels has testified, the prosecutor can use the confession either in cross-examining Daniels or during the government’s rebuttal.[83]

Scenario 3: Daniels invokes their Fifth Amendment right not to testify during the prosecutor’s case and does not testify in their own. Unless Daniels has thereby forfeited their confrontation or self-incrimination rights (the topic of this Section) or an exception applies (discussed in the next Section), their confession is inadmissible. Note, however, that only the confession is excluded. Evidence derived from it is not the Confrontation Clause’s concern.[84]

Of these scenarios, only the third requires our close attention. In it, a defendant makes two moves almost simultaneously: (1) they deny the prosecutor the opportunity to use their testimonial hearsay because the prosecutor hasn’t produced its declarant (i.e., Daniels) to testify, and (2) they block the prosecutor from producing the declarant. But can the defendant assert the confrontation right and the privilege against self-incrimination with regard to the same evidence? Or is the confession’s admissibility the price of silence?

If my argument thus far holds up—that is, if a defendant who made a testimonial confession is a “witness” for Confrontation Clause purposes—those are the critical next questions. Conceptually, courts (or even legislatures) could try to compel a defendant to choose between confrontation and self-incrimination in two different ways. First, they might deem defendants who refused to testify during the prosecutor’s case to have thereby forfeited the right to confront their confessions. Second, they might deem defendants who asserted a Confrontation Clause objection to have forfeited the privilege against self-incrimination.[85]

Below, I analyze these possibilities under the applicable case law. An alternative approach would be to analyze them from first principles as a “double constitutional conditions” problem.[86] Doing so would add complexity, but not much light. The reason is that a proper double conditions analysis involves taking account of the interests served by each constitutional provision.[87] Although the Confrontation Clause has relatively well-understood purposes,[88] the Self-Incrimination Clause does not.[89] Indeed, the prevailing, though not unanimous, view in the legal and philosophical academies is that the privilege against self-incrimination lacks a sound theoretical rationale.[90] Against that backdrop, constitutional conditions analysis will not take us very far. So, I’ll stick with the lawyer’s traditional fallback: precedent.

1.       Forfeiting Confrontation

First, could a court deem a defendant who refuses to testify in the prosecution’s case to have forfeited (or, for that matter, waived) an objection to the government’s use of their testimonial confession? Under current law, no.

The Supreme Court has a well-defined doctrine for evaluating when confrontation rights are “extinguished” on “equitable grounds.”[91] In Giles v. California, the Court held that under the doctrine of “forfeiture by wrongdoing,” a defendant loses the right to confront a witness by “engag[ing] in conduct designed to prevent the witness from testifying.”[92] The Court explained that the rule is “aimed at removing the otherwise powerful incentive for defendants to intimidate, bribe, and kill the witnesses against them.”[93] Giles had no need to delineate what sort of conduct—beyond intimidation, bribery, and murder—qualifies for forfeiture. But lower courts both before and after Giles confirm that, as the doctrine’s name indicates, a “wrongful act on the part of the defendant” is required.[94]

When defendants invoke the Fifth Amendment right not to testify, they intentionally prevent a witness (i.e., themselves) from testifying. The question is whether doing so is “wrongful.” Perhaps, as Justice Scalia argued in a dissenting opinion, “it is more virtuous for [a] wrongdoer to admit his offense” than to “stand silent.”[95] But that hardly makes exercising a constitutional right wrongful. Indeed, even deploying a subconstitutional privilege that keeps a witness from testifying isn’t a wrongful act.[96] In Crawford itself, the defendant (Michael Crawford) was able to invoke his confrontation right even though he’d prevented the declarant (Sylvia Crawford) from testifying by asserting spousal privilege.[97] Likewise, while a defendant’s refusal to testify might spoil a prosecutor’s case, it’s not wrongdoing. As such, it cannot be the basis for stripping a defendant of their Sixth Amendment rights.

2.       Forfeiting Self-Incrimination

Let’s go, then, to the inverse question. Could a court decide that a defendant forfeits (or waives) the right not to testify during the prosecution’s case by asserting a Sixth Amendment objection against their confession? Again, no, not consistently with Supreme Court precedent.

Since the 1950s, the Supreme Court has decided several compelled election cases involving the right not to testify, yielding six significant data points. Defendants can be forced, the Court has ruled, to choose between the right to remain silent at trial and (1) the right to testify at trial,[98] (2) the right to keep prior conviction evidence out of the jury’s hands,[99] and (3) (of historical interest only) the right to testify about punishment at a unified death penalty trial.[100] Meanwhile, they cannot be made to choose between the right to remain silent and (1) the right to testify in support of a pre-trial Fourth Amendment motion to suppress,[101] (2) the right to plead guilty,[102] or (3) the right to testify after other defense witnesses.[103]

Fortunately, the cases have a through line: They turn on whether compelling the election was necessary to protect the adversarial process of trial. It was in the first set of cases but not in the second. This line is most apparent in the juxtaposition of Brown v. United States and Mitchell v. United States. In Brown, the Court held that a defendant may not testify on their own behalf and then refuse to answer the prosecutor’s questions on cross.[104] Allowing that, the Court explained, would make the Fifth Amendment “a positive invitation to mutilate the truth.”[105] In Mitchell, the defendant admitted at a plea colloquy that she did “some” of what she’d been accused of, but later, at sentencing, tried to invoke her privilege to remain silent.[106] Had she testified at trial, she would have had no privilege to invoke at sentencing. The government argued that her statement at the plea colloquy had the same effect.[107] Disagreeing, the Court explained that Mitchell’s silence at sentencing, following her admissions at the colloquy, “did not pose a threat to the integrity of factfinding proceedings.”[108]

Our question is now clarified. Does a defendant threaten the integrity of the adversarial process of trial by refusing to testify after objecting to the use of their unconfronted hearsay? So is the answer: not at all. They have presented the jury with no “mutilated” information. Rather, the jury simply never learns that they confessed. While the jury collects less information on which to base its verdict, that’s par for exclusionary rules.[109] The government has to find another way to prove its case, but the adversarial process is safe.

C.      Crawford is Amenable to a Limited “Confession Exception”

So far we have seen, first, that a defendant can be a “witness[] against” themself under Crawford’s definition; and, second, that a defendant cannot be made to choose between their Fifth Amendment privilege to remain silent at trial and their Sixth Amendment right to confront the government’s witnesses. The remaining question is whether the Confrontation Clause is amenable to an “exception” that would allow a defendant’s confession to be admitted without confrontation.

Crawford did not announce a total ban on prosecutorial use of unconfronted testimonial hearsay. Rather, the Court recognized that the “Sixth Amendment incorporates . . . limitations” if they are grounded in history.[110] If unconfronted testimonial hearsay was admissible at common law in 1791, that is, then the Sixth Amendment does not bar it today.[111] That’s why the Court “accepted” (its word) the forfeiture by wrongdoing exception.[112] It’s also why it recognized that dying declarations might be admissible even when they’re testimonial, though it left that open for further (historical) analysis.[113]

Conventionally, the list of Crawford exceptions stops there, at two.[114] But the Court noted a third category of unconfronted testimonial hearsay admissible at common law—a suspect’s confession offered against its maker:

Several authorities also stated that a suspect’s confession could be admitted only against himself, and not against others he implicated. See 2 W. Hawkins, Pleas of the Crown ch. 46, § 3, pp. 603–604 (T. Leach 6th ed. 1787); 1 Hale, supra, at 585, n. (k) . . . .[115]

The Court did not label this an exception to the confrontation right, but the logic of admissibility was the same as for forfeitures by wrongdoing and dying declarations: Confessions were admissible against their makers without confrontation in 1791, so they are admissible today.[116]

At this point, an enthusiast of confession evidence might be tempted to breathe a sigh of relief. That would be premature. Because the justification for a “confession exception” to the confrontation right is historical, the exception can reach only those confessions that would actually have been admissible under the “common law . . . at the time of the founding.”[117] Enter the “confessions rule.”

1.       The Confessions Rule at Common Law

The confessions rule barred evidence of confessions induced by “hope” or “fear.” “[A] confession forced from the mind by the flattery of hope or the torture of fear,” the King’s Bench explained in 1783’s King v. Warickshall, “comes in so questionable a shape . . . that no credit ought to be given to it; and therefore it is rejected.”[118] While the confessions rule is well known to legal historians[119] and was once part of Fifth Amendment jurisprudence,[120] it’s never been thought relevant to the Sixth Amendment. Crawford made it so.

Common law judges have fretted about the methods used in interrogations since at least the seventeenth century,[121] but initially they left it up to jurors to decide whether to discount a confession induced by manipulative tactics.[122] That began to change in 1740, when a London judge refused to let a jury hear evidence of a confession prompted by the questioner’s promise to be a “Friend” if the suspect confessed.[123] “[A] Confession obtained on a Promise of Friendship,” the case report notes, “ought not to be given in Evidence.”[124] The common law of induced confessions then entered a state of flux. Over the next few decades, some judges excluded confessions induced by hope or fear, others admitted them but cautioned juries to be wary, and a third group admitted them without comment.[125] Eventually the practice of excluding induced confessions prevailed. As John Langbein explains, “Departures” from the exclusionary rule “cease to be reported, practice had become law.”[126]

The leading historians of the confessions rule differ about what motivated its rise. Langbein links it to a larger trend of eighteenth-century criminal procedure reforms targeting the “potential unreliability of prosecution evidence,” including the emergence of defense counsel.[127] Eighteenth-century judges had reason to doubt the veracity of confessions, Langbein explains, in a world in which many were tendered by domestic servants to employers who they relied on for economic survival, and in which private investigators were financially rewarded for apprehending, and sometimes interrogating, suspects.[128] George Thomas and Richard Leo, meanwhile, emphasize two other historical developments. They point first to the ascendence of Whig judges following the English Civil War and the Glorious Revolution who viewed criminal defendants more sympathetically than their sixteenth- and seventeenth-century predecessors.[129] But the most important antecedent for the rule, according to Thomas and Leo, was the emergence in eighteenth-century London of a small group of “proto-detectives,” i.e., magistrates who vigorously investigated crime, including by questioning suspects.[130] More than anything else, on Thomas and Leo’s telling, judicial distrust of confessions that these proto-detectives procured explains the rise of the confessions rule.[131]

Whatever its motivations, the confessions rule had reached mature form in England by the 1780s.[132] Consider the rule’s description in the first source Crawford cited for the proposition that confessions were admissible at common law—the 1787 Sixth Edition of Hawkins’s Pleas of the Crown, edited by Thomas Leach:

The human mind under the pressure of calamity is easily reduced; and is liable, in the alarm of danger, to acknowledge, indiscriminately, a falsehood or a truth, as different agitations may prevail. A confession, therefore, whether made upon an official examination or in discourse with private persons, which is obtained from a defendant either by the flattery of hope, or by the impressions of fear, however slightly the emotions may be implanted is not admissible evidence; for the law will not suffer a prisoner to be made the deluded instrument of his own conviction.[133]

English cases decided prior to 1791 corroborate Leach’s claim that any hope or fear rendered a confession invalid. For example:

·         In 1788, a judge rejected a confession because the defendant was told that “he should have some favour, if he told the truth.”[134] “[P]eople must understand,” the judge commented, “that when they are talking to persons in this situation, what they say, is to come from them without any bias.”[135]

·         In 1783, a judge held a confession inadmissible because the defendant had been told that “unless you give me a more satisfactory account I shall take you before a Magistrate.”[136] The judge remarked that while the comment “scarcely amounts to a threat,” it was “certainly a strong invitation to him to confess.”[137]

·         In 1774, a confession was inadmissible because the suspect was told by a justice of the peace that he “would do all that lay in his power” if the suspect confessed and testified for the prosecution.[138]

·         In 1773, a burglary defendant’s confession was inadmissible because he’d been encouraged to provide prosecution evidence.[139]

The confessions rule served multiple purposes. Leach’s insistence that the “law will not suffer a prisoner to be made the deluded instrument of his own conviction” reflects a concern about autonomy.[140] But the rule’s “formal rationale,” as Thomas and Leo note, was reliability.[141] Warickshall, which remains the canonical confessions rule case,[142] spurned the idea that confessions induced by fear or hope are “rejected from a regard to public faith.”[143] Rather, the court observed, confessions are “received” or “rejected” based solely on “whether they are or are not intitled to credit.”[144]

I can now state this Article’s proposed rule in full. If a defendant’s confession is testimonial, it should be inadmissible unless either (1) the defendant testifies or (2) the confession was obtained without resort to the “flattery of hope” or the “impressions of fear,” even “slightly.”[145] Note that this would not constitutionalize the confessions rule.[146] Many confessions condemned by the common law rule would remain beyond Sixth Amendment regulation because they were not “testimonial”—in particular, confessions to friends, relatives, and victims. But while the confessions rule is not freestanding constitutional law, it should be a limiting principle on any exception to the confrontation right.

2.       The Confessions Rule and Contemporary Interrogation

Interrogation practices have changed since the founding era. In the eighteenth century, most questioning of criminal suspects was conducted by amateur justices of the peace, “thieftakers” (i.e., private individuals who sought to apprehend criminals), or crime victims themselves.[147] Though, as we’ve seen, the first “proto-detectives” appeared in London, professionalized police forces lay in the future.[148] By contrast, typical interrogations today are conducted by trained detectives in purpose-built rooms.[149]

Courts would have to work out what “hope” and “fear” mean in the context of contemporary police interrogations.[150] Some cases may be difficult. For example, do “false friend” tactics induce “hope”?[151] Well, in the very first case (that we know of) in which a British court suppressed a confession because it was induced by hope or fear, the offending remark was that the interrogator would be a “Friend.”[152] So maybe. I’ll come back to that below.[153]

Borrowing a page from Crawford’s refusal to define “testimonial,”[154] I take no position on close cases. But nor do I wish to obfuscate—the rule this Article proposes would, if adopted, have enormous consequences for how police conduct interrogations. That’s because the common law confessions rule plainly condemns some routine modern interrogation practices.

The “Reid technique” is the leading interrogation method in the United States.[155] Hundreds of thousands of American investigators, by some estimates over half of all law enforcement officers, have received training on it.[156] The technique distinguishes “interviews,” where an investigator gathers information, from “interrogations,” which start once the investigator is “reasonably certain” that the suspect is guilty.[157] In an interrogation, the investigator seeks to “active[ly] persua[de]” the suspect to “tell the truth” through a nine-step process.[158] It begins with a “direct, positive confrontation”—i.e., an accusation backed up by “evidence,” which can be fabricated if necessary—and ends with a written confession.[159]

The Reid technique incorporates two key psychological strategies into the nine steps: maximization and minimization.[160] As Saul Kassin, a leading critic of the technique, and Karlyn McNall explain, maximization is “a ‘hard-sell’ technique in which the interrogator tries to scare and intimidate” the interrogatee by “making false claims about evidence” and “exaggerating the seriousness of the offense and the magnitude of the charges.”[161] In the parlance of the confessions rule, maximization induces fear.[162] Minimization, meanwhile, is a “‘soft-sell’ technique in which the police interrogator tries to lull the suspect into a false sense of security by offering sympathy, tolerance, face-saving excuses, and even moral justification, by blaming a victim or accomplice, by citing extenuating circumstances, or by playing down the seriousness of the charges.”[163] As maximization provokes fear, minimization offers hope—social, moral, and (at least by implication) legal hope.[164]

We need not, however, rely exclusively on the Reid technique’s critics. The technique’s manual, Criminal Interrogation and Confessions, explains how to leverage a suspect’s hopes and fears to obtain a confession. For example:

·         Investigators should “allow the suspect to sit in the interview room alone for about five minutes” prior to beginning interrogation to “increase the level of insecurity and apprehension the suspect experiences at the outset.”[165]

·         Investigators should, early in the interrogation, develop a “theme” explaining the suspect’s reasons for committing the crime. The theme should present a “moral excuse” for the crime or minimize its “moral implications.” Sometimes, the theme “may offer a ‘crutch’ for the suspect as he moves toward a confession.”[166]

·         Investigators should “discuss the benefits of telling the truth” when dealing with unemotional suspects. For instance, investigators might tell him that he “will be able to start over. He made some errors in judgment that led to his crime but this is an opportunity to start over with his life and follow a different path.”[167]

·         Investigators should try to “turn[] . . . around” any objections from the suspect. For instance, in a homicide case where the suspect says he didn’t own a gun: “You see, Joe, if you did own a gun and carried it in that night, ready to use it, to kill somebody if they got in your way, that’s one thing. But if the other guy stuck it in your hand, to use it just to scare everybody, that’s something else again.”[168]

·         Investigators should continue to “fend off the suspect’s denials and objections,” “focus more intently” on the theme, “display[] understanding and sympathy,” and attempt to “sell the suspect on [the investigator’s] genuine interest in having the matter resolved,” if a suspect “assume[s] a defeatist posture” or seems “downcast,” “until the suspect shows some physical sign of resignation.”[169]

To be sure, Criminal Interrogation and Confessions cautions investigators not to expressly threaten legal consequences or promise legal leniency.[170] Instead, they should imply a legal threat or promise. Thus, rather than tell a suspect: “Listen, Joe, if this is the first time you did something like this, I’ll talk to the judge and make sure that he gives you probation,” the manual advises investigators to say: “Joe, if this is something that happened on the spur of the moment, that would be important to include in my report.”[171] Social science research confirms the obvious—that Joe understands exactly what the interrogator is implying.[172] Criminal Interrogation and Confessions nonetheless disclaims responsibility: “Through wishful thinking a suspect might surmise . . . [that] he is due some leniency in court,” it allows, but investigators “cannot be held accountable for a guilty suspect’s wishful thinking.”[173]

Criminal Interrogation and Confessions insists when used correctly its methods do not render confessions involuntary for purposes of the Fifth and Fourteenth Amendments.[174] That is mostly true.[175] At the same time, some[176] of these methods exploit suspects’ hopes or fears.[177] The confessions they generate should be inadmissible unless the defendant testifies at trial.[178] Concurrently, law enforcement agencies should adopt interrogation practices that aren’t grounded in psychological manipulation. As we’ll see in the Conclusion, such practices exist and are widely used in other parts of the world.

I noted at the outset of this Section that applying a hope or fear standard to modern interrogation practices will yield contestable cases for courts to work out over time.[179] While I won’t try to anticipate an iterative process of doctrinal refinement, it may help to be a bit more concrete. So, let’s return to the “false friend” techniques mentioned above, in which interrogators falsely claim to be, or seek to be, the suspect’s friend.[180] Do these techniques exploit suspects’ hopes or fears? Sometimes, they surely do. When a confession is explicitly made the price of an interrogator’s “friendship”—and, hence, leniency—the interrogator has exploited hope.[181] Other times, false friend tactics surely do not involve hope or fear. Say an officer covertly records a confession from a suspect who thought the officer was their friend.[182] By itself, the gambit involves neither hope nor fear.

A more interesting scenario is the familiar “Mutt and Jeff” (a.k.a. “good cop, bad cop”) technique. Here’s how the Reid technique’s manual explains how to do it:

Interrogator A, after having employed a sympathetic, understanding approach through his interrogation, expresses regret over the suspect’s continued lying. A then leaves the room. Interrogator B enters and proceeds to make uncomplimentary statements to the suspect, by pointing out his objectionable characteristics or behavior . . . After Interrogator B (the unfriendly one) has been in the interview room for a short while, Interrogator A (the friendly one) reenters and scolds B for his unfriendly conduct. A asks B to leave, and B goes out of the door with a pretended feeling of disgust toward both the suspect and A. A then resumes his friendly, sympathetic approach.[183]

The manual explains that the “psychological reason” the technique works is that it “serves to accentuate the friendly, sympathetic attitude” of Interrogator A.[184] The Supreme Court criticized the technique in dicta in Miranda v. Arizona.[185] Yet, except in egregious cases where the technique is applied against an especially vulnerable suspect or in combination with other psychologically aggressive techniques, courts aren’t troubled by the confessions it produces.[186] For reasons that will become clear below, that makes sense under current due process and self-incrimination regimes.[187] Courts would have to worry about the constitutionality of such confessions in a confrontation-based regime. If Interrogator B’s verbal barrage induced fear, or if Interrogator A’s safe haven induced hope, the confession would be inadmissible unless the suspect, now a defendant, testifies at trial. While that inquiry requires case-by-case determinations, some confessions induced by the Mutt and Jeff tactic that are admissible today would become inadmissible under a confrontation-based approach.

II. Objections

Part I showed that applying Crawford to defendants’ confessions leads to a surprisingly simple rule: A testimonial confession (i.e., one given under police interrogation) is inadmissible against its maker unless either (1) they testify or (2) investigators avoided tactics leveraging hope or fear. The major task that remains is to anticipate objections.[188] This Part considers (and rejects) four: that confessions are not hearsay;[189] that confessions are already adequately regulated;[190] that it’s enough that defendants have the opportunity to confront their confessions during the defense case;[191] and that “self-confrontation” didn’t exist at common law, so it should not exist today.[192] I save a final objection—that the proposed rule would hinder law enforcement—for the Conclusion.

A.      Confessions Are Not Hearsay

Let’s start with what state and lower federal courts say about Crawford’s impact on defendants’ statements. They deny that it has one. Some just assert that a defendant has no “right to confront himself.”[193] Others gesture toward a rationale. Federal Rule of Evidence 801 and similar state provisions define a defendant’s statements as “not hearsay” when offered by the government.[194] Crawford is concerned only with hearsay, courts say, so Crawford has nothing to do with defendants’ statements.[195]

The leading case is United States v. Tolliver.[196] Here was the Seventh Circuit’s analysis:

There are two declarants at issue here. The first is [defendant] Dunklin himself. His statements on the tapes constitute admissions by a party-opponent, and, as such, those statements are, by definition, not hearsay under Federal Rule of Evidence 801(d)(2)(A). Consequently, since the prohibition annunciated in Crawford only applies to hearsay, that prohibition does not cover Dunklin’s statements on the tapes.[197]

By the Seventh Circuit’s logic, if party admissions were covered by an “exception” to the hearsay rule—i.e., if they were included in Federal Rules of Evidence 803 or 804—then Crawford would apply. But because admissions are instead carved out of the definition of hearsay by Rule 801, it does not. Tolliver makes no serious effort to explain why party admissions’ classification as non-hearsay under the Federal Rules of Evidence is constitutionally significant.[198] The court apparently believed the point was self-authenticating.

Even before the Supreme Court’s recent decision in Smith v. Arizona,[199] this reasoning didn’t make much sense. Party admissions are “not hearsay” because the committee that drafted the Federal Rules of Evidence labeled them that way.[200] In every sense but the label, party admissions are an exception to the hearsay rule.[201] A leading evidence casebook urges students not to “struggle to find meaning” in the “Orwellian labeling” of not hearsay, but rather to “think of all the rules” as “exceptions to Rule 802’s hearsay ban.”[202] Thus Tolliver and similar lower court rulings found meaning where none exists.

Smith, moreover, should put to rest any remaining doubt about whether a defendant’s confession is hearsay for purposes of the Confrontation Clause. Writing for the Court, Justice Kagan explained that “hearsay” has a constitutional meaning—it refers to an “out-of-court statement[] offered to prove the truth of the matter asserted.”[203] Whether a statement is hearsay in that sense determines the scope of a federal constitutional right, Kagan explained, so it ordinarily cannot be “expanded or contracted . . . by reference to non-constitutional bodies of law like evidence rules.”[204] Courts thus cannot “accept” an evidence rule’s “nonhearsay label at face value” but must independently determine if an out-of-court statement was “admitted for its truth.”[205]

Smith thus solidified earlier suggestions that the Confrontation Clause applies to out-of-court statements offered for their truth.[206] Confessions to the police, when offered by prosecutors against the defendants who made them, are exactly that, no matter what the Federal Rules of Evidence call them. So, let’s move to more serious objections.[207]

B.      Confessions Are Already Adequately Regulated by the Constitution

A more significant objection is that confessions to the police are already subject to pervasive constitutional regulation under two Fifth Amendment doctrines—voluntariness (from the Due Process Clause)[208] and Miranda (from the Self-Incrimination Clause).[209] Even if the Confrontation Clause could be read to encompass confessions offered against their makers, our imagined objector might say, the Fifth Amendment’s got this.[210]

Note that this is a pragmatic objection. The argument is that regulating confessions under the Sixth Amendment is unnecessary in light of existing doctrines. Pragmatic responses are thus in order. This Section’s is that neither voluntariness nor Miranda do an adequate job of regulating the reliability of confessions. That leaves a gap for the Confrontation Clause to fill.

This Section responds to the pragmatic objection in detail. Before turning to that, I pause briefly to note a more formal version of the objection. Whether or not due process and self-incrimination do a good job of regulating interrogations and confessions, a formalist objector might say that they occupy the field, thereby preempting further regulation from elsewhere in the Constitution. The formalist objector’s problem is that this is not how constitutional rights typically work. “We have rejected the view,” the Court explained in a 1993 case, “that the applicability of one constitutional amendment pre-empts the guarantees of another.”[211] That’s because when a party alleges multiple constitutional violations, the duty of courts is to “examine each constitutional provision in turn.”[212]

To be sure, there is a narrow exception. Sometimes where official action is governed by explicit constitutional text, the Court holds that the specific text preempts claims sounding in “generalized notion[s] of ‘substantive due process.’”[213] The logic is the familiar idea that specific legal provisions trump general ones.[214] It flips that logic upside down to suggest, as the formalist objection does, that “generalized notion[s] of substantive due process”—which are the principal source of constitutional regulations of confessions[215]—preempt an otherwise availing constitutional claim. Besides, none of the three clauses at issue—Due Process, Self-Incrimination, and Confrontation—speaks specifically to confessions made to the police during interrogation, so the logic that general rules yield to specific ones does not even come into play. The formal version of the objection from due process and self-incrimination thus doesn’t get out of the starting blocks. Let’s see if the pragmatic version does any better.

1.       Other Constitutional Regulation of Confessions

We can quickly dispense with the possibility that Miranda sufficiently regulates confessions. The reliability of confessions does appear, occasionally, in Miranda jurisprudence.[216] But Miranda cannot meaningfully protect against unreliable confessions because, as scholars generally agree, it just doesn’t matter much.[217] For one thing, suspects typically waive Miranda’s protections.[218] When people talk to police officers, Charles Ogletree observed, “there appears to be an almost irresistible impulse to respond to the accusations.”[219] For another, the Miranda doctrine is effectively silent about the interrogation tactics police use after a suspect waives their rights, no matter how prone they are to produce unreliable confessions.[220]

I’ll say much more about Miranda’s failure to regulate interrogation tactics later,[221] but for now, let’s focus on the voluntariness doctrine. The Supreme Court has long held that a court violates a defendant’s due process rights by admitting an “involuntary” confession.[222] But what makes a confession “involuntary”? The Court has framed the question enigmatically:

Is the confession the product of an essentially free and unconstrained choice by its maker? If it is, if he has willed to confess, it may be used against him. If it is not, if his will has been overborne and his capacity for self-determination critically impaired, the use of his confession offends due process.[223]

The elusiveness of voluntariness was partly what motivated the Court to adopt Miranda’s clear (but superficial) warn-and-waive protocol.[224] Still, due process remains “a backstop to the Miranda prophylactic rules,” as Guha Krishnamurthi notes.[225] Lower courts thus have to operationalize it. Here’s an effort from the First Circuit:

We determine the voluntary nature of the statements by considering the totality of the circumstances, including both the nature of the police activity and the defendant’s situation. Relevant considerations include the length and nature of the questioning, promises or threats made by investigators, and any deprivation of the suspect’s essential needs. They also include the defendant’s personal circumstances, including his age, education, and mental condition, as well as his prior experience with the criminal justice system. A defendant’s calm demeanor and the lucidity of his statements weigh in favor of finding his confession voluntary.[226]

This is better than the Supreme Court’s formulation, but between the number and variety of “circumstances” to consider, it offers few practical guideposts for police or judges.[227] And differences among jurisdictions abound, especially when it comes to “the nature of the police activity.”[228] In the First Circuit, for instance, it is “well settled” that “an officer does not impermissibly overbear a defendant’s will by . . . suggesting that cooperation may lead to more favorable treatment.”[229] Meanwhile, in the Ninth Circuit, “there are no circumstances in which law enforcement officers may suggest that a suspect’s exercise of the right to remain silent may result in harsher treatment by a court or prosecutor.”[230] Nonetheless, as Mark Godsey observes, the law’s “ambiguity most often works in [the government’s] favor,” and confessions are rarely deemed involuntary.[231]

Where does a confession’s reliability fit into the voluntariness picture? Early voluntariness cases were rooted in reliability,[232] but the Court has since taken a different tack. In Colorado v. Connelly, it held that due process’s concern in the interrogation context is “coercive government misconduct.”[233] Absent such misconduct, the Court said, “there is simply no basis for concluding that any state actor has deprived a criminal defendant of due process of law.”[234] Thus the reliability of confessions is not a freestanding due process interest.[235] “The aim of the requirement of due process is not to exclude presumptively false evidence,” the Court explained, “but to prevent fundamental unfairness in the use of evidence, whether true or false.”[236]

But reliability isn’t entirely irrelevant either.[237] If police intentionally extract a false confession, that’s obviously misconduct.[238] Further, a false confession might be a proxy for coercive techniques.[239] So while reliability isn’t a stand-alone due process concern, due process could promote reliability anyway. Let’s call that the “secondary regulation” hypothesis. It makes sense intuitively.[240] Why would anyone “voluntarily” confess to a crime they didn’t commit? Unfortunately, the hypothesis doesn’t withstand empirical scrutiny. Confessions that are voluntary and false are not aberrations in American courts.

Begin with evidence from DNA exonerations, which undermines the theory that only the guilty voluntarily confess. Brandon Garrett closely examined case records from two waves of DNA exonerations.[241] Of the initial 250 people exonerated by DNA evidence, forty-two falsely confessed.[242] At least twenty-eight had sought to suppress their confession, and none succeeded.[243] Likewise in the second wave, which included twenty-six more individuals who falsely confessed, Garrett found that judges had invariably rejected voluntariness challenges.[244]

DNA exonerations show that false confessions can fall through the voluntariness doctrine’s cracks. They do not tell us how often courts admit false confessions,[245] but they leave the secondary regulation hypothesis on the ropes. For the knockout, we turn to social psychology. Innocent people may not confess without provocation, but experiments reveal that it takes much less of it than our law or intuitions suppose.[246] These studies, which replicate law enforcement interrogation techniques as closely as possible in an experimental setting, reveal that the maximization and minimization tactics at the heart of the Reid technique lead many to falsely confess.[247]

Consider two examples. First, Melissa Russano and colleagues conducted a study in which undergraduates were accused of cheating on logic problems during an experiment.[248] Experimenters “interrogated” the students and told them that the professor running the study was “annoyed and upset by the situation.”[249] (The experimenters cleverly manipulated conditions so that some of the subjects had actually “cheated,” but we’ll focus on the subjects who were falsely accused.) Subjects were then asked to sign a document admitting that they cheated. Some were also exposed to standard police interrogation tactics, including minimization (“I’m sure you didn’t realize what a big deal it was”) and assurances of leniency if they confessed (“things could probably be settled pretty quickly”).[250] A small number of innocent subjects—6 percent—signed the confession without the interrogator using either tactic.[251] The percentage more than doubled when interrogators gave assurances of leniency.[252] It tripled when the interrogators used minimization. Nearly half of the innocent subjects confessed when interrogators applied both tactics.[253]

Jessica Klaver and her colleagues saw similar results in their study of interrogation techniques.[254] They gave their subjects, also undergraduates, a typing task.[255] They told them that if they pressed a particular key (either ALT or ESC), the computer would crash with a resulting loss of data.[256] Immediately after they instructed the subjects to press the Z key (which is close to ALT but far from ESC), the experimenters stopped the task and falsely accused them of pressing the forbidden key.[257] In the “interrogations” that followed, experimenters used either maximization or minimization techniques.[258] In the minimization condition, experimenters told subjects: “Don’t worry. It was just an accident. You didn’t mean to hit the ALT/ESC key. Several participants so far have pressed the ALT/ESC key during this task. Are you sure you didn’t press it?”[259] Meanwhile, in the maximization condition, experimenters said things like: “There is no way to recover the data. It looks like the entire project may be delayed now.”[260] These tame tactics were, of course, easily within the bounds of what’s permissible under the voluntariness doctrine. The experimenters then asked the subjects to sign a written confession.[261] Of subjects accused of pressing the ESC key—an implausible allegation, considering the distance from ESC to Z—5 percent of people exposed to the maximization condition and 23 percent of those exposed to minimization falsely confessed.[262] But the really dramatic findings were among the subjects accused of pressing ALT. Nearly half, 47 percent, exposed to the maximization tactic and 70 percent of those exposed to minimization falsely confessed.[263]

These studies demonstrate the faultiness of the intuition that innocent suspects don’t confess without “coercion.” So much for the secondary regulation hypothesis. The common law confessions rule was predicated on its own intuition—that even the “slightest” amount of “hope” or “fear” could induce an innocent person to confess.[264] The psychological intuitions of eighteenth-century jurists and lawyers, it turns out, were much closer to the empirical mark than those of their twentieth- and twenty-first-century successors.

To be sure, no one knows how frequently maximization and minimization tactics lead suspects to confess falsely.[265] Police do not randomly select individuals for interrogation, so it seems unlikely that most people who confess are innocent. That’s a function of base rates. If the meteorologist in Hilo, Hawai‘i forecasts rain every day instead of checking the radar, they’ll usually be right.[266] Likewise, if most people hailed into the interrogation room are guilty, then confessions will usually be true. But that doesn’t mean that either the forecast or the confession has epistemic value. Manipulative interrogation tactics that extract confessions from guilty and innocent suspects alike contribute little to the rational assessment of guilt or innocence.[267]

The existing constitutional law of confessions is unable to deal with unreliable confessions, but the Confrontation Clause could. By rendering confessions induced with minimization and maximization tactics inadmissible (unless the defendant testifies), the Confrontation Clause would draw a line between confessions motivated by remorse and those motivated by a suspect’s fear of not confessing or hope to get something for it. The former are at least potentially reliable; the latter are not.[268] If, as I’ve assumed, the Court’s due process and self-incrimination jurisprudences are fixed points, confrontation is the best constitutional strategy for ridding America’s courts of unreliable confessions.[269]

2.       A Note on Confrontation and “Reliability”

The previous Section argued that existing constitutional doctrines don’t pay enough attention to the reliability of confessions. That argument is only responsive to the objection (that confessions are sufficiently regulated) if the reliability of confessions is constitutionally significant. We have seen claims that it isn’t.[270] But the Confrontation Clause hasn’t been part of the conversation. As this Section explains, the reliability of confessions is in its bailiwick.

Few would disagree that reliability is a—if not the—central concern of the Confrontation Clause.[271] But what sort of reliability? Consider this passage:

At the core of [Ohio v.] Roberts was the proposition, rarely disputed, that the Confrontation Clause aims to ensure the reliability of evidence. That’s ‘the Clause’s ultimate goal,’ Justice Scalia said in Crawford. It follows that if a hearsay statement is highly likely to be reliable and cross-examination cannot readily take place, good sense and justice suggests the statement may be admitted in absence of confrontation.[272]

If the Confrontation Clause targets accurate trial outcomes, this logic—that really reliable hearsay should be admitted even without confrontation—would be spot-on. But the Confrontation Clause specifies a rule of criminal procedure, not a rule of evidence.[273] While evidence rules are (mostly) geared to helping factfinders make accurate decisions,[274] criminal procedure rules that confer rights on the accused typically have a different orientation. They do not minimize errors but bias them against the government.[275] So it is with the Confrontation Clause.[276] It aims to avoid erroneous convictions, not erroneous verdicts in general.

As the Court understands it, the Confrontation Clause promotes reliable convictions via a specific device: “testing in the crucible of cross-examination.”[277] The theory is that defendants can expose “infirmities” in the testimony of the government’s witnesses, “thereby calling to the attention of the factfinder the reasons for giving [it] scant weight.”[278] That’s exactly what a defendant who confronts their confession would try to do. The defendant would point out “infirmities” in their own out-of-court “testimony” to persuade the jury not to give it much weight.[279]

The rubber meets the road when out-of-court testimony can’t be tested because the declarant died, left town, or (most relevant here) asserted a privilege. In those scenarios, there are only two ways to proceed, and both risk an erroneous verdict. Either we withhold the testimony from the jury, risking a false acquittal, or we admit it without testing, risking a false conviction. We could choose by applying a rule that minimizes the total risk of error. If the Confrontation Clause announced an evidentiary principle, that’s exactly what we would do. Instead, to protect the reliability of convictions, we allocate the risk of error to the government by excluding the testimony. As Blackstone might say, it’s better to exclude ten true untested out-of-court accusations—including confessions—than to admit a single false one.[280]

C.      Defendants Can Already Confront Their Confessions

A third objection returns to the interplay between confrontation and self-incrimination. We saw previously that a defendant does not lose the privilege against self-incrimination by invoking confrontation during the prosecution’s case.[281] But that’s not the end of the story because the defendant also has the right to testify during their own case.[282] Thus, they have a right to tell jurors all the reasons they should not believe the confession, which is what confrontation means in our context.[283] Doesn’t a defendant’s opportunity to confront their hearsay satisfy the Confrontation Clause? Let the prosecutor introduce the confession in the government’s case, and let the defendant confront it in theirs, an objector might say.[284]

The objection may sound reasonable, but it runs aground at the Supreme Court’s decision in Melendez-Diaz v. Massachusetts.[285] There, the Court considered whether Massachusetts prosecutors could use a lab report at trial without calling the analysts who produced it.[286] The state argued that it did not have to produce the analysts during its case because the defendant “had the ability to subpoena” them during his.[287] Writing for the Court, Justice Scalia explained the flaw in the state’s argument:

[F]undamentally, the Confrontation Clause imposes a burden on the prosecution to present its witnesses, not on the defendant to bring those adverse witnesses into court. Its value to the defendant is not replaced by a system in which the prosecution presents its evidence via ex parte affidavits and waits for the defendant to subpoena the affiants if he chooses.[288]

The Court made two things clear: first, that the Confrontation Clause imposes a duty on the government to produce testimonial declarants, and second, that the defendant’s own opportunity to call a testimonial declarant is no substitute.[289] This “production rule” reflects that the Confrontation Clause’s point is to enable the defendant to test the government’s evidence.[290] That’s why confrontation is ordinarily something that happens during the government’s case.[291]

Current law thus doesn’t support the idea that a defendant’s right to testify in their case displaces the government’s duty to produce prosecution witnesses.[292] Still, Melendez-Diaz was not about declarants who are also defendants, so perhaps the Court should make an exception. In our scenario, after all, the defendant complains that their confession is untested and yet, when given the chance to test it, refuses. To our objector, that sounds like gamesmanship. Ought there be a “gamesmanship exception” to Melendez-Diaz’s production rule?[293]

The initial difficulty for a gamesmanship exception is that it lacks a legal hook. It sounds in generalized considerations of fairness, but the Confrontation Clause famously produces results that are, from the government’s vantage, unfair.[294] That’s because it excludes otherwise admissible hearsay when, through no fault of the government, the declarant is unavailable.[295] Take Crawford itself. As explained above, the only thing preventing the state of Washington from using the declarant’s testimony at trial was that the defendant invoked spousal privilege.[296] For better or worse, fairness to the government is not a guiding Confrontation Clause principle.[297]

Beyond that, and even if the Court one day changes course and holds that a defendant’s ability to subpoena a witness does avert the government’s obligation to produce them, the equities are complicated when it comes to the defendant’s testimony. A defendant might choose not to testify at trial for reasons having nothing to do with their confession. Most obviously, if a defendant testifies, the jury could learn about the defendant’s (otherwise inadmissible) prior convictions.[298] More broadly, as Stephen Schulhofer explains, a defendant—even an innocent one—may choose not to testify out of fear that they could “look sleazy” to the jury or that they could be “inarticulate,” “nervous,” “easily intimidated,” or “vulnerable to a clever cross-examination” by virtue of their “vague memory on some of the details.”[299] “Some people can handle this kind of situation,” Schulhofer notes, “but others, especially if they are poor, poorly educated or inarticulate, cannot.”[300]

The objection would make defendants risk all of that if they wished to confront their confessions. This puts a defendant who falsely confessed to a difficult choice between confrontation and self-incrimination. The defendant can testify, but if they do, they chance the jury judging them harshly for their record or demeanor. It’s a high price to make a defendant pay for confrontation.

That said, the objection generates an intriguing question, which I flag but do not resolve. Could a state adopt a process through which a defendant confronts their confession during the prosecution’s case without testifying? Imagine a statute providing that when a defendant (1) objects to the admissibility of their confession under the Confrontation Clause and (2) invokes their right not to testify during the prosecution’s case, they may then enter into evidence a written statement explaining in their own words why the jury should not believe the confession. The defendant would not, by making the statement, waive the privilege against self-incrimination, expose themself to impeachment by prior conviction, or invite the jury to negatively judge their demeanor. That’s all to their advantage. On the other hand, one could make a plausible argument that the writing would satisfy Crawford, allowing the confession to be admitted even when the defendant does not testify.[301] That benefits the prosecution. The critical question is whether the state could deem the defendant to have forfeited or waived their confrontation right if they declined to make a written statement. As this Article breaks enough ground, I leave that for another day.

D.     “Self-Confrontation” Did Not Exist at the Founding

In some ways, this Article’s proposal has originalist credentials.[302] It is grounded in Crawford, one of the Court’s most touted originalist criminal procedure decisions of recent decades,[303] and it takes common law practice seriously as a source of constitutional doctrine.[304] That said, a jurist in 1791 would have found the idea of making certain confessions admissible only if the defendant testifies very peculiar. So, the credentials go only so far.

This Section addresses two potential originalist objections to the proposal. The first is that the framers of the Confrontation Clause didn’t condition the admissibility of a confession on the defendant testifying, so the Sixth Amendment cannot require that we do so today. The second is that the relevant public (lawyers and judges) in 1791 did not understand the confrontation right to include confessions offered against their makers.[305] For both, my response is ultimately the same—the history “runs out” before it can undermine the proposal.

The first objection points to the lack of a historical antecedent for conditioning the admissibility of a confession on a defendant’s testifying. I concede the objection’s premise. No one in 1791 would have conditioned the admissibility of a confession or anything else on a defendant taking the witness stand, because criminal defendants on both sides of the Atlantic were categorically disqualified from testifying.[306] In that sense, the criminal trial of 1791 is unrecognizable to us. We have since seen the folly of silencing defendants,[307] so unless we are going to revisit the right of defendants to testify, we must move into the realm of counterfactuals.[308] The question is whether the Sixth Amendment’s authors would have conditioned the admissibility of a confession on a defendant’s testifying in a world in which defendants could testify.

For the sake of argument, I will stipulate that lawyers and jurists of the time would not have conditioned the admissibility of a confession on a defendant’s testifying even in a counterfactual world in which defendants were allowed to do so. But we’ve departed from the common law baseline in a second way. At common law, as we have seen, judges barred confessions induced by hope or fear.[309] So now we can pose the question in its full context: Would the Sixth Amendment’s authors have conditioned the admission of a confession induced by hope or fear on a defendant testifying, in a world in which defendants could testify? We have no idea how they would have thought about that question. More importantly, it seems implausible that they would have thought about it at all. “When the authors do not know what meaning they intended,” Larry Alexander notes, “then there is no intended meaning for the audience to discover.”[310]

The second originalist objection focuses on the meaning of “confrontation.” Relevant members of the public in 1791, the objector might say, would not have understood the confrontation rule to cover a criminal defendant’s own confession. And if the confrontation rule didn’t encompass confessions then, the Confrontation Clause doesn’t now.

The problem is that there was no single historical confrontation rule for the Sixth Amendment to incorporate.[311] Rather, there were a variety of common law rules and practices that today sit under the umbrella of confrontation. They include (or may include):

1)      The opportunity to cross-examine live witnesses.[312]

2)      The rule against using a non-suspect’s examination (taken under oath) without either live testimony from the witness or a prior opportunity for cross examination.[313]

3)      The rule that a suspect’s examination can be used against that suspect, but not others, and only if it was not induced by hope or fear.[314]

The crux of the question is whether the italicized portion of the third rule belongs under the umbrella. Perhaps only the unitalicized portion—i.e., “a suspect’s examination cannot be used against anyone else”—counts as “confrontation,” while the italicized portion is something else entirely.

There is evidence that Founding-era lawyers and jurists saw the two parts of the third rule as two sides of the same coin. Leach’s 1787 edition of Pleas of the Crown observed that confessions “hath always been allowed to be given in evidence against the party confessing, but not against others,” immediately after which it treated readers to its account of the confessions rule.[315] Gilbert’s The Law of Evidence, also cited in Crawford, likewise noted that confessions are “[e]vidence against the Party himself who made it; but cannot be made use of against others,” and then proceeded forthwith to a discussion of the requirement that the confession be made “[f]reely without menaces or any species of undue influence.”[316]

These treatises offer some support for the idea that the confession rule was intertwined with the rule against using a suspect’s confession against others. But because the right of confrontation is our term for loosely associated common law rules and practices, eighteenth-century treatises cannot determine its scope. “When history runs out,” Michael McConnell observes, “we have to use other methods of interpretation.”[317] With the history indeterminate (and the text even more so), we’re stuck assessing whether applying the confrontation right to confessions serves the Confrontation Clause’s purposes as we understand them.[318]

III. Crawford and Miranda

I have proposed and defended a reading of Crawford under which a defendant’s testimonial confession would be inadmissible unless either the defendant testifies or the interrogation was free of hope or fear. This final Part revisits the Court’s most famous confessions case—perhaps its most famous case on any topic—in light of that reading.[319]

A quick primer before we get into details: Miranda held that custodial interrogation is “compulsion” and, therefore, violates the Fifth Amendment’s privilege against compelled self-incrimination[320] unless interrogators inform the suspect of several rights (including to remain silent) and the suspect voluntarily waives them.[321] The Supreme Court has since adjusted the doctrine’s justification—now the warnings serve to prevent, rather than cure, Fifth Amendment violations[322]—and weakened it in important ways.[323] The warn-and-waive protocol, however, remains.[324]

The Miranda doctrine has disappointed just about everyone.[325] Unsurprisingly, many scholars have conducted post-mortems.[326] In one of the best, William Stuntz argued that Miranda’s “mistake” was “to sort suspects rather than police tactics.”[327] Miranda divided suspects into the few, those who exercise their rights, and the many, those who waive them.[328] The few, Stuntz explained, are not interrogated, while the many “have very nearly no protection at all.”[329]

Stuntz was right to highlight Miranda’s failure to protect unsophisticated suspects. Most interrogatees waive their Miranda rights.[330] Whether they do so because police cajole them or because people are naturally keen to consent to official requests is not important for these purposes.[331] What is important is that the Court imposed no meaningful back-end limitations on interrogation after waiver.[332] But sorting suspects rather than tactics was the consequence of Miranda’s mistake, not the mistake itself. The underlying error was treating psychologically manipulative interrogation as solely a concern of the Self-Incrimination Clause.

Miranda’s primary target was psychological manipulation in interrogation.[333] “[T]he modern practice of in-custody interrogation,” the Court emphasized, “is psychologically rather than physically oriented.”[334] For evidence, the Court canvassed police interrogation manuals, with particular attention on Criminal Interrogation and Confessions, the Reid technique’s manual.[335] Across six pages of the U.S. Reports,[336] the Court detailed the psychological ploys it found therein, which included isolation,[337] maximization,[338] minimization,[339] and deception.[340] While such tactics do not involve “physical intimidation,” the Court deemed them “equally destructive of human dignity.”[341] Therefore, it announced it would “concern [itself] primarily with this interrogation atmosphere and the evils it can bring.”[342]

The Court’s mitigation strategy—the warn-and-waive protocol—came at the end of its analysis.[343] The crucial step on the way was the Court’s holding that the psychological pressures of interrogation can constitute “compulsion” within the meaning of the Self-Incrimination Clause even when they do not amount to “coercion” for purposes of the Due Process Clause. As Stephen Schulhofer has explained in reconstructing the Court’s logic (which is difficult to discern from the face of the opinion), what made interrogation “compelling,” in the Court’s eyes, was its purpose.[344] If “governmental efforts [are] intended to pressure an unwilling individual to assist as a witness in his own prosecution,” Schulhofer notes, then they are compelling even if they do not overbear a suspect’s will, and thus would not offend due process.[345] Because pressuring people to assist the government is precisely the point of modern interrogation techniques, the Court ruled that the interrogation room is a compelling environment.[346]

Miranda’s distinction between compulsion and coercion (in the “overbear the will” sense) has the virtue of separating two discrete constitutional interests in interrogation.[347] On one hand, there is the individual and collective interest in avoiding misconduct by government officials. On the other, there is the purely individual interest in autonomously choosing whether to “assist as a witness in [one’s] own prosecution.”[348] In some situations, one set of facts will trigger both coercion and compulsion, but not always. Miranda recognized that modern interrogation methods, which subtly decompose a person’s choice about whether to talk to the police, almost always trigger the latter, but not the former.[349]

The Court’s mistake was stopping there. Preventing government abuse and protecting autonomy are valid concerns of the Due Process and/or Self-Incrimination Clauses. But they do not comprise a complete list of constitutional concerns implicated by modern interrogations because the reliability of convictions, the Confrontation Clause’s domain, is missing.[350] To be sure, the Court can hardly be faulted for overlooking the Confrontation Clause.[351] The omission comes into view only with Crawford’s innovation in Confrontation Clause jurisprudence,[352] the social psychological experiments on false confessions,[353] and the rise of DNA exonerations.[354] Still, with the benefit of hindsight, it was a mistake.

And it was consequential, because the Court’s heavy reliance on self-incrimination led almost inexorably to the warn-and-waive protocol. If “compulsion” turns on the purpose of an interrogation,[355] it is broad enough to reach almost any tactic.[356] As a practical matter, the Court was left with a binary choice: forbid all (or nearly all) interrogations or permit the taint of compulsion to be washed away with warnings and waivers. Abolishing interrogation was not politically viable, so the Court accepted the option that remained.[357] Manipulative tactics would be strictly off-limits without waiver, and virtually carte blanche with it.[358]

The Confrontation Clause (with the benefit of Crawford) allows for a narrower but more robust regulatory approach to interrogation. It would be narrower because the Confrontation Clause implicates only a subset of the tactics that self-incrimination does. Where basically any interrogation tactic can constitute compulsion under Miranda, only tactics that leverage hope or fear can render a confession inadmissible under Crawford.[359] That is, in some sense, a historical accident, but a salutary one.[360] And because a confrontation-based confession law would be targeted, it would not raise the specter of wholesale abolition of interrogation, avoiding the political conundrum the Court encountered in Miranda.[361]

A confrontation-based regime would also have been more robust. That’s because confrontation is not amenable to waiver in the interrogation room. Confrontation is a trial right belonging to defendants, not an investigation right belonging to suspects.[362] While the Supreme Court has allowed suspects to waive their investigation rights anywhere and everywhere,[363] I’m aware of no instance in which a suspect successfully waived a trial right without a judge being involved.[364] Absent a judge in the interrogation room, warn-and-waive thus does not offer a viable path around the Confrontation Clause.[365]

Miranda sorts suspects instead of tactics because its reliance on self-incrimination led it to a dysfunctional warn-and-waive protocol. Had the Court relied instead, or in addition, on the confrontation right as we know it after Crawford, a different remedy would have been available.[366] Once again, the point is not to fault the justices who decided Miranda for failing to anticipate Crawford. We can understand Miranda’s mistake, but we should still correct it.

Conclusion

This Article has argued that on a straightforward reading of Crawford v. Washington, a defendant’s testimonial confession ought to be inadmissible unless either they testify or the interrogation that produced it was devoid of tactics leveraging hope or fear. It also responded to several likely objections to that reading. I saved one more for the end.

Consider how the rule would play out in practice. Confessions made to police without psychological manipulation would be admissible, regardless of whether the defendant testified. Defendants would gain a veto, however, over the admissibility of confessions induced by hope or fear, which they would use to exclude confessions both true and false. To avoid that, investigators would have to remove hope- and fear-based tools from their interrogation toolkits.[367] In all likelihood, (some) law enforcement leaders would not be enthusiastic about doing so. Abolishing minimization and maximization tactics, they’d say, would interfere with their ability to solve crime.[368] They’d probably borrow a metaphor from the reaction to Miranda and argue that the rule would “handcuff” police.[369] They might be right. What’s more likely is that they would (perhaps under protest) update their interrogation practices. Fortuitously, interrogation tactics not based on leveraging suspects’ hopes and fears already exist and are widely used by investigators elsewhere in the world.

The counterpoint to the “accusatorial” model of interrogation typified by the Reid technique is the “information-gathering” model, and the most prominent version is PEACE.[370] From origins in Britain in the early 1990s, PEACE has spread to the rest of the United Kingdom, Australia, New Zealand, Norway, portions of Canada, and even the Vermont State Police.[371] Where the Reid technique emphasizes investigators’ control of suspects, the PEACE technique is designed to get suspects talking.[372] After eliciting a suspect’s narrative of events, PEACE-trained investigators break it down into its component parts, probing for inconsistencies and confronting the suspect with genuine evidence that contradicts their account.[373] The theory, as Margareth Etienne and Richard McAdams explain, is that talkative guilty suspects will “tend to start contradicting themselves.”[374]

The crucial question is how well information-gathering interrogation methods like PEACE “work” compared to accusatorial approaches.[375] In other words, do they produce fewer bad outcomes (false confessions) and/or more good ones (true confessions)? Empirical work on the comparative diagnosticity of the models is ongoing.[376] It’s a difficult question to answer, as studies can be conducted only in the psychologist’s laboratory, which can never truly replicate a police interrogation room, or in the field, where the ground-truth of a confession is unknowable.[377] The evidence we have, however, supports the information-gathering approach.[378] As the authors of a 2014 review concluded: “[I]nformation-gathering approaches showed superior diagnosticity by significantly increasing the elicitation of true confessions and significantly reducing the incidence of false confessions” when compared with accusatorial approaches.[379]

That said, the empirical work is too preliminary to definitively resolve anything.[380] So to bias the discussion against this Article’s proposal, I’ll assume that information-gathering interrogations today yield fewer true confessions than accusatorial interrogations.[381] If that assumption is correct, should the proposal be rejected on the grounds that it will hinder law enforcement? No, for two reasons.

First, if the Confrontation Clause compels law enforcement to use a somewhat less effective interrogation technique, what of it? Criminal procedure rules place certain investigative and prosecutorial practices out-of-bounds even when they’re effective.[382] That’s the price we pay for the values embedded in the Fourth, Fifth, and Sixth Amendments.[383]

Second, if non-manipulative approaches to interrogation are less effective today, that does not mean that they will be tomorrow. The problem is that American police currently have little incentive to change their approach to interrogation. While the psychologically manipulative tactics our law permits do not produce reliable results, the results seem reliable to juries and judges.[384] Defendants, anticipating that, often plead guilty in cases where an unreliable confession would be offered at trial.[385] But constraint can foster innovation.[386] If police learn that using hope- or fear-based interrogation tactics will render confessions inadmissible, they will have a reason to find and refine better methods. The ultimate outcome could be an interrogation regime that’s effective in obtaining true confessions and avoiding false ones.

Copyright © 2025 William Ortman, Professor of Law and David Adamany Research Scholar, Wayne State University. For helpful comments and conversations, thanks to Al Alschuler, Jeff Bellin, Don Dripps, Dan Epps, Stephen Ferruolo, Thomas Frampton, Rachel Harmon, Shelly Layser, Greg Mitchell, Hillel Nadler, Michael Oswalt, Lisa Ramsey, Maimon Schwarzschild, Steven Smith, Mila Sohoni, Jon Weinberg, Chris Williams, Steve Winter, and additional participants in workshops at the University of San Diego and the University of Virginia. Thanks also to three generous but anonymous scholars who engaged with a draft of the paper for purposes of Wayne State’s promotion process. For helping me plumb the Confrontation Clause’s depths, many thanks to my Evidence and Criminal Procedure students.

          [1].     John Henry Wigmore, A Treatise on the Anglo-American System of Evidence in Trials at Common Law § 835 (2d ed. 1923).

          [2].     See People v. Powell, 37 N.Y.3d 476, 490 (2021) (“[A]wareness of . . . false confessions has evolved to the point of common knowledge, if not conventional wisdom.”).

          [3].     See infra notes 241-245 and accompanying text; see also Brandon L. Garrett, Convicting the Innocent: Where Criminal Prosecutions Go Wrong 14–44 (2011).

          [4].     See infra notes 246–263 and accompanying text. For an overview of the evidence, see generally Saul Kassin, Duped: Why Innocent People Confess (2022).

          [5].     See Miranda v. Arizona, 384 U.S. 436, 461–62 (1966); Culombe v. Connecticut, 367 U.S. 568, 621 (1961). Note that the Self-Incrimination Clause yields two rights pertinent to interrogation: the (textual) right against compelled self-incrimination and the (prophylactic) right to pre-interrogation warnings. See Vega v. Tekoh, 597 U.S. 134, 141–42 (2022). Our focus will be the latter.

          [6].     See infra notes 217–220 and accompanying text.

          [7].     See infra notes 222–239 and accompanying text.

          [8].     U.S. Const. amend. VI (“In all criminal prosecutions, the accused shall enjoy the right . . . to be confronted with the witnesses against him . . . .”).

          [9].     541 U.S. 36 (2004).

        [10].     U.S. Const. amend. VI; see also Douglas v. Alabama, 380 U.S. 415, 418 (1965) (explaining that a “primary interest secured by it is the right of cross-examination”).

        [11].     Crawford declined to comprehensively define “testimonial” but held that statements to police during formal questioning “qualify under any definition.” 541 U.S. at 52.

        [12].     See id. at 52–54.

        [13].     Id. at 53–54.

        [14].     See infra Part I.A.

        [15].     See Colorado v. Connelly, 479 U.S. 157, 182 (1986) (Brennan, J., dissenting) (“[T]he introduction of a confession makes the other aspects of a trial in court superfluous.”).

        [16].     See Crawford, 541 U.S. at 68–69.

        [17].     See Salinas v. Texas, 570 U.S. 178, 184 (2013) (noting defendant’s “absolute right not to testify” (internal quotations omitted)).

        [18].     See infra Part I.B.

        [19].     See Crawford, 541 U.S. at 54.

        [20].     See id.

        [21].     Id. at 45.

        [22].     R v. Warickshall, 168 Eng. Rep. 234, 235 (K.B. 1783).

        [23].     See infra Part I.C.1.

        [24].     See infra Part I.C.2.

        [25].     The only prior scholarship to seriously consider Crawford’s implications for defendants’ confessions is Mark A. Summers, Taking Confrontation Seriously: Does Crawford Mean That Confessions Must Be Cross-Examined, 76 Alb. L. Rev. 1805 (2013). Summers acknowledged that Crawford could be read to exclude the confessions of non-testifying defendants, but he regarded this “literal reading” as a problem to be avoided via a historical exception. Id. at 1819–20. I applaud Summers for considering the issue, but I disagree with his analysis of it. The “literal reading” of Crawford is an opportunity, not a problem. See infra Part II.B.2. And while common law practices do justify an exception for confessions, that exception is narrow. See infra Part I.C.

        [26].     Federal Rule of Evidence 801(d)(2)(A) exempts a party’s statements (when offered by the party’s opponent) from the ordinary definition of hearsay as an out-of-court statement offered for its truth.

        [27].     See infra Part II.A.

        [28].     See supra notes 5–7 and infra Part II.B.

        [29].     See infra Part II.C.

        [30].     See infra Part II.D.

        [31].     See Louis M. Holscher, The Legacy of Rock v. Arkansas: Protecting Criminal Defendants’ Right to Testify in Their Own Behalf, 19 New Eng. J. on Crim. & Civ. Confinement 223, 226 (1993) (describing defendant disqualification).

        [32].     384 U.S. 436 (1966).

        [33].     See Welsh S. White, Miranda’s Failure to Restrain Pernicious Interrogation Practices, 99 Mich. L. Rev. 1211, 1211 (2001).

        [34].     See infra Part III.

        [35].     I’ve argued that Crawford erred in limiting confrontation to trials, thereby excluding the plea-bargaining stage of criminal prosecutions. See William Ortman, Confrontation in the Age of Plea Bargaining, 121 Colum. L. Rev. 451, 479–85 (2021). That line drawing isn’t important here, so I accept arguendo the Court’s view.

        [36].     The stipulation covers only the Supreme Court. See infra Part II.A. It does not, moreover, include everything the Court has ever said in dicta. For instance, in Colorado v. Connelly, 479 U.S. 157, 157 (1986), the Court said that the reliability of confessions is “a matter to be governed by the evidentiary laws of the forum.” That’s wrong, I will argue, because it fails to consider the Confrontation Clause. On the other hand, I take as fixed Connelly’s holding that the reliability of confessions is not a freestanding due process concern. See infra notes 233–236 and accompanying text. To be sure, reliability was once fundamental to due process analysis of confessions, and it could be again. See Eve Brensike Primus, The Future of Confession Law: Toward Rules for the Voluntariness Test, 114 Mich. L. Rev. 1, 34–55 (2015) (proposing revitalized voluntariness test). But that’s a topic for another day.

        [37].     For a list of cases in which the Supreme Court overruled itself, see Table of Supreme Court Decisions Overruled by Subsequent Decisions, Const. Annotated, https://constitution.congress.gov/resources/decisions-overruled/ [https://perma.cc/SGA3-XJSX].

        [38].     By one count, almost three hundred articles were written about Crawford in its first two years of life. Andrew King-Ries, State v. Mizenko: The Montana Supreme Court Wades into the Post-Crawford Waters, 67 Mont. L. Rev. 275, 280 (2006).

        [39].     U.S. Const. amend. VI.

        [40].     See, e.g., Mattox v. United States, 156 U.S. 237, 242 (1895) (addressing question in the nineteenth century).

        [41].     For an explanation of what makes hearsay “testimonial” in the Crawford framework, see infra notes 58–61.

        [42].     Crawford v. Washington, 541 U.S. 36, 61 (2004).

        [43].     See id. at 50–57.

        [44].     Id. at 50–53.

        [45].     Id. at 44.

        [46].     Id. at 54.

        [47].     Justice Scalia’s historical analysis is controversial. For critical assessments, see id. at 69–73 (Rehnquist, C.J., concurring); see generally Thomas Y. Davies, What Did the Framers Know, and When Did They Know It? Fictional Originalism in Crawford v. Washington, 71 Brook. L. Rev. 105 (2005).

        [48].     Crawford, 541 U.S. at 51 (“[N]ot all hearsay implicates the Sixth Amendment’s core concerns.”).

        [49].     Id. at 68.

        [50].     Id.

        [51].     Id. (“Where testimonial evidence is at issue . . . the Sixth Amendment demands what the common law required: unavailability and a prior opportunity for cross examination.”).

        [52].     Id. at 51 (explaining that “witnesses” are those who “bear testimony,” which “in turn, is typically [a] solemn declaration or affirmation made for the purpose of establishing or proving some fact” (internal quotations omitted)). On the proposition that declarants are not Sixth Amendment witnesses until the government uses their statements, see Ortman, supra note 35, at 478–79.

        [53].     I elaborate this point below. See infra notes 285–291 and accompanying text.

        [54].     See Crawford, 541 U.S. at 53–54.

        [55].     Id. at 62 (noting the “rule of forfeiture by wrongdoing (which we accept)”).

        [56].     Id. at 56 n.6.

        [57].     Id. at 45.

        [58].     Id. at 68.

        [59].     I have summarized these difficulties elsewhere. See Ortman, supra note 35, at 459–60. They are ongoing. See Smith v. Arizona, 602 U.S. 779, 800-01 (2024) (declining to resolve whether a lab analyst’s statements were testimonial because the question was not presented in the certiorari petition, but “offer[ing] a few thoughts, based on the arguments made here, about the questions the state court might usefully address” if it needs to decide it on remand).

        [60].     Crawford, 541 U.S. at 52 (noting that statements during interrogation are “testimonial under even a narrow standard”); see also Michigan v. Bryant, 562 U.S. 344, 358 (2011) (observing that “formal, out-of-court interrogation of a witness” is core Confrontation Clause concern).

        [61].     Crawford, 541 U.S. at 52.

        [62].     See Ohio v. Roberts, 448 U.S. 56, 66 (1980).

        [63].     See Ortman, supra note 35, at 457 (summarizing Roberts’s approach to confrontation).

        [64].     See Crawford, 541 U.S. at 62.

        [65].     See id.; see also Whorton v. Bockting, 549 U.S. 406, 420 (2007) (observing that it is “unclear whether Crawford, on the whole, decreased or increased the number of unreliable out-of-court statements that may be admitted in criminal trials”).

        [66].     Crawford, 541 U.S. at 61.

        [67].     Id. at 62.

        [68].     Id.

        [69].     The Confrontation Clause is understood to regulate the admissibility of confessions of people other than the defendant. See Bruton v. United States, 391 U.S. 123, 126 (1968) (holding that even with a limiting instruction, the government’s use at trial of the confession of a non-testifying codefendant violates a defendant’s confrontation right). This Article is about the admissibility of confessions offered against their makers.

        [70].     U.S. Const. amend. VI.

        [71].     See supra notes 51–53 and accompanying text.

        [72].     Nearly all confessions made during custodial interrogation qualify as testimonial. See Crawford, 541 U.S. at 52 (“Statements taken by police officers in the course of interrogations are also testimonial under even a narrow standard”). So do some noncustodial confessions, with the precise line of demarcation turning on the still-uncertain definition of testimonial. One sticking point is that the Court distinguishes interrogation for purposes of law enforcement, which leads to testimonial statements, from interrogation for purposes of emergency management, which does not. See Michigan v. Bryant, 562 U.S. 344, 358 (2011). Custodial interrogations that qualify for Miranda’s public safety exception thus might produce confessions that are exempt from the confrontation right. See New York v. Quarles, 467 U.S. 649, 655–56 (1984) (creating public safety exception).

        [73].     In some cases, such as where an oral confession was not recorded, self-confrontation could mean explaining that they didn’t say what the police claim they did.

        [74].     Defendants can attack their confessions without testifying, for instance by offering expert testimony on the unreliability of confession evidence. See Brandon L. Garrett, Contaminated Confessions Revisited, 101 Va. L. Rev. 395, 425–29 (2015). Courts are mixed in their receptivity to such expert evidence. Id. at 425–26.

        [75].     A study utilizing simulated trials found that recantations did little to temper participants’ estimate of whether defendants committed the crime. See generally Linda Henkel, Jurors’ Reactions to Recanted Confessions: Do the Defendant’s Personal and Dispositional Characteristics Play a Role?, 14 Psych., Crime & L. 565 (2008).

        [76].     See supra notes 67–68 and accompanying text.

        [77].     U.S. Const. amend. V (“No person . . . shall be compelled in any criminal case to be a witness against himself.”).

        [78].     Id. at amend. VI.

        [79].     Rock v. Arkansas, 483 U.S. 44, 52 (1987) (finding right to testify is “[l]ogically included in the accused’s right to call witnesses”).

        [80].     Peter Westen, The Future of Confrontation, 77 Mich. L. Rev. 1185, 1197 (1979); see also Randolph N. Jonakait, “Witnesses” in the Confrontation Clause: Crawford v. Washington, Noah Webster, and Compulsory Process, 79 Temp. L. Rev. 155, 172 (2006) (“The interrelationship of confrontation and compulsory process is particularly close.”).

        [81].     The term “witnesses” also appears in the Treason Clause of Article III, which provides that “No Person shall be convicted of Treason unless on the Testimony of two Witnesses to the same overt Act, or on Confession in open Court.” U.S. Const. art. III., § 3, cl. 1. That provision does appear to draw a distinction between defendants (who confess) and “Witnesses” (who give “Testimony”). On the other hand, it could be read so that a defendant who confessed out of court counts as one of the required “Witnesses.” That would render the term “Witnesses” in the Treason Clause consistent with Crawford’s understanding of a witness, but it’s an awkward reading. See Jonakait, supra note 80, at 167 (arguing that applying Crawford’s definition of witness to the Treason Clause “produces a redundancy”). I have found only one case considering the question, a district court ruling from 1861 that rejected the idea that a confession and one other witness is enough. See United States v. Greiner, 26 F. Cas. 36, 40 (E.D. Pa. 1861). Given the scarcity of Treason Clause case law and the specialized focus, it seems not especially important here.

        [82].     See infra notes 285–291 and accompanying text.

        [83].     There are two sub-scenarios: if Daniels raises the confession on their own, and if they do not. If Daniels raises the confession, they waive any objection to its admissibility under the Confrontation Clause. If they do not, the prosecutor will presumably wish to raise it on cross. Here we run into a minor complexity. Crawford contemplates that testimonial declarants testify and be subject to cross-examination by the defense lawyer. Thus, for Daniels’s confession to be admissible, they need to be subject to cross-examination about it by their lawyer. (That makes sense when one recalls that for purposes of the Confrontation Clause, Daniels is a prosecution witness.) If the prosecutor raises the confession in cross-examining Daniels during the defense case, their lawyer’s opportunity to confront the confession will come in a round of questioning conventionally labeled “redirect.” See Black’s Law Dictionary (11th ed. 2019) (defining “redirect examination”). Courts could deal with that in two ways. First, they could simply recognize that redirect counts as “cross-examination” for Sixth Amendment purposes. Second, they could allow the prosecutor to recall the defendant during the government’s rebuttal to ask about the confession, with the defense lawyer cross-examining Daniels about it in the ordinary course. See Robert P. Mosteller, Kenneth S. Broun, George E. Dix, Edward J. Imwinkelried, David H. Kaye, Eleanor Swift, McCormick on Evidence § 129 (“The loss of protection [against self-incrimination] has traditionally been regarded as effective throughout the trial in which the accused testifies.”). The first alternative seems less cumbersome and fully consistent with Crawford.

        [84].     Evidence derived from a confession is not the Confrontation Clause’s concern because the confrontation right belongs to criminal defendants, not suspects. That means that it’s impossible for police investigators to violate a person’s confrontation right. (Only prosecutors and judges can do that.)

        [85].     This Section considers the interaction of confrontation and self-incrimination during the prosecution’s case. For the related (but analytically distinct) question of their interaction during the defense case, see infra Part II.C.

        [86].     See Peter Westen, Incredible Dilemmas: Conditioning One Constitutional Right on the Forfeiture of Another, 66 Iowa L. Rev. 741, 753–54 (1981). A double constitutional conditions problem exists when the exercise of one constitutional right is conditioned on forfeiting a different constitutional right. Id.

        [87].     See McGautha v. California, 402 U.S. 183, 213 (1971) (“The threshold question is whether compelling the election impairs to an appreciable extent any of the policies behind the rights involved.”); see also Westen, supra note 86, at 757–58 (explaining that to analyze double constitutional conditions, courts should “shut, first, the left mind’s eye and focus on the constitutional interest underlying the option on the right (so to speak),” and then repeat the process on the other side).

        [88].     See infra Part II.B.2.

        [89].     See Orin S. Kerr, Compelled Decryption and the Privilege Against Self-Incrimination, 97 Tex. L. Rev. 767, 793 (2019) (observing that the “theoretical justification of the privilege is disputed territory”).

        [90].     See Stephen J. Schulhofer, Some Kind Words for the Privilege Against Self-Incrimination, 26 Val. U. L. Rev. 311, 311 (1991) (“It is hard to find anyone these days who is willing to justify and defend the privilege against self-incrimination.”). For classic articles attacking the privilege, see generally David Dolinko, Is There a Rationale for the Privilege Against Self-Incrimination?, 33 UCLA L. Rev. 1063 (1986); Henry J. Friendly, The Fifth Amendment Tomorrow: The Case for Constitutional Change, 37 U. Cin. L. Rev. 671 (1968). For a compelling defense of the privilege, see Schulhofer, supra, at 325–33. Scholars have also attempted to construct descriptive theories of the Court’s self-incrimination caselaw. E.g., William J. Stuntz, Self-Incrimination and Excuse, 88 Colum. L. Rev. 1227, 1228 (1988) (arguing that “there is an increasingly strong pattern that runs through a great deal of fifth amendment law” which “can be derived from notions of excuse”); Nita A. Farahany, Incriminating Thoughts, 64 Stan. L. Rev. 351, 365 (2012) (arguing that Stuntz’s excuse model “has great descriptive force, but only when used in conjunction” with evidence from neuroscience).

        [91].     Crawford v. Washington, 541 U.S. 36, 62 (2004).

        [92].     Giles v. California, 554 U.S. 353, 359 (2008). The Court described this as a forfeiture doctrine, but commentators have suggested that it’s better understood as waiver. See, e.g., Monica J. Smith, Goodbye Forfeiture, Hello Waiver: The Effect of Giles v. California, 13 Barry L. Rev. 137, 150 (2009).

        [93].     Giles, 554 U.S. at 374.

        [94].     Commonwealth v. Edwards, 830 N.E.2d 158, 168 (Mass. 2005) (emphasis added); see also People v. Reneaux, 50 Cal. App. 5th 852, 872–73 (2020) (analyzing cases that apply Giles’s definition of wrongdoing). The wrongful act need not be independently criminal. See Commonwealth v. Szerlong, 933 N.E.2d 633, 638 (Mass. 2010).

        [95].     Minnick v. Mississippi, 498 U.S. 146, 167 (1990) (Scalia, J., dissenting); see also I. Bennett Capers, Criminal Procedure and the Good Citizen, 118 Colum. L. Rev. 653, 668 (2018) (describing norm in Fifth Amendment cases that “good citizen[s]” should “not keep silent, notwithstanding any Fifth Amendment privilege against self-incrimination they may have”).

        [96].     See Richard D. Friedman, Forfeiture of the Confrontation Right After Crawford and Davis, 19 Regent U. L. Rev. 489, 493 (2007) (contending that “if the defendant has a privilege and keeps the witness off the stand by virtue of that privilege . . . there should not be a forfeiture”). Cf. Bordenkircher v. Hayes, 434 U.S. 357, 363 (1978) (“To punish a person because he has done what the law plainly allows him to do is a due process violation of the most basic sort . . . .”).

        [97].     To be sure, at the Supreme Court, the state abandoned its argument that Michael waived confrontation by asserting spousal privilege, so Justice Scalia formally “express[ed] no opinion” about it. Crawford v. Washington, 541 U.S. 36, 42 n.1 (2004). In rejecting the state’s waiver argument below, the Washington Supreme Court observed that “forcing the defendant to choose between the marital privilege and confronting his spouse presents an untenable Hobson’s choice.” State v. Crawford, 54 P.3d 656, 660 (Wash. 2002); see also State v. Jones, 984 N.E.2d 948, 972–73 (Ohio 2012) (conducting a Crawford inquiry into statements made by the spouse of the defendant and finding that she was unavailable after the defendant invoked spousal privilege). Courts have found that defendants forfeited confrontation by encouraging someone to refuse to testify. E.g., Cody v. Commonwealth, 812 S.E.2d 466, 482 (Va. App. 2018). But I am aware of no case in which a court ruled that confrontation was forfeited or waived based on the defendant invoking a valid privilege.

        [98].     Brown v. United States, 356 U.S. 148, 153–56 (1958).

        [99].     McGautha v. California, 402 U.S. 183, 215 (1971) (citing Spencer v. Texas, 385 U.S. 554, 561 (1967)).

      [100].     Id. at 214–17.

      [101].     Simmons v. United States, 390 U.S. 377, 394 (1968).

      [102].     Mitchell v. United States, 526 U.S. 314, 322–25 (1999).

      [103].     Brooks v. Tennessee, 406 U.S. 605, 610–12 (1972).

      [104].     Brown v. United States, 356 U.S. 148, 153–56 (1958).

      [105].     Id. at 156.

      [106].     Mitchell, 526 U.S. at 318–19.

      [107].     See id. at 321.

      [108].     Id. at 323.

      [109].     See Mirjan Damaska, Evidence Law Adrift 12 (1997) (observing that “extrinsic exclusionary rules” reject “probative information for the sake of values unrelated to the pursuit of truth”).

      [110].     Crawford v. Washington, 541 U.S. 36, 54 (2004).

      [111].     Crawford made numerous references to the year 1791. See id. at 46, 54, 55. At this point, I have a confession of my own: I am not an originalist. Rather, I’m making a doctrinal argument that’s grounded in an originalist decision. I’m relying on the common law circa 1791 because that’s what Crawford directs.

      [112].     Id. at 62; see also Giles v. California, 554 U.S. 353, 359 (2008) (observing that doctrine of forfeiture by wrongdoing “has roots in the 1666 decision in Lord Morley’s Case”).

      [113].     Crawford, 541 U.S. at 56 n.6 (noting that the “existence of [the dying declaration] exception as a general rule of criminal hearsay law cannot be disputed”).

      [114].     See, e.g., Giles, 554 U.S. at 358 (“[T]wo forms of testimonial statements were admitted at common law even though they were unconfronted.”); Aviva Orenstein, Her Last Words: Dying Declarations and Modern Confrontation Jurisprudence, 2010 U. Ill. L. Rev. 1411, 1440 (2010) (“Dying declarations seem to be one of only two exceptions to the rule announced in Crawford.”).

      [115].     Crawford, 541 U.S. at 45 (emphasis added).

      [116].     Readers may wonder whether the admissibility of confessions could be regarded as a manifestation of the principle that confrontation at trial wasn’t required at common law when the accused previously confronted the accuser, rather than as an exception. The logic of that view would be that suspects could personally confront themselves during their interrogations, just as they could personally confront third-party witnesses during their interviews. See Donald A. Dripps, Controlling the Damage Done by Crawford v. Washington: Three Constructive Proposals, 7 Ohio St. J. Crim. L. 521, 542 (2010) (explaining that at the founding, confrontation was accomplished personally, not through counsel). The principal reason why this view is unavailing is that self-confrontation in the interrogation room is very different from either confrontation of other witnesses during their interviews or self-confrontation at trial. When a defendant confronts themself at trial, they tell the jury that an earlier version of themself was, for understandable reasons, untruthful. See supra notes 73–76 and accompanying text. That can’t happen in the interrogation room at the very moment that a suspect accuses themself. Beyond that, common law courts justified the admissibility of confessions on precisely the same reliability grounds that they justified the admissibility of dying declarations, which we understand as exceptions to the confrontation right. Compare R v. Warickshall, 168 Eng. Rep. 234, 235 (K.B. 1783) (“A free and voluntary confession is deserving of the highest credit . . . and therefore it is admitted as proof of the crime to which it refers”) with R v. Woodcock, 168 Eng. Rep. 352, 353 (1789) (“[T]he general principle on which [dying declarations are] admitted is . . . the mind is induced by the most powerful considerations to speak the truth . . . .”).

      [117].     Crawford, 541 U.S. at 54; see also Giles, 554 U.S. at 358–59 (holding that the forfeiture by wrongdoing exception has the same limitations it did when the Sixth Amendment was adopted).

      [118].     R v. Warickshall, 168 Eng. Rep. 234, 235 (K.B. 1783).

      [119].     See, e.g., John H. Langbein, The Origins of Adversary Criminal Trial 218–33 (2005); George C. Thomas & Richard A. Leo, Confessions of Guilt: From Torture to Miranda and Beyond 40–54 (2012); J. M. Beattie, Crime and the Courts in England, 1660–1800 365–66 (1986); Lawrence Herman, The Unexplored Relationship between the Privilege Against Compulsory Self-Incrimination and the Involuntary Confession Rule (Part I), 53 Ohio St. L.J. 101, 147–62 (1992).

      [120].     See Bram v. United States, 168 U.S. 532, 543 (1897). The confessions rule proved to be a poor fit for the Self-Incrimination Clause, see Mark A. Godsey, Rethinking the Involuntary Confession Rule: Toward a Workable Test for Identifying Compelled Self-Incrimination, 93 Calif. L. Rev. 465, 485–88 (2005), and Bram was widely ignored, see Ronald J. Allen, Miranda’s Hollow Core, 100 Nw. U. L. Rev. 71, 74 (2006). Eventually, the Court rejected it. See Arizona v. Fulminante, 499 U.S. 279, 285 (1991) (“[Bram] does not state the standard for determining the voluntariness of a confession.”). Jeffrey Bellin argues that by locating the confessions rule in the privilege against self-incrimination, Bram erroneously removed it from the realm of ordinary evidence law, where it could check unreliable confessions. See Jeffrey Bellin, The Evidence Rules That Convict the Innocent, 106 Cornell L. Rev. 305, 335–36 (2021). Because Fifth Amendment protections against unreliable confessions fell away, Bellin explains, Bram led to modern law’s laissez-faire approach to confession evidence. See id. at 337 (“Once entangled with the constitutional inquiry, what remained of the common law’s reliability-focused evidence rule slowly bent toward process-focused constitutional doctrine, and then vanished entirely.”).

      [121].     See 2 Matthew Hale, History of the Pleas of the Crown 283 (1736) (noting that “it must be testified” that a defendant who confessed upon examination “did it freely without any menace, or undue terror imposed upon him.”). There is some dispute over whether this passage was written by Hale, who died in 1676, or by his eighteenth-century editor, Sollom Emlyn. Herman attributed the relevant text to Emlyn. Herman, supra note 119, at 148 & n. 262. Thomas and Leo respond, I think persuasively, that Hale was the author. Thomas & Leo, supra note 119, at 42–43, 248–49 n. 24.

      [122].     See Thomas & Leo, supra note 119, at 140–43 (discussing pre-1740 cases).

      [123].     Tobias Isaacs & Rachel Isaacs, Old Bailey Sessions Papers (July 1740, # 324–25), at 193–94, https://www.oldbaileyonline.org/record/17400709 [https://perma.cc/3UH6-EKWB].

      [124].     Id.; see also Langbein, supra note 119, at 220 (“Suddenly, in July 1740, the confession rule pops out.”); Thomas & Leo, supra note 119, at 43 (“[T]he remedy for a confession induced by fear would change, beginning with the 1740 case Tobias and Rachel Isaacs.”).

      [125].     See Langbein, supra note 119, at 221.

      [126].     Id.

      [127].     Id. at 233.

      [128].     See id. at 230–32.

      [129].     Thomas & Leo, supra note 119, at 50.

      [130].     Id. at 51.

      [131].     Id. at 52.

      [132].     The same historians disagree about exactly when it reached mature form. Langbein argues that by the 1760s, “exclusion had become a rule of law, no longer subject to judicial discretion.” Langbein, supra note 119, at 222. Thomas and Leo, on the other hand, suggest that the rule was not settled until the 1780s. See Thomas & Leo, supra note 119, at 47–48. The rule may have lagged behind a few years in America, though we lack adequate colonial court records to say so with confidence. See Herman, supra note 119, at 165; see also Thomas & Leo, supra note 119, at 67–69. At least some early American courts treated fear- and hope-based interrogation methods as grounds for juries to reject confessions, rather than as reason to withhold them from juries entirely. See Herman, supra note 119, at 165; e.g., Commonwealth v. Dillon, 4 Dall. 116, 117 (Pa. 1792) (admitting confession despite “irregular” pressure upon suspect). Contemporaries, moreover, attributed acquittals to the fact that a confession was induced by fear or hope, suggesting that judges may have cautioned juries not to trust such confessions or even directed not guilty verdicts. See, e.g., 1 Newspaper Reports of Decisions in Colonial, State, and Lower Federal Courts Before 1801 150 (Stanton D. Krauss ed., 2018) (reprinting contemporaneous newspaper account of 1797 Georgia case: “[B]ut this confession appearing to have been made while the prisoner was under the impressions of fear, and the flatteries of hope; the jury found the prisoner not guilty.”); id. at 503 (same for 1791 Massachusetts case: “But, fortunately for the Prisoners, the Jury were convinced that this confession of Hilton was not fairly obtained; but that it was drawn from him through fear or persuasion: Their Verdict was therefore NOT GUILTY.”). In any event, an exclusionary rule corresponding to the British rule emerged by the first decade of the nineteenth century. See Herman, supra note 119, at 166–68; see also Thomas & Leo, supra note 119, at 69. This timeline corresponds neatly to the historical evidence cited in Crawford on whether a non-testifying accuser’s deposition could be offered against a defendant without a prior opportunity for cross-examination. In support of the proposition that it could not (which was central to its holding), the Court cited (1) English authorities mostly pre-dating 1791 and (2) American authorities post-dating 1791, nearly all of which were from the nineteenth century. Compare Crawford v. Washington, 541 U.S. 36, 46–47 (2004) (citing British authorities) with id. at 49–50 (citing American authorities).

      [133].     2 William Hawkins, A Treatise of the Pleas of the Crown c. 46, § 3 n. 2 (6th ed. 1787) (Thomas Leach, ed.). Leach added the quoted language to the treatise. See Thomas & Leo, supra note 119, at 8.

      [134].     Samuel Cheshum & James Sherrard, Old Bailey Sessions Papers (Jan. 1788), at 171–72, https://www.oldbaileyonline.org/record/t17880109-38?text=cheshum [https://perma.cc/T3TG-J2LC].

      [135].     Id.

      [136].     R v. Jacob Thompson, 1 Leach’s Cr. C. 291 (1783).

      [137].     Id. at 293.

      [138].     Joshua Coster & Peeling Herne, Old Bailey Sessions Papers (Jan. 1774), at 86, https://www.oldbaileyonline.org/record/t17740112-44 [https://perma.cc/XN4C-6RLV].

      [139].     Thomas Ashby, et al., Old Bailey Sessions Papers (Sep. 1773), at 353–55, https://www.oldbaileyonline.org/record/17730908 [https://perma.cc/6KWF-25UZ]. The cases cited in this and the preceding note are examples of what Langbein calls “crown witness confessions,” in which a witness confesses in hopes of “persuading the authorities to designate him as the crown witness, with the result that he would escape prosecution.” Langbein, supra note 119, at 223. Such confessions were a “remarkable blind spot” for the confessions rule, as judges permitted them to be used against the confessor until the 1770s. Id. at 223–25. In the 1770s, however, Langbein explains, “the tension between the confession rule and the crown witness competition was squarely acknowledged and resolved in favor of exclusion under the confession rule.” Id. at 225–26.

      [140].     Hawkins, supra note 133, at c. 46, § 3 n.2; see also Thomas & Leo, supra note 119, at 53 (noting that Leach “captures the Whig concern with human autonomy”).

      [141].     Thomas & Leo, supra note 119, at 53.

      [142].     See 2 Wayne R. LaFave, Jerold H. Isreal, Nancy J. King, and Orin S. Kerr, Crim. Proc. § 6.2(a) (4th ed.) (“A formal rule of exclusion appears to have been first stated in 1783 in the case of R v. Warickshall.”); see also Wesley MacNeil Oliver, Magistrates’ Examinations, Police Interrogations, and Miranda-Like Warnings in the Nineteenth Century, 81 Tul. L. Rev. 777, 787 (2007) (“An accident of history, however, led nineteenth-century judges and treatise writers to trace the origins of the rule to a 1783 case, R v. Warickshall, heard at the Justice Hall of the Old Bailey in London.”).

      [143].     R v. Warickshall, 168 Eng. Rep. 234, 234 (K.B. 1783).

      [144].     Id.

      [145].     See supra note 133 and accompanying text. There is a third circumstance in which a testimonial confession would be admissible against its maker, but it would arise so rarely that I omit it in the text. Crawford recognized that if a defendant had a prior opportunity to confront a testimonial declarant, and the declarant is unavailable, the Confrontation Clause poses no barrier to prosecutorial use of the statement. Crawford v. Washington, 541 U.S. 36, 54–56 (2004). So, if a defendant (1) previously confronted (i.e., testified about) their confession, (2) invokes their right not to testify at a subsequent trial, and (3) can introduce their prior testimony at the subsequent trial, then the prosecutor would not be barred from using the confession.

      [146].     Cf. Thomas Y. Davies, Farther and Farther from the Original Fifth Amendment: The Recharacterization of the Right Against Self-Incrimination as a Trial Right in Chavez v. Martinez, 70 Tenn. L. Rev. 987, 1020 (2003) (describing common law confession rule as a “nonconstitutional doctrine”).

      [147].     Under the Marian Committal Statute of 1555, as John Langbein explains, “when a suspect was apprehended and brought to the [justice of the peace], the JP should, before committing him to pretrial detention, examine him and his accusers about the charges.” Langbein, supra note 119, at 41. On the qualifications of justices of the peace, Langbein notes that they were “substantial citizens who rendered unpaid parttime service in law enforcement,” and were “seldom lawyers, seldom legally trained.” Id. at 46. The same basic structure existed in colonial America. See Eben Moglen, Taking the Fifth: Reconsidering the Origins of the Constitutional Privilege Against Self-Incrimination, 92 Mich. L. Rev. 1086, 1095 (1994) (explaining that “[w]ithout exception, administrators of criminal justice in British North America made use of the process established by the Marian committal statute,” and the “nonprofessional character of the JP was assumed a fortiori in North America”). On thieftakers, see Ruth Paley, Thief-takers in London in the Age of the McDaniel Gang, c. 1745–1754, in Policing and Prosecution in Britain 1750–1850, at 311, 315–17 (Douglas Hay & Francis Snyder eds., 1989). As for crime victims, specifically household employers, as interrogators, see Langbein, supra note 119, at 230–31.

      [148].     Lawrence M. Friedman, Crime and Punishment in American History 67–68 (1993) (“One of the major social inventions of the first half of the nineteenth century was the creation of police forces: full-time, night-and-day agencies whose job was to prevent crime, to keep the peace, and to capture criminals.”).

      [149].     On police training for interrogation, see Margareth Etienne & Richard McAdams, Criminogenic Risks of Interrogation, 98 Ind. L.J. 1031, 1037–38 (2023). On interrogation rooms, see Dan Simon, In Doubt: The Psychology of the Criminal Justice Process 134 (2012) (“Interrogations are conducted in specially designed rooms that are small, windowless, and secluded.”)

      [150].     The text focuses on substance, but there are procedural questions lurking as well. Courts would need to receive evidence about whether a particular interrogation induced fear or hope. That could happen outside the jury’s presence during trial, but the better practice would be pre-trial evidentiary hearings on motions in limine. Either the state could move to admit a confession on the grounds that the interrogation did not involve hope or fear, or the defendant could move to exclude it (if they did not testify) on the grounds that it did.

      [151].     In the “false friend” technique, “the interrogator represents that he is a friend acting in the suspect’s best interest.” LaFave et al., supra note 142, at § 6.2(c).

      [152].     See Isaacs, supra note 123, at 324.

      [153].     See infra notes 180–187 and accompanying text.

       154.      See Crawford v. Washington, 541 U.S. 36, 68 (2004).

      [155].     See Etienne & McAdams, supra note 149, at 1037–38 (noting the Reid technique’s “dominant position in the field”); Alan Hirsch, Going to the Source: The “New” Reid Method and False Confessions, 11 Ohio St. J. Crim. L. 803, 803 (2014) (similar).

      [156].     See Etienne & McAdams, supra note 149, at 1037–38.

      [157].     Fred E. Inbau, John E. Reid, Joseph P. Buckley & Brian C. Jayne, Criminal Interrogation and Confessions 4–5 (5th ed. 2013).

      [158].     Id. at 5.

      [159].     Id. at 5, 192, 310; see also Simon, supra note 149, at 135 (“The evidence may well be genuine, but in its absence, interrogators often fabricate it and deceive the suspect into believing that it exists.”).

      [160].     See Etienne & McAdams, supra note 149, at 1035–36 (“The confrontational method prominently includes the tactics known as maximization and minimization.”); Saul M. Kassin, The Psychology of Confessions, 4 Annu. Rev. Law. Soc. Sci. 193, 200 (2008) (observing that “police employ a number of tactics that involve using a combination of negative and positive incentives.”).

      [161].     Saul M. Kassin & Karlyn McNall, Police Interrogations and Confessions: Communicating Promises and Threats by Pragmatic Implication, 15 Law & Hum. Behav. 233, 234–35 (1991).

      [162].     See Simon, supra note 149, at 135 (“The implicit message [of maximizing strategies] is that the suspect is bound to be convicted even absent a confession . . . .”); Steven A. Drizin & Richard A. Leo, The Problem of False Confessions in the Post-DNA World, 82 N.C. L. Rev. 891, 911 (2004) (“Because it is designed to break the anticipated resistance of an individual who is presumed guilty, police interrogation is stress-inducing by design; it is intentionally structured to promote isolation, anxiety, fear, powerlessness, and hopelessness.”).

      [163].     Kassin & McNall, supra note 161, at 235.

      [164].     See Drizin & Leo, supra note 162, at 912 (noting that Reid interrogations use “positive incentives (i.e., tactics that suggest the suspect will in some way feel better or benefit if he confesses, such as appealing to the suspect’s self-interest or minimizing the seriousness of the offense)”); see also Etienne & McAdams, supra note 149, at 1036–37 (distinguishing legal and moral minimization).

      [165].     Inbau et al., supra note 157, at 191.

      [166].     Id. at 202; see also Simon, supra note 149, at 135 (“The effect of minimization is usually achieved by means of presenting the suspect with a theme that reduces the import of the crime.” (emphasis in original)).

      [167].     Inbau et al., supra note 157, at 243.

      [168].     Id. at 279–80.

      [169].     Id. at 278–88, 291.

      [170].     E.g., id. at 203 (“If the theme present threats of inevitable consequences coupled with promises of leniency, it could jeopardize the validity of the confession.”); see also Etienne & McAdams, supra note 149, at 1054 (noting that the manual “warns interrogators against making promises of leniency in exchange for a confession”).

      [171].     Inbau et al., supra note 157, at 345.

      [172].     See Simon, supra note 149, at 135 (noting that “experiments show that minimizing themes are understood by lay people as implicit promises of leniency”); Saul M. Kassin, Allison D. Redlich, Fabiana Alceste & Timothy J. Luke, On the General Acceptance of Confessions Research: Opinions of the Scientific Community, 73 Am. Psych. 63, 72 (2018) (reporting that 91 percent of expert psychologist survey respondents endorsed as reliable the proposition that “minimization leads people to infer leniency upon confession”).

      [173].     Inbau et al., supra note 157, at 213.

      [174].     Id. at 415–30.

      [175].     See Simon, supra note 149, at 135 (“[T]he Reid Method advocates the use of interrogative techniques . . . which have been . . . deemed permissible by the Court.”); see also Etienne & McAdams, supra note 149, at 1046 n.87 (finding “only three state cases in which courts suppressing a confession recognized that minimization was one relevant factor in a totality of the circumstances test for voluntariness”). See also infra Part II.B.1.

      [176].     Some, but not all. Richard Leo analyzed the prevalence of twenty-five techniques in interrogations conducted by two police departments. See Richard A. Leo, Inside the Interrogation Room, 86 J. Crim. L. & Criminology 266, 278 (1996). Many tactics Leo observed leveraged hope or fear, but many did not, or at least did not necessarily do so, including “confront[ing] suspect[s] with existing evidence of guilt,” the second most-used tactic, which appeared in 85 percent of interrogations, and “identify[ing] contradictions in suspect’s story” (42 percent of interrogations). See id.

      [177].     In perhaps the only Supreme Court decision to apply the confessions rule, Bram v. United States, 168 U.S. 532 (1897), the Court had no difficulty finding that hope was induced by implication, rather than directly. See id. at 564–65. For more on Bram, see supra note 120.

      [178].     Before we leave the confessions rule, one more detail warrants exploration. Custodial interrogations today are ordinarily preceded by a Miranda warning. See infra notes 320–324 and accompanying text. Do Miranda warnings bear on the confessions rule analysis? It depends on whether similar warnings were given at common law and, if so, what legal effect they had. Justices of the peace gave warnings to suspects as early as the late eighteenth century. See Oliver, supra note 142, at 785; Robert Davidson, Old Bailey Sessions Papers (Nov. 1796), at 80–82, https://www.oldbaileyonline.org/record/t17961130-61?text=davidson [https://perma.cc/ZM5T-XGKT]. From the second decade of the nineteenth century, treatises reported that justices of the peace had a duty to inform suspects that they were not required to incriminate themselves. See Thomas & Leo, supra note 119, at 57; e.g., 1 J. Chitty, A Practical Treatise on the Criminal Law 85 (1816); 1 William Dickinson, A Practical Exposition of the Law Relative to the Office and Duties of a Justice of the Peace 457 (1813). Later in the nineteenth century, the legal consequences of warnings came into focus, as treatises related that they could dissipate the taint of improper interrogation. See Oliver, supra note 142, at 789. But there are two major caveats. First, the treatises provided that a proper warning could undo the effect of prior appeals to hope or fear, such that post-warning confessions would be admissible. See Simon Greenleaf, A Treatise on The Law of Evidence 288 (1888) (“[W]here an inducement has been held out . . . but the prisoner is subsequently warned by the magistrate . . . his confession, afterwards made, will be received as a voluntary confession.”); Esek Cowen, Notes to Phillipps’ Treatise on the Law of Evidence 429 (1850) (“[I]mproper influence once exerted may yet be countervailed by a subsequent influence, so as to make an after confession receivable.”). I have found no evidence that warnings immunized subsequent manipulative tactics, as Miranda warnings do. See infra note 332 and accompanying text. Second, warnings were only one factor courts looked to in determining whether a confession was induced by hope or fear. A warning made it “more likely” that a confession would be admitted, Oliver, supra note 142, at 789–90, but warning or not, courts allowed confessions only if it was clear that the effects of the inducement had completely abated. See, e.g., People v. Robertson, 1 Wheel. Crim. Cas. 66 (N.Y. Gen. Term 1822); Edward Hyde East, 2 A Treatise of the Pleas of the Crown 658 (1806) (describing 1800 case of Richard Harvey).

                  Miranda warnings thus have limited bearing on how courts should analyze contemporary confessions under the confessions rule. In an unusual case where investigators first induced a confession, then thoroughly reversed the effects of the inducement, and then obtained an unmanipulated confession, the warning might matter. In a case where interrogators perfunctorily read Miranda and then induced a confession with manipulative interrogation tactics, the warning would be neither here nor there.

      [179].     See supra note 150 and accompanying text.

      [180].     See supra note 151 and accompanying text.

      [181].     Such cases would simply be modern versions of the Isaacs case discussed above. See supra notes 123–124 and accompanying text.

      [182].     Cf. LaFave et al., supra note 142, at § 6.2(c) (describing the false friend situation “in which by deception the defendant is made unaware that the person with whom he is conversing is a police officer or police agent”).

      [183].     Inbau et al., supra note 157, at 272.

      [184].     Id.

      [185].     See Miranda v. Arizona, 384 U.S. 436, 452 (1966).

      [186].     See David Abney, Mutt and Jeff Meet the Constitution, 22 Crim. L. Bull. 118, 118–19 (1986) (“The conclusion is that this interrogation technique is permissible as long as it is used moderately and not combined with oppressive ancillary methods.”); see also LaFave et al., supra note 142, at § 6.2(c) (noting “dictum in the Miranda case [that] was critical of the ‘Mutt and Jeff’ routine,” but observing that “courts have not generally disapproved of the police giving ‘friendly’ advice to the defendant or expressing sympathy for him.”).

      [187].     See infra Part II.B.1 (explaining the limited reach of due process and self-incrimination jurisprudence as applied to interrogations and confessions). For a cogent argument that courts should already be wary of the method, however, see Welsh S. White, Police Trickery in Inducing Confessions, 127 U. Pa. L. Rev. 581, 625–28 (1979).

      [188].     I indirectly addressed two already—that defendants cannot simultaneously invoke confrontation and self-incrimination, see supra Part I.B, and that Crawford is amenable to a categorical exception for confessions offered against their makers, see supra Part I.C.

      [189].     See infra Part II.A.

      [190].     See infra Part II.B.

      [191].     See infra Part II.C.

      [192].     See infra Part II.D. Still more anticipated objections are addressed in notes elsewhere in the Article. See supra notes 116 and 178 and infra note 361.

      [193].     See, e.g., United States v. DeLeon, 287 F. Supp. 3d 1187, 1252 n.17 (D.N.M. 2018) (“Perez’ testimonial statements . . . are admissible against Perez notwithstanding the Confrontation Clause, because Perez does not have the right to confront himself.”); Brown v. State, 2008 WL 2152557, *5 (Ind. Ct. App. May 23, 2008) (similar); see also United States v. Allen, 10 F.3d 405, 413 (7th Cir. 1993) (similar but pre-Crawford).

      [194].     See Fed. R. Evid. 801(d)(2) (“[A statement] is not hearsay . . . [if it is] offered against an opposing party and[] was made by the party in an individual or representative capacity.”); Clifford S. Fishman & Anne Toomey McKenna, Jones on Evidence § 27:7 (7th ed.) (noting equivalent state evidence codes).

      [195].     See, e.g., United States v. Spencer, 592 F.3d 866, 879 (8th Cir. 2010) (“The tape recordings do not implicate Crawford. Derrick’s statements on the tape are ‘admissions by a party-opponent’ and are admissible on that basis.”); United States v. Debrew, 2010 WL 11523673, at *9 (D.N.M. Jan. 26, 2010) (similar); United States v. Andrews, 2007 WL 1749221, at *5 (N.D. Ind. June 15, 2007) (similar); State v. Johnson, 771 N.W.2d 360, 368 (S.D. 2009) (similar); Brown, 2008 WL 2152557, at *5 (similar); People v. Thompson, 2005 WL 3304096, at *1 (Mich. Ct. App. Dec. 6, 2005) (similar).

      [196].     454 F.3d 660 (7th Cir. 2006).

      [197].     Id. at 665 (citation omitted).

      [198].     Nor does any other opinion that I have found. Summers reached the same conclusion. See Summers, supra note 25, at 1806–07.

      [199].     602 U.S. 779 (2024).

      [200].     See Fed. R. Evid. 801(d)(2) advisory committee’s note (“Admissions by a party-opponent are excluded from the category of hearsay on the theory that their admissibility in evidence is the result of the adversary system rather than satisfaction of the conditions of the hearsay rule . . . .”). The Advisory Committee believed that party admissions differ from other hearsay exceptions in that admissibility is not based on the statements being especially reliable. See Roger Park, The Rationale of Personal Admissions, 21 Ind. L. Rev. 509, 509 n.3 (1988). Yet reliability is not a plausible rationale for every hearsay exception in Rules 803 and 804. See, e.g., Fed. R. Evid. 804(b)(6) and advisory committee’s note (explaining that the exception for statements offered against a party who wrongfully caused the declarant’s unavailability aims to “deal with abhorrent behavior”).

      [201].     See United States v. Smith, 578 F.2d 1227, 1231 (8th Cir. 1978) (“The distinction between a statement which is not hearsay and a statement which is an exception to the hearsay rule is semantic . . . .”). The distinction between hearsay exemptions and exceptions is, moreover, excessively confusing. See Sam Stonefield, Rule 801(d)’s Oxymoronic “Not Hearsay” Classification: The Untold Backstory and a Suggested Amendment, 5 Fed. Cts. L. Rev. 1, 3 (2011) (describing history of complaints about the labels).

      [202].     George Fisher, Evidence 407 (3d ed. 2013).

      [203].     Smith, 602 U.S. at 785 (internal quotation omitted).

      [204].     Id. at 794. In a footnote, Smith marked one “qualification,” writing that “[i]f an evidentiary rule reflects a long-established understanding, then it might shed light on the historical meaning of the Confrontation Clause.” Id. at 794 n.4. The classification of party opponent statements as “nonhearsay” in modern evidence codes is hardly such a rule. See supra notes 200–202 and accompanying text. For the classification’s (twentieth-century) history, see Stonefield, supra note 201.

      [205].     Smith, 602 U.S. at 794.

      [206].     Crawford v. Washington, 541 U.S. 36, 59 n.9 (2004) (noting that the Confrontation “Clause also does not bar the use of testimonial statements for purposes other than establishing the truth of the matter asserted”). But see Jeffrey L. Fisher, The Truth About the ‘Not for Truth’ Exception to Crawford, 32 Champion 18, 19 (2008) (observing that a “casual reading of the Crawford dicta might suggest that anything goes when the prosecution posits a non-hearsay purpose for introducing an out-of-court testimonial statement”).

      [207].     I have found no good justifications for excluding confessions from confrontation analysis in scholarly commentary either. Aside from Summers, supra note 25, the most sustained treatment is in the Federal Rules of Evidence Manual (the “Manual”), authored by three eminent scholars:

[I]t is clear that if the statement is by the defendant himself, or is adopted by him, then it is the defendant’s own hearsay statement that is being admitted. Clearly, a person has no constitutional right to confront himself. Crawford does not change this rule, because even if the accused’s statement is testimonial, there is no right to cross-examine yourself. Applying Crawford literally would mean that a defendant who made a voluntary confession pursuant to police interrogation would have a right to exclude this “testimonial” statement under the Confrontation Clause. Clearly Crawford did not contemplate such a result.

Stephen A. Saltzburg, Michael M. Martin & Daniel Capra, Federal Rules of Evidence Manual § 801.02(3) (2023). As a description of where the courts are on this issue, the Manual is unimpeachable. And it avoids the mistake of invoking the distinction between hearsay exceptions and exemptions. But if this passage is meant to supply a rationale, it comes up short. Crawford, it says, “does not change” the “rule” that “a person has no constitutional right to confront himself” because “there is no right to cross-examine yourself.” But why not? The Manual leaves that puzzle to readers. And it cites only two cases in support of the “clear” proposition that there is no right to confront one’s own confession: United States v. Rollins, 862 F.2d 1282 (7th Cir. 1988), which predates Crawford by more than a decade, and United States v. Hansen, 434 F.3d 92 (1st Cir. 2006), which involved a defendant’s statements but was decided on other grounds. Other scholarly takes on this issue are even briefer than the Manual’s. See, e.g., Michael H. Graham, Winning Evidence Arguments § 801:16 (2023); Artem M. Joukov, Isn’t That Hearsay Anyway? How the Federal Hearsay Rule Can Serve as a Map to the Confrontation Clause, 63 Wayne L. Rev. 337, 350–51 (2018).

      [208].     See Culombe v. Connecticut, 367 U.S. 568, 602 (1961).

      [209].     See Miranda v. Arizona, 384 U.S. 436, 444 (1966).

      [210].     And the Fourteenth Amendment, for state prosecutions. See Malloy v. Hogan, 378 U.S. 1, 6 (1964) (incorporating Self-Incrimination Clause against states); Brown v. Mississippi, 297 U.S. 278, 287 (1936) (rejecting confession in state case on Fourteenth Amendment due process grounds).

      [211].     United States v. James Daniel Good Real Prop., 510 U.S. 43, 49 (1993).

      [212].     Id. (quoting Soldal v. Cook County, 506 U.S. 56, 70 (1992)).

      [213].     Graham v. Connor, 490 U.S. 386, 395 (1989). For an explanation and criticism of this exception, see generally Toni M. Massaro, Reviving Hugo Black? The Court's “Jot for Jot” Account of Substantive Due Process, 73 N.Y.U. L. Rev. 1086 (1998). As Massaro explains, Graham “violates the Court’s well-established interpretative principle that multiple constitutional claims may apply to a given scenario.” Id. at 1113.

      [214].     See Masaro, supra note 213, at 1087 (explaining that Graham’s “rationale was a disarmingly simple principle of statutory interpretation: an explicit textual provision ‘on point’—here, the Fourth Amendment—trumps any more general provisions of the Constitution”).

      [215].     See infra Part II.B.1.

      [216].     In Oregon v. Elstad, for instance, the Court identified the “trustworthiness” of evidence and “deterrence” of police misconduct as the doctrine’s two purposes. 470 U.S. 298, 308 (1985). The Court’s attention to reliability in Miranda cases is inconsistent. See Tonja Jacobi, Miranda 2.0, 50 U.C. Davis L. Rev. 1, 10 (2016) (“The central problem with Miranda is that it was not crafted specifically to prevent false confessions, but rather to regulate interrogations more generally.”).

      [217].     See Thomas & Leo, supra note 119, at 190 (“Miranda changed the interrogation world in ways that were dramatic and profound, but ultimately superficial.”); William J. Stuntz, Miranda’s Mistake, 99 Mich. L. Rev. 975, 976 (2001) (noting that Miranda’s “effects are probably small, perhaps vanishingly so”). But see Paul G. Cassell & Bret S. Hayman, Police Interrogation in the 1990s: An Empirical Study of the Effects of Miranda, 43 UCLA L. Rev. 839, 917 (1996) (“[S]ignificant concern remains that Miranda has seriously hampered the prosecution and conviction of criminals.”).

      [218].     Thomas & Leo, supra note 119, at 190 (noting high waiver rates); see also Primus, supra note 36, at 16–18 (explaining breadth of implied waiver doctrine).

      [219].     Charles J. Ogletree, Are Confessions Really Good for the Soul: A Proposal to Mirandize Miranda, 100 Harv. L. Rev. 1826, 1828 (1987); see also Laura Smalarz, Kyle C. Scherr & Saul M. Kassin, Miranda at 50: A Psychological Analysis, 25 Current Directions Psych. Sci. 455, 456 (2016) (explaining that “[e]xperiments using physiological and self-report data have demonstrated that individuals experience stress upon being accused of wrongdoing” that “can substantially undermine Miranda comprehension”).

      [220].     See Welsh S. White, Miranda’s Failure to Restrain Pernicious Interrogation Practices, 99 Mich. L. Rev. 1211, 1213 (2001) (“Once [Miranda] rights have been waived, the restraints Miranda imposes on police interrogators are minimal.”); see also Thomas & Leo, supra note 119, at 190 (observing that waiver is Miranda’s “critical moment”).

      [221].     See infra Part III.

      [222].     Culombe v. Connecticut, 367 U.S. 568, 602 (1961).

      [223].     Schneckloth v. Bustamonte, 412 U.S. 218, 225–26 (1973) (quoting Culombe, 367 U.S. at 602). On the vagueness of this test, see Yale Kamisar, What Is an Involuntary Confession? Some Comments on Inbau and Reid’s Criminal Interrogation and Confessions, 17 Rutgers L. Rev. 728, 746 (1962) (“The trouble with Culombe is not that it fails to achieve precision in the area—this is only a failure to do the impossible—but that it threatens to enhance existing uncertainty and confusion.”); see also Stephen J. Schulhofer, Confessions and the Court, 79 Mich. L. Rev. 865, 869 (1981) (“The vagueness of the voluntariness test spawned several interrelated difficulties for the courts.”).

      [224].     See Geoffrey R. Stone, The Miranda Doctrine in the Burger Court, 1977 Sup. Ct. Rev. 99, 102–03 (“Given the [limitations of the voluntariness doctrine], it seemed inevitable that the Court would seek some automatic device by which the potential evils of incommunicado interrogation [could] be controlled.” (internal quotations omitted)).

      [225].     Guha Krishnamurthi, The Case for the Abolition of Criminal Confessions, 75 SMU L. Rev. 15, 56 (2022).

      [226].     United States v. Jacques, 744 F.3d 804, 809 (1st Cir. 2014) (citations omitted).

      [227].     See Godsey, supra note 120, at 470 (“[T]he theoretical ambiguity inherent in the voluntariness standard leaves police officers with little guidance in the field.”).

      [228].     See LaFave et al., supra note 142, at § 6.2(c) (noting split of authority on the “question of what the result should be when a confession has been obtained in response to a police assertion that cooperation would facilitate prompt release on bail or would mean that the defendant would fare better in subsequent proceedings”).

      [229].     Jacques, 744 F.3d at 809–10.

      [230].     United States v. Harrison, 34 F.3d 886, 891–92 (9th Cir. 1994).

      [231].     Godsey, supra note 120, at 470.

      [232].     See Donald A. Dripps, Guilt, Innocence, and Due Process of Plea Bargaining, 57 Wm. & Mary L. Rev. 1343, 1368 (2016) (“Reliability, however, was the irreducible core of the voluntariness standard, even though the standard swept more broadly.”); Schulhofer, supra note 223, at 867 (“The early cases required exclusion of [involuntary] confessions primarily (and perhaps exclusively) because of their unreliability . . . .”).

      [233].     479 U.S. 157, 163 (1986).

      [234].     Id. at 164.

      [235].     See Garrett, supra note 74, at 423 (noting that “reliability is not a separate ground for excluding a confession under the U.S. Constitution”).

      [236].     Connelly, 479 U.S. at 167.

      [237].     See Lisa Kern Griffin, Silence, Confessions, and the New Accuracy Imperative, 65 Duke L.J. 697, 741 (2016) (noting that “fairness, voluntariness, and reliability overlap” and that while they are “different values, they often move in the same direction”).

      [238].     See Primus, supra note 36, at 33 (“After Connelly . . . [if] the police use interrogation tactics that they know or should know are likely to lead to a false confession [whether across cases or in the case at hand], they engage in precisely the type of ‘police overreaching’ that Connelly condemns.”); see also Welsh S. White, Miranda’s Waning Protections 199 (2001) (“[A]n interrogation method that is substantially likely to induce an untrustworthy statement should still be impermissible under the due process test.”).

      [239].     See Jackson v. Denno, 378 U.S. 368, 386 (1964) (noting the “probable unreliability of confessions that are obtained in a manner deemed coercive”); see also Joseph D. Grano, Voluntariness, Free Will, and the Law of Confessions, 65 Va. L. Rev. 859, 922 (1979) (“[T]actics that would overcome the person of ordinary firmness or that would, in a constitutional sense, take undue advantage of an accused also are likely to create an unnecessary risk of unreliability.”).

      [240].     Saul M. Kassin, False Confessions: How Can Psychology So Basic Be So Counterintuitive?, 72 Am. Psych. 951, 952 (2017) (“The notion that anyone of sound mind would confess to a crime he or she did not commit is not intuitive to the average person.”).

      [241].     See Garrett, supra note 74, at 399–415; Brandon L. Garrett, The Substance of False Confessions, 62 Stan. L. Rev. 1051, 1062–66 (2010) [hereinafter Garrett, False Confessions].

      [242].     Garrett, False Confessions, supra note 241, at 1052.

      [243].     Id. at 1099. As for the remaining twelve exonerees, nine pleaded guilty, one did not seek to suppress the statement, and two may have challenged their confessions but the records were unclear. Id. Every exoneree had waived Miranda. Id. at 1092.

      [244].     Garrett, supra note 74, at 402.

      [245].     See Garrett, False Confessions, supra note 241, at 1054 (“These forty cases cannot speak to how often people confess falsely.”).

      [246].     See Kassin, supra note 161, at 201 (describing “three essential elements” of “the practice of interrogation” that “if overused, may induce innocent persons to confess,” none of which generally defeat voluntariness).

      [247].     That false confessions can be easily manipulated has achieved “general acceptance” among psychological researchers. See generally Kassin et al., supra note 172.

      [248].     See generally Melissa B. Russano, Christian A. Meissner, Fadia M. Narchet & Saul M. Kassin, Investigating True and False Confessions Within a Novel Experimental Paradigm, 16 Psych. Sci. 481 (2005).

      [249].     Id. at 483.

      [250].     Id.

      [251].     Id. at 484.

      [252].     Id.

      [253].     Id.

      [254].     See generally Jessica R. Klaver, Zina Lee & V. Gordon Rose, Effects of Personality, Interrogation Techniques and Plausibility in an Experimental False Confession Paradigm, 13 L. & Crim. Psych. 71 (2008).

      [255].     Id. at 75–77.

      [256].     Id. at 77.

      [257].     Id. at 77–78.

      [258].     Id. at 77.

      [259].     Id. at 78.

      [260].     Id.

      [261].     Id.

      [262].     Id. at 79.

      [263].     Id. For a cautionary note on the external validity of these studies, see Christopher Slobogin, Manipulation of Suspects and Unrecorded Questioning: After Fifty Years of Miranda Jurisprudence, Still Two (or Maybe Three) Burning Issues, 97 B.U. L. Rev. 1157, 1183–85 (2017). Notwithstanding that concern, Slobogin notes that “where the laboratory and field research is convergent—say, with respect to the impact of minimization techniques or the combined impact of the false evidence ploy and lengthy interrogation,” evidence from the psychological studies should be admissible via expert witnesses. Id. at 1185.

      [264].     See supra notes 133–139 and accompanying text.

      [265].     See Richard A. Leo & Richard J. Ofshe, The Consequences of False Confessions: Deprivations of Liberty and Miscarriages of Justice in the Age of Psychological Interrogation, 88 J. Crim. L. & Criminology 429, 431 (1998) (“[N]o one knows precisely how often false confessions occur in the United States . . . .”). The statistic is unknowable, rather than just unknown, because the “ground truth (what really happened) may remain in genuine dispute even after a defendant has pled guilty or been convicted.” Id. at 431–32.

      [266].     Hilo averages 276 rainy days per year, the most of any U.S. city. See Jeanne Yacoubou, What is the Rainiest City in the US? Top 11 Wettest Cities, Homeowner (Oct. 16, 2024), https://www.homeowner.com/weather/rainiest-city-in-the-us [https://perma.cc/CT6E-L938].

      [267].     This isn’t to say that they contribute literally nothing. If manipulative tactics increase the rate of true confessions more than they increase the rate of false confessions, the confessions they produce have some epistemic value. See Slobogin, supra note 263, at 1163 n.34 (observing that, in lab studies, manipulative tactics increased both true and false confessions but had a larger effect on true confessions). But not much, and surely far less than factfinders give them. See Kassin, supra note 160, at 209 (describing psychological research); Saul Kassin, Why Confessions Trump Innocence, 67 Am. Psych. 431, 439 (2012).

      [268].     See Christian A. Meissner, Christopher E. Kelly & Skye A. Woestehoff, Improving the Effectiveness of Suspect Interrogations, 11 Ann. Rev. L. & Soc. Sci. 211, 215 (2015) (describing research finding that “perceived pressure to confess, intimidation by the interrogator, and fear associated with the consequences of not confessing were significantly higher for false confessions” while “true confessions were more likely to result from a desire to relieve distress and feelings of guilt”).

      [269].     On the reasons for the assumption, see supra notes 36–37 and accompanying text.

      [270].     See supra notes 233–236 and accompanying text; see also Brandon L. Garrett, Constitutional Law and the Law of Evidence, 101 Cornell L. Rev. 57, 97 (2015) (noting that since the 1960s, “the Court has disclaimed constitutional regulation of reliability of confessions”).

      [271].     See Pamela R. Metzger, Cheating the Constitution, 59 Vand. L. Rev. 475, 501 (2006) (“Practitioners and academics generally agree that the primary goal of the Confrontation Clause is to ensure the reliability of evidence.”). Reliability may not be the Confrontation Clause’s only objective. For arguments that confrontation serves dignitary interests, see Sherman J. Clark, An Accuser-Obligation Approach to the Confrontation Clause, 81 Neb. L. Rev. 1258, 1270–71 (2003); Toni M. Massaro, The Dignity Value of Face-to-Face Confrontations, 40 Fla. L. Rev. 863, 897–917 (1988).

      [272].     George Fisher, The Crawford Debacle, 113 Mich. L. Rev. First Impressions 17, 27–28. (2014).

      [273].     I have developed this point in more detail elsewhere. See Ortman, supra note 35, at 470–76. To be precise, confrontation is a “substantive rule of criminal procedure,” id. at 470 n. 128, but the additional verbiage seems unnecessary here.

      [274].     See Richard D. Friedman, Truth and Its Rivals in the Law of Hearsay and Confrontation, 49 Hastings L.J. 545, 545 (1998) (“In most settings, the law of hearsay should be concerned nearly exclusively with achieving the goal of truth determination, subject to constraints related to the cost of litigation.”).

      [275].     See Daniel Epps, The Consequences of Error in Criminal Justice, 128 Harv. L. Rev. 1065, 1073 (2015) (providing examples of rules that “tilt in favor of defendants and thus, at least in theory, in favor of false acquittals and against false convictions”: the burden of proof beyond a reasonable doubt, unanimous juries, and asymmetric access to appellate courts); Richard D. Friedman, Anchors and Flotsam: Is Evidence Law Adrift, 107 Yale L.J. 1921, 1940 (1998) (observing that exclusionary rules that find a “firm basis in protecting the criminal defendant”).

      [276].     See Friedman, supra note 274, at 550 (“The confrontation right . . . is [an] example of . . . a constraint that operates in derogation of truth.”).

      [277].     Crawford v. Washington, 541 U.S. 36, 61 (2004)

      [278].     Delaware v. Fensterer, 474 U.S. 15, 22 (1985). It is controversial whether such testing is actually a useful mechanism for identifying false accusations. Compare Dripps, supra note 116, at 535 (“Crawford . . . has made it significantly more difficult to convict the guilty, without improving the chances of vindicating the innocent.”), with Friedman, supra note 274, at 563 (“To allow testimony, whether given in court or not, to be presented to the factfinder without giving the accused the opportunity to confront the witness would deprive the accused of the testing that a trial is supposed to provide.”). The Court’s current view is that the Sixth Amendment resolves the question. See Dripps, supra note 116, at 539 (explaining that Crawford jurisprudence is based on “a supposed constitutional command” and not “a contemporary assessment of competing interests”).

      [279].     See supra notes 73–74 and accompanying text.

      [280].     See 4 William Blackstone, Commentaries on the Laws of England 352 (1769). I am (obviously) paraphrasing liberally, but Blackstone was suspicious of confessions: “[E]ven in cases of felony at the common law, [confessions] are the weakest and most suspicious of all testimony; ever liable to be obtained by artifice, false hopes, promises of favour, or menaces . . . .” Id. at 357.

      [281].     See supra Part I.B.

      [282].     See Rock v. Arkansas, 483 U.S. 44, 49 (1987) (holding that “it cannot be doubted that a defendant in a criminal case has the right . . . to testify in his or her own defense.”).

      [283].     See supra notes 73–74 and accompanying text.

      [284].     A related objection is that conditioning the admissibility of confession on defendants waiving the privilege against self-incrimination would discourage defendants from testifying, thus depriving juries of information. Cf. Jeffrey Bellin, Improving the Reliability of Criminal Trials Through Legal Rules that Encourage Defendants to Testify, 76 U. Cin. L. Rev. 851, 881 (2008) (arguing that when legal rules discourage defendants from testifying, the “justice system suffers because it is deprived of the information defendants would provide”). The objection’s premise is certainly right. Under this Article’s reading of Crawford, defendants who confessed under psychologically manipulative interrogation would need a very good reason to take the stand at trial. Surely, most would not. In that sense, the rule bears some similarity to other rules that permit the government to use certain evidence only against testifying defendants, including the rules allowing the prosecution to impeach testifying defendants with their prior convictions or with their statements to the police obtained in violation of Miranda. See id. at 863–71. The major difference between those rules and the one proposed here is that manipulated confessions are currently admissible against all defendants, not just those who testify. Besides, juries are not deprived of information merely by the existence of rules that make testimony costly but because those rules co-exist with the defendant’s privilege not to testify.

      [285].     557 U.S. 305 (2009). The definitive work on the Confrontation Clause as imposing a duty of production is Pamela R. Metzger, Confrontation as a Rule of Production, 24 Wm. & Mary Bill Rts. J. 995 (2016).

      [286].     557 U.S. at 308–09.

      [287].     Id. at 324.

      [288].     Id. at 324–25. The Court had a second reason for rejecting Massachusetts’s argument: It would “shift[] the consequences of adverse-witness no-shows from the State to the accused.” Id. at 324. That obviously does not apply when the declarant is the defendant.

      [289].     See Metzger, supra note 285, at 1026 (“The Melendez-Diaz majority stressed that prosecutorial witness production was a non-negotiable mandate of the Confrontation Clause.”).

      [290].     See supra notes 277–278 and accompanying text.

      [291].     The objection would reimagine confrontation as a tool for a defendant to use in developing their case, but that’s what the Compulsory Process Clause is for. See Melendez-Diaz, 557 U.S. at 313–14 (“The text of the [Sixth] Amendment contemplates two classes of witnesses—those against the defendant and those in his favor. The prosecution must produce the former; the defendant may call the latter.”); Richard D. Friedman, Confrontation: The Search for Basic Principles, 86 Geo. L.J. 1011, 1037 (1998) (explaining that interpreting the Confrontation Clause in this way “would render [it] virtually superfluous, because that Clause would only be, in effect, the flip side of the Compulsory Process Clause”).

      [292].     The Supreme Court recently confirmed that Melendez-Diaz’s production rule is current law. Smith v. Arizona, 602 U.S. 779, 792 n.3 (2024). There had been some question about that because Justice Alito’s opinion for a four-justice plurality in Williams v. Illinois, 567 U.S. 50 (2012), contained language inconsistent with it. See id. at 58–59 (“This conclusion will not prejudice any defendant who really wishes to probe the reliability of the DNA testing done in a particular case because those who participated in the testing may always be subpoenaed by the defense and questioned at trial.”). Justice Thomas, however, who supplied the fifth vote for the Court’s judgment in Williams, explicitly endorsed the rule. Id. at 117 n.6 (Thomas, J., dissenting) (“Melendez–Diaz rejected this reasoning . . . holding that the defendant’s subpoena power is no substitute for the right of confrontation.” (internal quotations omitted)). In Smith, the Court put any lingering doubts about the production rule to rest. The state court in Smith had seemingly grounded its ruling in part on the idea that Smith could not complain about the government’s use of statements by an absent laboratory analyst because he “did not independently seek to subpoena the analyst.” Smith, 602 U.S. at 792 n.3 (quoting certiorari petition). In a portion of her opinion garnering seven votes, Justice Kagan wrote that the Court “need not spend much time on that rationale because the State rightly does not defend it.” Id. Melendez-Diaz, she explained, had settled this. Id.

      [293].     To be clear, the production rule predates Melendez-Diaz. See Metzger, supra note 285, at 1003–22 (describing production rule’s rise in the nineteenth century through incorporation and its fall between incorporation and Crawford).

      [294].     See Emily Spottswood, Truth, Lies, and the Confrontation Clause, 89 U. Colo. L. Rev. 565, 606 (2018) (explaining that “unconfronted testimonial hearsay is generally excluded even when the witness’s unavailability is no one’s fault” and arguing that “statements fitting into this category” should “be freely admissible against the defendant”).

      [295].     See David Alan Sklansky, Confrontation and Fairness, 45 Tex. Tech L. Rev. 103, 109 (2012) (“If we defined the right to confrontation with an eye to its underlying purpose of ensuring a fair trial, it is unlikely we would make it turn on [the government’s reasons] for interviewing a witness who is now, through no fault of the government, unavailable.”).

      [296].     See supra note 97 and accompanying text.

      [297].     Sklansky argues that confrontation jurisprudence’s inattention to fairness is for the worse. See Sklansky, supra note 295, at 109–11.

      [298].     See, e.g., Fed. R. Evid. 609; see also Jeffrey Bellin, The Silence Penalty, 103 Iowa L. Rev. 395, 398 (2018) (explaining that “the most concrete deterrent to testifying is a product of the evidence rules concerning prior crimes”).

      [299].     Schulhofer, supra note 90, at 330.

      [300].     Id.

      [301].     Crawford held that live confrontation at trial is not required if the defendant had a prior opportunity to confront the declarant and the declarant is unavailable. See 541 U.S. 36, 59 (2004) (stating that “[t]estimonial statements of witnesses absent from trial [are admissible] only where the declarant is unavailable, and only where the defendant has had a prior opportunity to cross-examine”). In this scenario, the declarant is unavailable by virtue of asserting their privilege against self-incrimination at trial, and the written statement arguably counts as a prior confrontation. To be sure, written confrontation would obviously not suffice for other out-of-court accusers. But defendants are (perhaps) different in this respect. A defendant’s self-confrontation could be a declaration explaining the circumstances of the interrogation in narrative form. For other testimonial declarants, written confrontation could only be done with interrogatories, which are not an effective substitute for cross-examination. That’s why civil discovery features both interrogatories and depositions. See Andrew J. Ruzicho, Louis A. Jacobs & Andrew J. Ruzicho II, § 7:11 Interrogatories, in 1 Litigating Age Discrimination Cases (Aug. 2024 update) (explaining that some matters are “far better suited to the conversational narrative of a deposition than the concise, written nature of an interrogatory answer”). In this sense, cross-examination, whether at or before trial, simply matters more for confronting others than for confronting oneself. There is obviously much more to say about this issue, which I leave for future analysis.

      [302].     Those are useful in a world in which “originalists are now dominating the federal judiciary and directing the exercise of our law.” Reva B. Siegel, Memory Games: Dobbs’s Originalism as Anti-Democratic Living Constitutionalism—and Some Pathways for Resistance, 101 Tex. L. Rev. 1127, 1197 (2023).

      [303].     See Stephanos Bibas, Originalism and Formalism in Criminal Procedure: The Triumph of Justice Scalia, the Unlikely Friend of Criminal Defendants?, 94 Geo. L.J. 183, 192 (2005) (describing Crawford as “a successful blend of originalism and formalism”).

      [304].     See supra Part I.C.1.

      [305].     The objections correspond, at least roughly, to the original intention and original meaning schools. See Lawrence B. Solum, What Is Originalism? The Evolution of Contemporary Originalist Theory, in The Challenge of Originalism: Theories of Constitutional Interpretation 12, 16–24 (Grant Huscroft & Bradley W. Miller eds., 2011) (placing schools in historical context); Larry Alexander, Constitutional Theories: A Taxonomy and (Implicit) Critique, 51 San Diego L. Rev. 623, 638 (2014) (distinguishing the approaches).

      [306].     Holscher, supra note 31; Criminal defendants did talk during their trials, but not under oath. In the eighteenth century “accused speaks” model of the criminal trial, the “accused’s merged roles as defender and witness were inextricable,” such that “the accused found himself constantly speaking about his role in events.” Langbein, supra note 119, at 13–14.

      [307].     See generally Rock v. Arkansas, 483 U.S. 44 (1987) (holding that criminal defendants have a constitutional right to testify in their own defense).

      [308].     See Randy E. Barnett, The Gravitational Force of Originalism, 82 Fordham L. Rev. 411, 412 (2013) (“[B]ecause new constitutional cases typically involved factual situations unknown to the Framers, old originalists typically engaged in a hypothetical and counterfactual inquiry into how the Framers would have addressed an issue had they thought of it.”).

      [309].     See supra Part I.C.1.

      [310].     Alexander, supra note 305, at 639.

      [311].     See Stephanos Bibas, The Limits of Textualism in Interpreting the Confrontation Clause, 37 Harv. J.L. & Pub. Pol’y 737, 741 (2014) (“It is not clear that there was a preexisting concept of what ‘confrontation’ meant in a criminal trial.”); Margaret A. Berger, The Deconstitutionalization of the Confrontation Clause: A Proposal for a Prosecutorial Restraint Model, 76 Minn. L. Rev. 557, 562 (1992) (“Little information exists about precisely what the concept of confrontation signified in the seventeenth and eighteenth centuries in England and the colonies.”); see also California v. Green, 399 U.S. 149, 173–74 (1970) (Harlan, J., concurring) (“[T]he Confrontation Clause comes to us on faded parchment.”).

      [312].     See David Alan Sklansky, Anti-Inquisitorialism, 122 Harv. L. Rev. 1634, 1644 (2009) (“[T]he Supreme Court has long interpreted the Confrontation Clause to guarantee . . . cross-examination of prosecution witnesses by defense counsel in front of the jury.”).

      [313].     Crawford v. Washington, 541 U.S. 36, 45-47 (2004) (“One recurring question was whether the admissibility of an unavailable witness’s pretrial examination depended on whether the defendant had had an opportunity to cross-examine him”).

      [314].     See id. at 45; see also supra Part I.C.1.

      [315].     See Hawkins, supra note 133, at c. 46, § 3; see also supra note 133 and accompanying text.

      [316].     1 Geoffrey Gilbert, The Law of Evidence 216 (London 1791); see also Crawford, 541 U.S. at 45 (citing Gilbert).

      [317].     Michael W. McConnell, Time, Institutions, and Interpretation, 95 B.U. L. Rev. 1745, 1761 (2015).

      [318].     See Clark, supra note 271, at 1269 (“[G]iven that the Framers left us little or no direct indication of what the confrontation right meant to them, even the most thorough-going originalist would have to acknowledge the necessity for us to craft our own understandings.”). On the purposes of the Confrontation Clause, see generally supra notes 277–278 and accompanying text.

      [319].     See Miranda v. Arizona, 384 U.S. 436 (1966). On Miranda’s fame, see Frederick Schauer, The Miranda Warning, 88 Wash. L. Rev. 155, 155 (2013) (observing that “Miranda v. Arizona may well be the most famous appellate case in the world”).

      [320].     Miranda, 384 U.S. at 458 (“Unless adequate protective devices are employed to dispel the compulsion inherent in custodial surroundings, no statement obtained from the defendant can truly be the product of his free choice.”); see also Donald A. Dripps, About Guilt and Innocence 80 (2003) (“Police interrogation in general, [Chief Justice] Warren wrote, constitutes compulsion.”).

      [321].     See 384 U.S. at 467–77.

      [322].     See Oregon v. Elstad, 470 U.S. 298, 306 (1985) (indicating that the “Miranda exclusionary rule . . . sweeps more broadly than the Fifth Amendment itself” and “may be triggered even in the absence of a Fifth Amendment violation”).

      [323].     See Primus, supra note 36, at 3 (“In the last decade . . . the Roberts Court has gutted criminal suspects’ Miranda . . . protections.”).

      [324].     One observer sees the Court’s decision in Vega v. Tekoh, 597 U.S. 134 (2022), as a shot across Miranda’s bow. See Leading Case, Vega v. Tekoh, 136 Harv. L. Rev. 430, 430 (2022).

      [325].     Of course, different people are disappointed for different reasons. See Stuntz, supra note 217 at 975–76.

      [326].     Many of the postmortems are cited in the footnotes to LaFave et al., supra note 142, at § 6.5(d).

      [327].     Stuntz, supra note 217, at 977.

      [328].     Id. at 982.

      [329].     Id.

      [330].     See supra note 218 and accompanying text.

      [331].     See Roseanna Sommers & Vanessa K. Bohns, The Voluntariness of Voluntary Consent: Consent Searches and the Psychology of Compliance, 128 Yale L.J. 1962, 2015 (2019) (“[W]arnings are ineffective because they fail to address the psychology of compliance.”); Donald A. Dripps, Against Police Interrogation—And the Privilege Against Self-Incrimination Supreme Court Review, 78 J. Crim. L. & Criminology 699, 722 (1988) (“The Miranda majority never explained how a suspect could make a ‘knowing and intelligent waiver’ while under the influence of the same inherently compelling environment that justified the warnings.”).

      [332].     See Allen, supra note 120, at 75 (“[After Miranda] things went on more or less as before, with the primary difference that the police henceforth had to recite the warnings before obtaining waivers and proceeding to the interrogation.”); Richard A. Leo & K. Alexa Koenig, The Gatehouses and Mansions: Fifty Years Later, 6 Ann. Rev. L. & Soc. Sci. 323, 334 (2010) (“Once issued and waived, Miranda neither regulates nor restricts psychologically manipulative or deceptive interrogation techniques . . . .”).

      [333].     See Stephen J. Schulhofer, Miranda’s Practical Effect: Substantial Benefits and Vanishingly Small Social Costs, 90 Nw. U. L. Rev. 500, 561 (1996) (“Miranda’s stated objective was not to eliminate confessions, but to eliminate compelling pressure in the interrogation process.”).

      [334].     Miranda v. Arizona, 384 U.S. 436, 448 (1966).

      [335].     Id. at 448–49 n.9 (citing Fred E. Inbau & John E. Reid, Criminal Interrogation and Confessions (1st ed. 1962)); see also Albert W. Alschuler, Constraint and Confession, 74 Denv. U. L. Rev. 957, 971 (1997) (noting that Criminal Interrogation and Confessions “was exhibited in the Miranda opinion like a relic from a medieval torture chamber”).

      [336].     384 U.S. at 448–55.

      [337].     Id. at 449 (“The officers are told by the manuals that the ‘principal psychological factor to a successful interrogation is privacy—being alone with the person under interrogation.’”).

      [338].     Id. at 450 (“[T]he manuals instruct the police to display an air of confidence in the suspect’s guilt and from outward appearance to maintain only an interest in confirming certain details.”).

      [339].     Id. (“The officers are instructed to minimize the moral seriousness of the offense, to cast blame on the victim or on society.”).

      [340].     Id. at 453 (“The interrogators sometimes are instructed to induce a confession out of trickery.”).

      [341].     Id. at 457.

      [342].     Id. at 456.

      [343].     See Stephen J. Schulhofer, Reconsidering Miranda, 54 U. Chi. L. Rev. 435, 436 (1987).

      [344].     Id. at 444–45.

      [345].     Id. at 445.

      [346].     Id. at 446 (“Custodial interrogation brings psychological pressure to bear for the specific purpose of overcoming the suspect’s unwillingness to talk, and it is therefore inherently compelling within the meaning of the fifth amendment.”); see also Miranda, 384 U.S. at 457 (“It is obvious that such an interrogation environment is created for no purpose other than to subjugate the individual to the will of his examiner.”).

      [347].     It may have had vices as well, of course. For critical takes on the compulsion/coercion distinction, see Albert W. Alschuler, Miranda’s Fourfold Failure, 97 B.U. L. Rev. 849, 851 n.10 (2017); Godsey, supra note 120, at 503–04; Joseph D. Grano, Miranda’s Constitutional Difficulties: A Reply to Professor Schulhofer, 55 U. Chi. L. Rev. 174, 182–83 (1988). Alschuler, for instance, points out that in ordinary language there is no relevant distinction between interrogation tactics that “compel” speech and tactics that “coerce” it. Alschuler, Miranda’s Fourfold Failure, supra, at 851 n.10; see also Albert W. Alschuler, A Peculiar Privilege in Historical Perspective: The Right to Remain Silent, 94 Mich. L. Rev. 2625, 2627 n.6 (1996). For better or worse, neither Miranda nor Schulhofer grounded their account of compulsion in ordinary language. See Schulhofer, supra note 333, at 551 (“The English language explains some of the confusion here, because terms like coercion and compulsion have virtually interchangeable meanings . . . .”).

      [348].     Schulhofer, supra note 343, at 445; see also Louis Michael Seidman, Brown and Miranda, 80 Calif. L. Rev. 673, 720 (1992) (“[L]iberal justifications for punishment have also relied on a second argument from consent . . . : state coercion can be reconciled with individual autonomy precisely because the state does not compel consent to the punishment.”).

      [349].     To be sure, the Court has not consistently maintained the distinction. See Schulhofer, supra note 343, at 440 (noting decisions conflating compulsion and voluntariness); see also Kyron Huigens, Custodial Compulsion, 99 B.U. L. Rev. 523, 525 (2019) (“From the beginning, the Supreme Court has analyzed Miranda cases using ‘coercion’ and ‘compulsion’ . . . interchangeably.”). Such confusion is on display in the Court’s most recent Miranda ruling, Vega v. Tekoh, 597 U.S. 134 (2022). To support the conclusion that Miranda violations are not “necessarily” violations of the Fifth Amendment privilege against self-incrimination, the Court asserted that the Miranda “opinion suggested that there might not have been any actual Fifth Amendment violations in the four cases that were before the Court.” Id. at 143. Miranda, however, suggested that there may not be a due process violation in the cases. See Miranda, 384 U.S. at 457 (“[W]e might not find the defendants’ statements to have been involuntary in traditional terms.”).

      [350].     See sources cited supra note 216. To be sure, that a particular interrogation tactic is likely to lead to false confessions can sometimes make it abusive and thus fair game for due process analysis. See supra note 238 and accompanying text.

      [351].     For the same reason, the Court cannot really be faulted for minimizing that psychological interrogation tactics may yield false confessions. This possibility was relegated to a footnote. 384 U.S. at 455 n.24 (“Interrogation procedures may even give rise to a false confession.”).

      [352].     See supra notes 41–68 and accompanying text.

      [353].     See supra notes 246–263 and accompanying text.

      [354].     See supra notes 241–244 and accompanying text.

      [355].     See supra notes 344–355 and accompanying text.

      [356].     See Godsey, supra note 120, at 500 (“[T]he Court in Miranda undeniably interpreted the term ‘compelled’ very broadly.”).

      [357].     See Dripps, supra note 320, at 81 (“The only explanation for the inconsistency in the Miranda rules is concern for effective law enforcement.”); Seidman, supra note 348, at 742 (noting that the “elimination of confessions altogether” was “clearly a politically unacceptable result”).

      [358].     See sources cited supra note 332.

      [359].     See supra Part I.C.2.

      [360].     See supra text accompanying note 264.

      [361].     I argued above that the admissibility of confessions at common law should be regarded as an exception to the confrontation right, rather than as a manifestation of the prior confrontation rule. See supra note 116. I’m wrong about that; there’s a plausible argument that a confrontation-based confession law would threaten interrogation generally. (Thanks to Don Dripps for bringing the argument to my attention.) Since the founding era, confrontation has shifted from something done by defendants to something done by defense counsel. See Dripps, supra note 116, at 542. If confessions are admissible based on “prior confrontation” in the interrogation room, then perhaps only prior confrontation in the interrogation room by defense counsel suffices today. Lawyers almost invariably instruct their clients to invoke the Fifth Amendment in interrogation, so that would mean a serious curtailment of interrogation. See Dripps, supra note 320, at 75 (describing “incompatibility of defense counsel and police interrogation”). That said, the premise of this view—that confessions are admitted on a prior confrontation logic—is mistaken for reasons I explained above. See supra note 116. The proposal thus does not demand lawyers in interrogation rooms, though that might be an excellent idea for other reasons. See Ogletree, supra note 219, at 1830.

      [362].     As noted, while I think Crawford erred in limiting confrontation to the trial setting, I accept the Court’s view for purposes of this Article. See supra note 35.

      [363].     E.g., Schneckloth v. Bustamonte, 412 U.S. 218, 219 (1973) (“It is . . . well settled that one of the specifically established exceptions to the requirements of both a warrant and probable cause is a search that is conducted pursuant to consent.”).

      [364].     See William T. Pizzi, Waiver of Rights in the Interrogation Room: The Court’s Dilemma, 23 Conn. L. Rev. 229, 244–45 (1991) (suggesting that suspect cannot intelligently “waive his sixth amendment right to a jury trial” at a police station); see also Patterson v. Illinois, 487 U.S. 285, 304–06 (1988) (justifying more stringent requirements for waiver of counsel at trial than in pretrial questioning).

      [365].     Some argue that judges should be present for pre-trial questioning of defendants. See Dripps, supra note 331, at 730 n.120 (collecting citations).

      [366].     The text hedges on whether Miranda should have relied on confrontation instead of or in addition to self-incrimination. It depends, at least in part, on whether one finds Miranda’s self-incrimination analysis persuasive. Note that a version of Miranda based solely on confrontation would result in an interrogation regime close to Stuntz’s ideal, where “everyone talks—or, at the least, one in which everyone submits to questioning—but where police tactics are effectively regulated.” Stuntz, supra note 217, at 977.

      [367].     Because Crawford’s exclusionary rule does not cover the fruits of a manipulated confession, investigators could continue interrogating as usual, hoping that derivative evidence will materialize. See supra note 84 and accompanying text. That would be very risky.

      [368].     See Inbau et al., supra note 157, at 344 (arguing that the “only way” to “persuade a suspect whom the investigator believes to be lying” to “tell the truth” is “by allowing the suspect to believe that he will benefit in some way by telling the truth”); see also Miranda v. Arizona, 384 U.S. 436, 517 (Harlan, J., dissenting) (1966) (“We do know that some crimes cannot be solved without confessions . . . .”).

      [369].     See John A. MacKerron III, Miranda: Crime, Law and Politics, by Liva Baker, 35 Hastings L.J. 551, 557 n.44 (1984) (book note) (“The post-Miranda debate has often focused on whether the Miranda rule handcuffs police and creates another ‘loophole’ that allows criminals to be let loose on society.”); see generally Paul G. Cassell & Richard Fowles, Handcuffing the Cops? A Thirty-Year Perspective on Miranda’s Harmful Effects on Law Enforcement, 50 Stan. L. Rev. 1055 (1998).

      [370].     See Etienne & McAdams, supra note 149, at 1085 (“Broadly speaking, the alternative to the accusatory method is the information-gathering method.”). On the history and development of the PEACE technique, see Colin Clarke & Rebecca Milne, National Evaluation of the PEACE Investigative Interviewing Course 1–3 (Home Office, Report No: PRAS/149, 2001). PEACE is a mnemonic, with each letter standing in for one step in the method’s process: P, “planning and preparation,” E, “engage and explain,” A, “account,” C, “closure,” and E, “evaluation.” Id. at 1. The investigator completes the first step (P) prior to the interrogation and the last step (E) after it is complete. The remaining steps constitute the interrogation itself. Id. at 2.

      [371].     See Mary Schollum, Bringing PEACE to the United States: A Framework for Investigative Interviewing, The Police Chief 30 (Nov. 2017); Lisa Fallon, Brent Snook, Todd Barron, Angela Baker, Mike Notte, Jeff Stephenson & Dan Trotter, Evaluating the Vermont State Police’s PEACE Model Training Program: Phase 1, 28 Psych. Crime & L. 59, 63 (2022).

      [372].     See Christian A. Meissner, Allison D. Redlich, Stephen W. Michael, Jacqueline R. Evans, Catherine R. Camilletti, Sujeeta Bhatt & Susan Brandon, Accusatorial and Information-Gathering Interrogation Methods and Their Effects on True and False Confessions: A Meta-Analytic Review, 10 J. Experimental Criminology 459, 461–63 (2014).

      [373].     See id. at 462–63 (“Only after suspects have been given a full opportunity to provide information are they questioned and presented with any inconsistencies/contradictions [e.g., information known to the interviewer but not yet revealed to the suspect].”); see also Clarke & Milne, supra note 370, at 36–37 (explaining that after “encouraging the interviewee to give their account of events,” the investigator “breaks this account into small topic areas which are examined in a logical sequence to elicit more detailed information”).

      [374].     Etienne & McAdams, supra note 149, at 1085.

      [375].     A secondary question concerns the comparative expense. There is reason to think that the PEACE technique may “work better” than the Reid technique because it requires more skilled (and thus more expensive) investigators. See Dave Walsh & Ray Bull, What Really is Effective in Interviews with Suspects? A Study Comparing Interviewing Skills Against Interviewing Outcomes, 15 Legal & Criminological Psych. 305, 309 (2010) (“The PEACE model . . . requires a higher domain of skills in obtaining an accurate account of what happened than, say, those employed in trying to trick guilty [or otherwise] suspects into confessions.”).

      [376].     See Meissner et al., supra note 372, at 460–61 (defining diagnosticity in interrogation).

      [377].     See Slobogin, supra note 263, at 1183–84 (observing that the “generalizability of [lab] findings to warned criminal defendants charged with serious crimes and potentially subject to imprisonment has been called into question”); Meissner et al., supra note 372, at 463 (explaining that “only experimental studies offer a sound perspective on the diagnostic value of an interrogative method—field studies cannot distinguish the ‘ground truth’ that is necessary to assess the accuracy of a confession or the culpability of a suspect”).

      [378].     See Etienne & McAdams, supra note 149, at 1086 (“The existing social science research is unable to provide a definitive answer, although the existing results favor the information-gathering model.”).

      [379].     Meissner et al., supra note 372, at 481.

      [380].     See id. at 482 (urging caution until more studies “refine and solidify our understanding of the effects of various interrogative methods in eliciting truth and false confessions”); see also Slobogin, supra note 267, at 1162 (“More importantly, the jury is still out as to how effective these third-generation [interrogation] tactics are.”).

      [381].     I’m not assuming that information-gathering interrogations produce more false confessions, as that seems quite implausible. See Slobogin, supra note 263, at 1163 (observing that it is “likely true” that “third-generation techniques are less likely to cause false confessions”).

      [382].     See Williams v. Florida, 399 U.S. 78, 113–14 (1970) (Black, J., dissenting) (“The Framers decided that the benefits to be derived from the kind of trial required by the Bill of Rights were well worth any loss in ‘efficiency’ that resulted.”)

      [383].     See Stephen J. Schulhofer, Bashing Miranda Is Unjustified—and Harmful, 20 Harv. J.L. & Pub. Pol’y 347, 348 (1997) (“[D]id the Framers think that the Fifth Amendment would not make law enforcement more difficult? Of course not.”); see also California v. Acevedo, 500 U.S. 565, 601 (1991) (Stevens, J., dissenting) (“Even if the warrant requirement does inconvenience the police to some extent, that fact does not distinguish this constitutional requirement from any other procedural protection secured by the Bill of Rights.”).

      [384].     See Kassin, supra note 160, at 209 (describing psychological research).

      [385].     See Saul Kassin, Why Confessions Trump Innocence, 67 Am. Psych. 431, 439 (2012) (“Perhaps an even more dramatic effect on process concerns the possibility that false confessions undermine a defendant’s opportunity to get his or her proverbial day in court.”).

      [386].     See generally Joseph P. Fishman, Creating Around Copyright, 128 Harv. L. Rev. 1333, 1359 (2015) (“Since the early 1990s, however, empirical evidence has been mounting that, contrary to the freer-is-better account, certain forms of constraint in fact enhance creativity.”).