The Crawford Debacle

George Fisher*

First a toast—to my colleague Jeff Fisher and his Crawford[1]compatriot, Richard Friedman, on the tenth anniversary of their triumph: What they achieved in Crawford is every lawyer’s dream. By dint of sheer vision and lawyerly craft, they toppled what many saw as a flawed confrontation-law regime and put in its place one that promised greater justice. For that, much applause is due.

Still there’s no denying their doctrine’s a muddle, if not as conceived, then as realized. Consider the count: Four justices almost agree on Crawford’s contours but patch over the issues that divide them. A fifth justice defends the doctrine but scrimps on its scope. And the other four seek every chance to slip this listing ship and swim to dry land. After ten years and eight major rulings and mounting confusion on the Court, it’s time to reset and reassess: How might we have avoided this mess?

I. Some Common Ground

To plot our differences, it’s useful first to survey common ground. No one disputes, first of all, the critical importance of a criminal defendant’s “right . . . to be confronted with the witnesses against him.” All agree the Confrontation Clause grants defendants the right to cross-examine those who testify against them in court. And almost everyone assumes that the clause extends further—that it barsthe prosecution from introducing at least some hearsay unless the declarant appears in court for cross-examination or, if the declarant is unavailable at trial, the defendant had the chance to cross-examine the declarant before.

Yet almost no one reads the right as barring all hearsay offered against criminal defendants lacking the chance to cross-examine the declarant. Several sorts of hearsay survive almost every Confrontation Clause analysis. Chief among these are classic business records—not the records of crime labs working in league with prosecutors but those of commercial or nonprofit entities kept routinely and without criminal prosecution in view. Public records kept in similar circumstances likewise evade objection. With somewhat less confidence I’ll add dying declarations, made without hope of survival and with death close at hand and telling how death came about. And fourth I’ll include statements made in the throes of danger, desperately, when the declarant’s safety hangs on the listener’s aid. Other hearsay statements survive scrutiny under one or another conception of the Confrontation Clause, but I think these four categories are common ground and therefore a starting point for analysis.

There are in contrast two sorts of hearsay that all believe pose dangers. Take first the hearsay at issue in Crawford—an accomplice’s custodial statement offered to inculpate the accused. The Supreme Court has subjected such statements, “motivated by a desire to curry favor with the authorities,” to “special suspicion.”[2] Even more troubling are statements of child sexual-abuse victims, which the Supreme Court has agreed to address this term.[3] The frequency and gravity of sex crimes against children, together with children’s fragile memories and psyches, make their hearsay accusations a critical test of any confrontation-law regime. A wise regime would address these two forms of hearsay cautiously, generally excluding blame-shifting statements of accomplices and distinguishing those children’s statements that can stand on their own from those that, absent the child’s testimony, must stay out, scuttling the case.

It’s common ground, then, that some rule or standard must distinguish those hearsay statements meeting no Confrontation Clause objection from those the clause bars absent the defendant’s chance to cross-examine the declarant. The nature of that rule or standard—its provenance, constitutional justification, and contours at the margins—is the nub of Confrontation Clause controversy. It’s what divides the current justices into three contingents and separates the Crawford regime from the preceding regime of Ohio v. Roberts[4] and from other imaginable confrontation-law schemes.

Yet even here, in choosing among these schemes, we can find several points in common. First we must start with text. Perhaps the clause’s words admit too much ambiguity to guide us, but no analysis will prove credible that doesn’t start there. We must be guided as well by principle. Textualism is one principle; historical fidelity is another; deference to just results and the means to achieve them is yet another. The choice among them is fraught, but without some guiding principle no scheme can claim respect. And whatever the principle, it must yield predictable results in recurring situations and, ideally, in emerging ones too.

By all these criteria I believe Crawford and the law regime it launched have fallen short. Let me say why—and then suggest a fix.

II. Crawford’s Flawed Originalism

Writing for the Court in Crawford, Justice Scalia sought to divine the scope of the Confrontation Clause’s command. Rightly he began with text. Here he confessed—again correctly—the text’s hopeless ambiguity: “One could plausibly read ‘witnesses against’ a defendant to mean those who actually testify at trial, those whose statements are offered at trial, or something in-between.” Pages later, however, Justice Scalia teased from the text the meaning he had confessed was lacking: “The text of the Confrontation Clause . . . applies to ‘witnesses’ against the accused—in other words, those who ‘bear testimony.’ ” Here Justice Scalia cited the 1828 edition of Noah Webster’s American Dictionary of the English Language, which in turn defined testimony as “[a] solemn declaration or affirmation made for the purpose of establishing or proving some fact.” Hence a witness against the defendant, Justice Scalia concluded, can be “[a]n accuser who makes a formal statement to government officers” but not “a person who makes a casual remark to an acquaintance.” The former’s out-of-court statement would be testimonial hearsay; the latter’s would not.

Nowhere in this seemingly simple bit of lexicology did Justice Scalia confess that Webster had supplied not one but five definitions of witness and that Justice Scalia had selected the fifth—“One who gives testimony; as, the witnesses in court . . . .” Webster’s first and fourth entries clearly did not apply to witness as used in the Confrontation Clause. His second definition—“That which furnishes evidence or proof”—possibly applied to documents or physical artifacts, not persons. But what’s wrong with Webster’s third definition? It defined a witness as “[a] person who knows or sees any thing; one personally present; as, he was witness; he was an eye-witness.” As Professor Jonakait[5] and others have noted, this definition of witness suggests nearly all hearsay—not merely Justice Scalia’s narrow class of “testimonial” hearsay—falls under the clause’s command. Because almost all hearsay declarants knew or saw something, the Sixth Amendment could bar virtually all hearsay offered against criminal defendants unless they can cross-examine the declarant.

Justice Scalia acknowledged this uncertainty—but not in Crawford. Dissenting in Maryland v. Craig fourteen years before, he argued that Webster’s broader definition of witness could not be what the Framers had in mind. That definition is “excluded in the Sixth Amendment by the words following the noun: ‘witnesses against him.’ The phrase obviously refers to those who give testimony against the defendant at trial.”[6] But as Professor Shaviro asks, “Why cannot the term ‘witnesses against him’ refer to all persons having knowledge about the case and whose statements reporting such knowledge the prosecution uses as evidence against the defendant?”[7] At bottom, then, the Confrontation Clause’s text cannot sustain Justice Scalia’s distinction between testimonial hearsay, generally excluded unless the defendant can cross-examine the declarant, and nontestimonial hearsay, regulated only by evidence rules.

Nor does the clause’s history support this distinction. Consider the 1603 prosecution of Sir Walter Raleigh, which figures so prominently in the Crawford Court’s analysis. It’s true that Sir Walter fumed against admitting Lord Cobham’s confession accusing Raleigh of joining with Cobham in treason. “[L]et Cobham be here,” Raleigh cried. “[L]et him speak it. Call my accuser before my face . . . .”[8] It’s also true that by any of the many definitions of testimonial hearsay appearing in the Crawford canon, Cobham’s accusation would qualify.

But what of the other notoriously rank hearsay used to condemn Sir Walter? For all the attention the Crawford Court lavished on Raleigh’s complaints about his absent accuser Cobham, it paid none at all to Raleigh’s second absent accuser—the unnamed “gentleman” whom the witness Dyer, a boat pilot, encountered while visiting a merchant’s house in Lisbon. On hearing Dyer was English, the Portuguese gentleman asked if the King was crowned. “I answered, No,” Dyer testified, “but that I hoped he should be so shortly. Nay, saith [the gentleman], he shall never be crowned; for Don Raleigh and Don Cobham will cut his throat ere that day come.”[9] Here followed perhaps the trial’s most memorable moment—Raleigh’s outraged cry, “This is the saying of some wild Jesuit or beggarly Priest; but what proof is it against me?” A saying in the day’s legal argot was an unsworn statement[10]—here unsworn hearsay. Ignoring Raleigh’s charge, the attorney general rejoined coldly, “It . . . shows that your treason had wings.”[11]

Memorable as this scene was, it would have unsettled Justice Scalia’s carefully wrought historical argument. The Crawford Court focused on Raleigh’s trial to make a point—that the historical concern underlying the Confrontation Clause was admission of accusations made in formalized ex parte affidavits, the Court’s paradigm of testimonial hearsay. Dyer’s account of the Portuguese gentleman’s words decidedly did not fit this mold. The distant gentleman had made no “solemn declaration or affirmation . . . for the purpose of establishing or proving some fact.” Rather he made his accusation to a boat pilot in a private, unrecorded conversation far from the nearest English court. It was “a casual remark to an acquaintance,” the clearest kind of nontestimonial evidence. Yet Raleigh railed in outrage against its admission—and I know of no observer in Raleigh’s time or since who has suggested his outrage was misplaced.

Indeed I know of no observer in Raleigh’s day or the next three centuries who distinguished between testimonial and nontestimonial hearsay. Professor Davies sought such a distinction in the framing era and found none. In the day’s jargon, he writes, all unsworn statements—hence almost all hearsay—were simply “no evidence.” He notes Justice Scalia “did not identify any framing-era source that distinguished between testimonial and nontestimonial hearsay. So far as I can tell, none did.” And framing-era authorities surely did not view admission of nontestimonial hearsay against criminal defendants with complacency. On the contrary, Davies argues, the Framers never contemplated the matter “because they never anticipated that informal hearsay statements could come to be viewed as valid evidence in criminal trials.”[12]

Here Chief Justice Marshall lends Davies support. Presiding at Aaron Burr’s 1807 trial, the Chief Justice excluded a claimed coconspirator’s statement, hearsay Crawford deemed nontestimonial. In what appears to be the first confrontation interpretation by a Supreme Court justice, he wrote, “I know not why . . . a man should have a constitutional claim to be confronted with the witnesses against him, if mere verbal declarations, made in his absence, may be evidence against him.”[13] A “mere verbal declaration[]” sounds like what Crawford would call nontestimonial hearsay. Unlike the Crawford Court, Chief Justice Marshall seemingly believed admitting such hearsay offended the Confrontation Clause.

It may well be, as Justice Scalia wrote in Crawford, that “the principal evil at which the Confrontation Clause was directed was the . . . use of ex parte examinations as evidence against the accused.” But to sustain his distinction between testimonial and nontestimonial statements, Justice Scalia must show that nontestimonial hearsay was received without complaint. This he has not done. Raleigh’s outrage at the boatman Dyer’s testimony and Chief Justice Marshall’s rejection of a nontestimonial statement—together with the utter lack of citations to framing-era cases admitting “casual remark[s] to an acquaintance” against an accused[14]—suggest the Court’s testimonial/nontestimonial distinction found no footing in the framing era. Even Robert Kry, Justice Scalia’s law clerk when Crawford was decided and one of few prominent defenders of his originalist arguments, lends no support here. In a long article rebutting Davies, Kry offers no defense of the Court’s claim that the Framers and lawyers of their era accepted admission of nontestimonial hearsay.[15]

Nor did lawyers of the founding era always object to admission of testimonial hearsay. As Justice Scalia confessed in Crawford, there was authority for admitting dying declarations even when clearly testimonial. “If this exception must be accepted on historical grounds,” he said, “it is sui generis.” But when analysis rests heavily on history, mounting historical anomalies—Raleigh’s outrage at Dyer’s nontestimonial hearsay, Chief Justice Marshall’s rejection of a nontestimonial coconspirator’s statement, and admission of the entire category of testimonial dying declarations—erode confidence. That’s true especially when the dying-declarations exception was among the first and best-known hearsay exceptions—and when the theory behind that exception made nothing of the statements’ (non)testimonial nature but stressed instead their reliability.

III. Crawford’s Ambiguity

Crawford’s testimonial/nontestimonial distinction, lacking both textual and historical support, suffers from a third flaw: no one quite knows what that distinction is. The Crawford Court didn’t say, “grandly declar[ing],” as Chief Justice Rehnquist scolded, “We leave for another day any effort to spell out a comprehensive definition of ‘testimonial.’ ” A decade on, the Court still hasn’t embraced a single, comprehensive definition of testimonial hearsay.

The Court came nearest this goal in footnote 6 of its 2011 ruling in Bullcoming v. New Mexico, authored by Justice Ginsburg: “To rank as ‘testimonial,’ a statement must have a ‘primary purpose’ of ‘establish[ing] or prov[ing] past events potentially relevant to later criminal prosecution.’ ”[16] Although admirably lean, this definition suffers from two shortcomings. First it’s not law. While most of Justice Ginsburg’s opinion commanded five votes, one member of her majority, Justice Thomas, withheld support from footnote 6.

Moreover, the Bullcoming definition fails to say whose “primary purpose” counts. Should trial courts look to the declarant’s purpose in speaking or, if the declarant was answering questions, the interrogator’s purpose in asking? By writing cleverly about the statement’s primary purpose, Justice Ginsburg ducked the issue. Her dodge was deliberate. Only four months earlier, two members of her fragile coalition had tussled over whose perspective counts when assessing a hearsay statement’s primary purpose. Justice Sotomayor, writing for the Court in Michigan v. Bryant, said “the statements and actions of both the declarant and interrogators provide objective evidence of the primary purpose of the interrogation.”[17] Although Justice Sotomayor devoted several paragraphs to defending this view, Justice Scalia’s response in dissent was curt: “[B]ecause the Court picks a perspective so will I: The declarant’s intent is what counts. . . . For an out-of-court statement to qualify as testimonial, the declarant must intend the statement to be a solemn declaration . . . .”

Justice Ginsburg needed both of these sparring justices to hold her majority in Bullcoming. Her diplomacy in defining testimonial hearsay in terms of the statement’s primary purpose perhaps mended the rift between her colleagues and delivered five votes for reversing Bullcoming’s conviction. But her definition’s reference to the statement’s purpose leaves unresolved whose intent controls. That uncertainty in turn makes confrontation analysis wholly unpredictable in a great mass of cases likely to come before trial courts. Consider a jailhouse informant’s conversations with a cellmate implicating both the cellmate and the accused. The informant’s purpose was to produce evidence for trial; the cellmate’s purpose was not. Whose controls? Or consider a group of officers who, as in Bryant, come upon a gunshot victim. The officers suspect the shooter lurks dangerously nearby and ask questions to find and disarm him; the victim knows the shooter has fled and poses no threat and seeks to ensure his arrest and prosecution. Whose purpose controls? Consider too a three-year-old victim of sex abuse who speaks in a playhouse-themed interview room with a social worker employed by the D.A. The interviewer poses questions with a prosecutorial purpose; the child answers with no such purpose. Whose purpose controls?[18]

In all these scenarios I’ve assumed both actors’ purposes are knowable. But as Bryant made plain, real life is rarely so tidy. Where Justice Sotomayor saw officers desperate to find a potential mass shooter whose motives, intentions, and whereabouts were unknown, Justice Scalia saw wannabe detectives seeking to solve a crime. And where Justice Sotomayor saw a mortally wounded man so weak he “may have [had] no purpose at all in answering questions posed,” Justice Scalia saw a savvy druggie who knew the shooter was distant and posed no danger and who realized the officers’ questions sought to gather evidence for trial. Nor does it help to speak, as the Court often does, of an actor’s primary purpose, “objectively considered.” What was the primary purpose, objectively considered, of the gaggle of cops who questioned the dying Anthony Covington in Bryant? What was Mr. Covington’s primary purpose, objectively considered, in speaking? These questions offer no clarity. If I may borrow a criticism Justice Scalia leveled at the old Roberts regime, “the [primary-purpose] test is inherently, and therefore permanently, unpredictable.”

Nor is it even clear the primary-purpose test controls the analysis. Although a version of that test commanded a Court majority in Davis v. Washington in 2006, those days seem gone. In Bullcoming four members of the Court—the Chief Justice and Justices Kennedy, Breyer, and Alito—joined no part of Justice Ginsburg’s opinion and embraced no primary-purpose test. Justice Thomas’s fifth vote gave Justice Ginsburg a majority, but rather than endorse her definition of a testimonial statement, Justice Thomas instead hewed to the same cramped notion of testimonial hearsay he first advanced when concurring in White v. Illinois in 1992: statements “contained in formalized testimonial materials, such as affidavits, depositions, prior testimony, or confessions.”[19] Having once joined in this view, Justice Scalia dismissed in Davis the notion that “the scope of the Clause is limited to that very formal category” of testimonial hearsay. But while Justice Thomas may be a curious outlier, he now defines the limit of Confrontation Clause protection in the post-Crawford world.

Recent cases have shown how very narrowly he defines that limit. The Court’s last foray into this realm, Williams v. Illinois of 2012,[20] concerned a commercial laboratory’s report of a DNA analysis commissioned by the Illinois State Police during a rape investigation. The report detailed a male DNA profile that laboratory technicians derived from semen-stained swabs collected from the victim. In many ways the lab report resembled the cocaine analysis at issue in Melendez-Diaz v. Massachusetts[21] and the blood-alcohol analysis at issue in Bullcoming—both of which Justice Thomas had deemed testimonial. Yet he spied differences in the formality of these documents and deemed those differences significant. The Melendez-Diaz analysis was sworn before a notary, he said, whereas the Williams report was unsworn. And while the Bullcoming report also was unsworn, Justice Thomas noted it contained a “Certificate of Analyst” affirming the technician followed proper protocol. The Williams report bore no such certificate and therefore, he concluded, was nontestimonial. Although Justice Kagan poked fun at such hairsplitting—there’s “(maybe) a nickel’s worth of difference”—Mr. Williams’s fate turned on these distinctions.

Justice Thomas wields such outsize influence because he holds the critical fifth vote for any Crawford-style reading of the Confrontation Clause. To his left sit Justices Scalia, Ginsburg, Sotomayor, and Kagan, all of whom remain true to Crawford and some version of the primary-purpose test. The four remaining justices have distanced themselves from Crawford’s framework, inching instead toward the old Roberts regime and its focus on the challenged hearsay’s reliability. The result is a Court so badly splintered that when it came time for Justice Alito to summarize Williams from the bench on the day the Court ruled, he all but confessed his inability: “Anyone interested in understanding the Court’s holding will have to read our opinions.”[22]

IV. Crawford’s Poor Sense

Yet the greatest failing of the Crawford framework and its testimonial/nontestimonial distinction is not the primary-purpose test’s ambiguity and inability to generate predictable results. Rather the Crawford framework’s greatest failing is its stubborn refusal to make sense. Here the Court’s failure to deliver a comprehensive definition of testimonial statement is not so much the problem as a symptom of the problem. The problem is the failure to explain why we should want to distinguish between testimonial and nontestimonial hearsay. Two answers seem plausible; neither explains the Crawford doctrine in a satisfying way.

The first answer emerges from hints scattered throughout the Crawford line of cases. Of the three “formulations of [the] core class of ‘testimonial’ statements” laid out in Crawford, two looked to the expectation of declarants (or of objective witnesses) that their statements would be used prosecutorially (or more generally at trial). Elsewhere in Crawford Justice Scalia suggested that some testimonial statements involve “government officers in the production of testimony with an eye toward trial”—a formula he said “presents unique potential for prosecutorial abuse.” And in Bryant Justice Scalia wrote that for a statement to be testimonial, the declarant must speak with the understanding that the statement “may be used to invoke the coercive machinery of the State against the accused.” All these hints suggest a common component of testimonial statements: the declarant’s or interrogator’s intent to create trial evidence while evading cross-examination.

Surely the law should frustrate such procedural ploys. Just as forfeiture doctrine saps the incentive for wrongdoers to eliminate witnesses, the Confrontation Clause should thwart those who contrive to plant trial evidence that eludes confrontation. If the Crawford doctrine targets testimonial hearsay for this reason, however, it’s both overinclusive and wildly underinclusive. It’s overinclusive because lots of hearsay deemed testimonial by post-Crawford courts is not created with an expectation of denying defendants the chance to cross-examine declarants. In typical crime investigations police officers canvass for witnesses. They ask for names and phone numbers precisely because they know prosecutors need those witnesses at trial. Most witnesses likewise know a trial may lie ahead and expect to testify if called. Sylvia Crawford herself was apparently willing to testify had her husband not invoked a marital privilege silencing her. Condemning all this hearsay as testimonial makes no sense if the aim is to discourage officers and witnesses from contriving to plant evidence while ducking cross-examination.

And if that’s the Crawford doctrine’s aim, the doctrine is radically underinclusive. For the most prolific actors in creating trial evidence that eludes cross-examination are not police officers or crime witnesses, but prosecutors. Every prosecutor who offers hearsay instead of calling an available declarant to testify intentionally strips the defendant of the chance to cross-examine the declarant. And if that hearsay is deemed nontestimonial because neither declarant nor interrogator aimed to create trial evidence, the Confrontation Clause leaves the defendant powerless to combat the prosecutor’s contrivance. Indeed if Justice Scalia is right that “[t]he declarant’s intent is what counts,” the Confrontation Clause is truly perverse: it protects defendants against crime witnesses (typically private persons) when they create trial evidence evading confrontation but fails to protect defendants against prosecutors (consummate state actors) when they engage in the same contrivance.

Of course there’s another—and far better—explanation for Crawford’s distinction between testimonial and nontestimonial hearsay and its general condemnation of the former: when crime witnesses speak (and officers question) “with an eye toward trial,” they have an incentive to lie (and elicit lies). This explanation, rooted in testimonial hearsay’s unreliability, makes good sense. Yet in Crawford Justice Scalia slammed the Roberts Court for looking to a statement’s reliability when assessing whether the Confrontation Clause allows admission absent cross-examination: “Reliability is an amorphous, if not entirely subjective, concept.” Instead of scrapping reliability as a constitutional touchstone, however, the Court replaced one sort of reliability analysis with another. It replaced Roberts and its plenary consideration of factors suggesting reliability or unreliability with a formula staking the entire analysis on one possible source of unreliability—the declarant’s or interrogator’s intent to create trial evidence. It’s hard to find sense here.

Little wonder, then, that four justices have divorced themselves from the Crawford framework, leaving us with the voting pattern that marked the Court’s last two confrontation ventures, Bullcoming and Williams: Four justices almost agree on a primary-purpose test dictated by Crawford and its kin, but avoid deciding whose purpose controls the analysis, about which they disagree. Justice Thomas holds to his anemic conception of the clause’s protection, which reaches only highly formalized hearsay. And the remaining four justices, having largely disavowed Crawford, hint at reviving an analysis rooted in contested hearsay’s reliability. If I may borrow again from Justice Scalia’s condemnation of Roberts, the Court’s current disarray “reveals a fundamental failure on [the justices’] part to interpret the Constitution in a way that secures its intended constraint on judicial discretion.” It’s time to suggest a fix.

V. A Way Out

Consider three propositions: At its core, if not in its particulars, Roberts was right. While Roberts was flawed, moreover, it was fixable—and was in the process of repair when the Court abandoned it. And third, a reformed Roberts regime would deliver the same results the Court reached in Crawford and every major post-Crawford case and would secure a sounder basis for analyzing two troubling questions not yet reached—the admissibility of dying declarations and of statements of child-abuse victims.

At the core of 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 suggest the statement may be admitted in absence of confrontation. That dying declarations, deemed reliable, proved admissible at the founding lends this approach historic precedent.

Starting from this sound core, however, Roberts made several fundamental errors: In most cases it abdicated to hearsay rules—and specifically “firmly rooted” hearsay rules—the judgment that contested hearsay is reliable enough to admit without cross-examination. Although our hearsay rules reflect largely sound notions of reliability, they took form in both civil and criminal courts without the deliberation that a denial of confrontation warrants. Moreover, in those cases not governed by a firmly rooted hearsay rule, Roberts endowed trial judges with nearly unguided discretion to admit hearsay absent cross-examination if it bore “particularized guarantees of trustworthiness”—a standard too flabby to ensure reasonably consistent results. And although Roberts wisely sought to pressure prosecutors to produce available declarants, its demand in most cases that “the prosecution must either produce, or demonstrate the unavailability of, the declarant whose statement it wishes to use” was too rigid. Inevitably this “rule of necessity” folded. After just six years the Court declared Roberts “cannot fairly be read to stand for the radical proposition that no out-of-court statement can be introduced by the government without a showing that the declarant is unavailable.”[23]

Despite these defects the Roberts doctrine was under repair when Crawford aborted the regime. In the Court’s last major Roberts-era confrontation case, Lilly v. Virginia,[24] a plurality of four led by Justice Stevens issued perhaps the wisest ruling in this realm. Like Crawford the case concerned the custodial statement of an accomplice implicating the accused and admitted under state law as a statement against interest. Anticipating Crawford, Justice Stevens wrote that such statements, “when offered in the absence of the declarant, function similarly to those used in the ancient ex parte affidavit system.” But he said nothing of the statements’ “testimonial” nature. Instead he declared such hearsay “inherently unreliable” and a threat to “the truthfinding function of the Confrontation Clause.” A codefendant’s “strong motivation to implicate the defendant and to exonerate himself” renders his statements to authorities about the defendant’s actions “presumptively suspect” and therefore inadmissible without cross-examination. Here Sir Walter Raleigh would agree. As he said of his absent accuser, “Cobham is absolutely in the King’s mercy; to excuse me cannot avail him; by accusing me he may hope for favour.”[25]

It’s true that under Lilly the presumptive unreliability of accomplice statements could face rebuttal. But Justice Stevens cautioned that effective rebuttal is “highly unlikely” when accomplices’ blame-shifting confessions “are given under conditions that implicate the core concerns of the old ex parte affidavit practice—that is, when the government is involved in the statements’ production, and when the statements . . . have not been subjected to adversarial testing.” Hence Lilly delivered much the same result as Crawford, but without the paradoxical, ahistorical division of hearsay into testimonial and nontestimonial statements.

Yet in Crawford Justice Scalia dismissed Lilly as a feeble hedge against wrongful admission of accomplice statements that shift blame onto the accused: “One recent study found that, after Lilly, appellate courts admitted accomplice statements to the authorities in 25 out of 70 cases . . . .” This attack on Lilly’s potency founders on several fronts. For one thing, the sampling of 70 appellate cases did not include all those cases, perhaps hundreds of them, in which trial judges, heeding the Lilly plurality’s counsel, excluded accomplice statements to authorities and therefore never gave rise to appeals on this score. Nor did Justice Scalia mention that of the 45 cases—or 64%—in which error was found, 26 resulted in reversals, suggesting the error was deemed not harmless.[26] As fewer than 15% of all federal criminal appeals result in full or partial reversals,[27] the 37% reversal rate in this sampling of 70 cases was strikingly high, suggesting the force of the Lilly plurality’s opinion. That 17 of those 26 reversals took place in murder cases[28] sharpens the point. And recall that Justice Stevens wrote in Lilly for only four justices. Four others pointedly distanced themselves from his suggestion that admission of accomplice statements to authorities implicating the accused was “highly unlikely.” The ninth justice, Scalia himself, spurned Justice Stevens’s Roberts-based analysis and thereby denied the plurality’s opinion the force of law he later knocked it for lacking.

If instead of panning Lilly the justices embrace it as a model, we can begin to imagine a recrafted confrontation-law regime. Because the Confrontation Clause aims to ensure the presence of witnesses for cross-examination and the reliability of their statements, this recrafted regime would look to declarants’ availability and their statements’ reliability when identifying the rather rare instances when hearsay may be admitted without cross-examination. Under this regime appellate courts and ultimately the Supreme Court would follow Justice Stevens’s lead in Lilly in assessing the reliability of identifiable classes of hearsay statements, an approach that would constrain trial judges’ discretion and make rulings more regular and predictable. When reaching these class judgments, courts may find guidance in current hearsay rules but would not be bound by them.

Several classes of statements the Supreme Court could declare presumptively unreliable and inadmissible absent cross-examination. Among these are grand jury testimony conducted by the prosecution “with an eye toward trial”; accomplices’ blame-shifting or blame-spreading statements made to police or prosecutors or to the court during plea allocutions; statements intended to evade cross-examination at trial, as with letters sent anonymously to authorities or arranged for delivery to law enforcement after the declarant’s death; casual gossip uttered to an acquaintance; and statements made by lab technicians employed or commissioned by police or prosecutors and able to discern the test result (positive for cocaine, for example) desired by them. Here notice-and-demand statutes of the sort the Court approved in Melendez-Diaz and Bullcoming could moderate the inconvenience to prosecutors and lab technicians of deeming these lab reports inadmissible without confrontation.

Other classes of statements the Court could declare presumptively reliable and admissible even without cross-examination. These include business records of private entities made routinely and without prosecutorial needs or specific litigation in view; public records made in the same circumstances; dying declarations made classically “in the hush of [death’s] impending presence”;[29] statements uttered in the throes of danger while seeking aid; and reports of accredited labs produced by expert technicians ignorant of the results prosecutors desire (for example, the DNA profile of a crime-scene biological sample submitted without a suspect sample and without naming a known suspect).

Other statements may prove hard to treat as a class and may require a closely factual, case-by-case analysis. In this group might fall statements against interest made privately; excited utterances and present-sense statements not made at death’s door or with danger lurking; medical statements; and statements of child victims. When assessing statements in this group, trial courts typically should exclude those made by available declarants not produced for trial, including those children able to testify without substantial trauma. Despite the Roberts-era ruling in Idaho v. Wright, there appears no sound reason to disregard corroborative evidence when judging reliability. On another score, though, Wright supplies sound guidance: a useful standard of reliability would demand that “hearsay evidence admitted under the Confrontation Clause be so trustworthy that cross-examination of the declarant would be of marginal utility.”[30] The Supreme Court and courts of appeals would need to police this standard vigilantly to ensure that trial judges apply it with rigor and substantial uniformity.

If I may borrow one last time from Justice Scalia’s opinion in Crawford, this recrafted regime provides “an empirically accurate explanation of the results [the Court’s] cases have reached.” That many of the class judgments I suggest above match results reached in Crawford-era cases is no surprise. As I noted earlier, the best rationale for the Crawford regime’s suspicion of testimonial statements is that declarants who speak in anticipation of trial have reason to lie, rendering their statements unreliable. Indeed the one Crawford-era case that would be stranded by a reliability standard’s return is Whorton v. Bockting. The Court’s perverse boast that its confrontation case law “has no application to [nontestimonial] statements and therefore permits their admission even if they lack indicia of reliability” would not survive.[31]

Happily, under this recrafted regime the Court’s case law at last would align with the lessons of Sir Walter’s trial. A reliability standard would justify Raleigh’s outraged protests at admission not only of Cobham’s statement to authorities but also of the Portuguese gentleman’s accusation made in passing to the boatman Dyer. Against the words of that “wild Jesuit or beggarly Priest,” Sir Walter no longer would stand naked before the court.


*        Judge John Crown Professor of Law, Stanford Law School. My thanks to Jeremy Matz, Jennifer Mnookin, and David Sklansky for wise comments and to Eileen Scallen, who coined “Crawford debacle” in Confrontation of Children and Other Challenging Witnesses, 35 Wm. Mitchell L. Rev. 1558, 1559 (2009). Thanks as well to Marcel Rosner and Brian Tengel of the Michigan Law Review for their sharp eyes and sensitive editing.

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

[2].     Williamson v. United States, 512 U.S. 594, 601 (1994) (internal quotation marks omitted).

[3].     See Ohio v. Clark, 135 S. Ct. 43 (2014) (mem.) (granting certiorari).

[4].     448 U.S. 56 (1980).

[5].     See Randolph N. Jonakait, “Witnesses” in the Confrontation Clause: Crawford v. Washington, Noah Webster, and Compulsory Process, 79 Temp. L. Rev. 155, 159–61 & nn.26–27 (2006).

[6].     497 U.S. 836, 864–65 (1990) (Scalia, J., dissenting).

[7].     Daniel Shaviro, The Confrontation Clause Today in Light of Its Common Law Background, 26 Val. U. L. Rev. 337, 365 (1991).

[8].     The Trial of Sir Walter Raleigh, (1603) 2 How. St. Tr. 1, 15–16 (Eng. 1816).

[9].     Raleigh, 2 How. St. Tr. at 25.

[10].     See, e.g., The Trial of Stephen Colledge, (1681) 8 How. St. Tr. 550, 641 (Eng. 1816) (quoting notorious George Jeffreys, who dismissed Titus Oates’s unsworn trial statement by declaring, “Here is Dugdale’s oath against Dr. Oates’s saying.”).

[11].     Trial of Sir Walter Raleigh, (1603) 1 D. Jardine, Crim. Trs. 400, 436 (Eng. 1832). Professor Park highlights Dyer’s testimony and Raleigh’s outrage in A Subject Matter Approach to Hearsay Reform, 86 Mich. L. Rev. 51, 90 & n.161 (1987).

[12].     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, 107, 119, 191 (2005).

[13].     United States v. Burr, 25 F. Cas. 187, 193 (C.C.D. Va. 1807).

[14].     See Jeffrey Bellin, The Incredible Shrinking Confrontation Clause, 92 B.U. L. Rev. 1865, 1891 (2012) (noting “the Davis Court’s failure to identify even one” framing-era case admitting nontestimonial hearsay).

[15].     See Robert Kry, Confrontation Under the Marian Statutes: A Response to Professor Davies, 72 Brook. L. Rev. 493 (2007).

[16].     131 S. Ct. 2705, 2714 n.6 (2011) (quoting Davis v. Washington, 547 U.S. 813, 822(2006)).

[17].     131 S. Ct. 1143, 1160 (2011).

[18].     Park poses similar hypotheticals in Is Confrontation the Bottom Line?, 19 Regent U. L. Rev. 459, 462–63 (2007).

[19].     502 U.S. 346, 365 (1992) (Thomas, J., concurring).

[20].     132 S. Ct. 2221 (2012).

[21].     557 U.S. 305 (2009).

[22].     Oral Opinion of Justice Alito, Williams v. Illinois, 132 S. Ct. 2221 (2012) (No. 10-8505). I thank David Kaye for this reference.

[23].     United States v. Inadi, 475 U.S. 387, 394 (1986).

[24].     527 U.S. 116 (1999).

[25].     Trial of Sir Walter Raleigh, (1603) 1 D. Jardine, Crim. Trs. 400, 435 (Eng. 1832).

[26].     See Roger W. Kirst, Appellate Court Answers to the Confrontation Questions in Lilly v. Virginia, 53 Syracuse L. Rev. 87, 109 (2003).

[27].     See Michael Heise, Federal Criminal Appeals: A Brief Historical Perspective, 93 Marq. L. Rev. 825, 829 tbl.1 (2009).

[28].     See Kirst, supra note 26, at 109.

[29].     Shepard v. United States, 290 U.S. 96, 100 (1933).

[30].     497 U.S. 805, 823 (1990).

[31].     549 U.S. 406, 420 (2007).


Download as PDF.