Come Back to the Boat, Justice Breyer!

Richard D. Friedman*

I want to get Justice Breyer back on the right side of Confrontation Clause issues.

In 1999, in Lilly v. Virginia,[1] he wrote a farsighted concurrence, making him one of the first members of the Supreme Court to recognize the inadequacy of the then-prevailing doctrine of the Confrontation Clause. That doctrine, first announced in Ohio v. Roberts,[2] was dependent on hearsay law and made judicial assessments of reliability determinative. In Crawford v. Washington,[3] the Court was presented with an alternative approach, making the key inquiry whether the statement in question was testimonial in nature. During the oral argument, Justice Breyer seemed to endorse the test I had articulated, in an amicus brief, for what makes a statement testimonial—“[W]ould a reasonable person in the position of declarant anticipate that the statement would likely be used for evidentiary purposes?”[4] Ultimately, he was one of seven members of the Court to support Crawford’s dramatic adoption of a testimonial approach. And two years later, in Hammon v. Indiana (decided with Davis v. Washington[5]), Justice Breyer was one of eight justices to treat as testimonial a woman’s statement accusing her husband of assaulting her, given that it was made to a police officer in the family living room a considerable time after the alleged event, while another officer held the accused at bay. (I like to think that I argued the case for the accused because I believed this was obviously right, not the other way around.)

But consider what Justice Breyer has done more recently. In 2011, he helped form a majority in Michigan v. Bryant,[6] taking a view far more restrictive than that reflected in Hammon of what is testimonial in the context of a fresh accusation. And three times over the last five years—in Melendez-Diaz v. Massachusetts,[7] Bullcoming v. New Mexico,[8] and Williams v. Illinois[9]—he has joined Chief Justice Roberts and Justices Kennedy and Alito in opinions resisting the proposition that forensic laboratory reports are ordinarily testimonial. Each of these opinions gives off a strong sense of buyer’s remorse, arising at least in large part from fear of undue burdens on the criminal justice system; the opinions seem to be looking around for any theory, however strained, that would substantially limit Crawford. Each has fallen just one vote short of gaining a majority of the Court, but in Williams—thanks to Justice Thomas, who (as in Hammon) takes an idiosyncratically formalistic view of what statements are formal enough to be deemed testimonial—the four justices became a plurality. Williams has, as predicted by Justice Kagan in a sparkling dissent, sown considerable confusion in the lower courts.

What happened?  At least in part, I think the answer is simple: Giles v. California.[10] At the argument in Giles, Justice Breyer noted that he had joined in Crawford, but he then added that he was “rapidly leaving” “the boat.”[11] And so it has been: since Giles, in every divided case, Justice Breyer has been on the side that would limit the confrontation right.

I. Giles and the Law of Unintended Consequences

Giles was the Court’s first chance to address in depth the doctrine under which an accused might forfeit his or her confrontation right by engaging in misconduct that prevents a witness from testifying subject to cross-examination. I have long believed that a robust forfeiture doctrine is essential to the development of a sound confrontation doctrine. And so when the Court took the case I was delighted, complacent in the belief that the justices would surely hold that by his misconduct Giles had forfeited the confrontation right. Boy, was I wrong!

Giles was charged with murdering his former girlfriend, Brenda Avie. The prosecution wished to introduce a statement she had made to the police after a violent incident a few weeks earlier. He objected on the ground that he had not had a chance to examine her. But the undisputed reason for that was that he had killed her.

I thought that all that was necessary to conclude that Giles had forfeited his confrontation right was a determination by the trial judge that Giles had killed Avie without justification. True, that was the very question before the jury, but this should not matter. The jury’s job in determining whether the accused was guilty of the crime charged was entirely separate from the judge’s job in determining whether the accused forfeited his confrontation right. Each of those functions may require deciding the same factual issue—but so what? This, indeed, is just what happens as an everyday matter in a prosecution for conspiracy when the judge has to determine whether a statement qualifies for the conspirator’s exemption from the rule against hearsay.[12]  And it seems clear where the equities lie: it is an abomination to allow the accused to keep the jury from hearing a witness’s statement on the grounds that he did not have an opportunity to cross-examine her if the reason he did not is that he murdered her.

But that’s not the way a majority of the Court saw it. The Court said that Giles could not be deemed to have forfeited his confrontation right unless his misconduct—killing Avie—not only prevented her from testifying but also was designed to do so.

I have thought ever since the Court decided Giles that the outcome was a disaster for the Confrontation Clause. This is sadly ironic, because this blow came at the hands of Justice Scalia, who wrote Crawford and who is intensely proud of it; indeed, he has said it is his favorite among his opinions for the Court.[13] I hope it does not turn out that, like the doomed man Oscar Wilde described in The Ballad of Reading Gaol, he “killed the thing he loved.”  Some, Wilde said, “do it with a bitter look,” others “with a flattering word,” the coward “with a kiss,” and the brave man “with a sword.” Wilde did not note that some do it in a more prosaic way, by failing to establish proper limitations on doctrine and thus causing that doctrine to be limited in other, inappropriate ways.

Indeed, one of the most predictable consequences of Giles was that, in compensation for undue narrowing of forfeiture doctrine, the term “testimonial” would be given a constricted reading. And that is just what happened in Bryant.

II. Bryant: Paying for Giles

Bryant concerned a statement made to uniformed police officers, with no apparent danger in sight, identifying the perpetrator of a shooting committed half an hour before and six blocks away. Given just those facts, it seems obvious that the statement was testimonial: clearly it was made in the anticipation—and indeed, one might say for the sole or at least dominant purpose—of bringing the shooter to justice. If Anthony Covington, the accuser, were readily available at trial, would it not be obvious that the prosecution should be required to produce him as a live witness rather than relying on the officers’ accounts of what he said?

But Bryant held that Covington’s accusations were not testimonial, which means that, so far as the Confrontation Clause is concerned, there would be no need to bring him to the trial or a deposition, no matter how readily available he was. How could this possibly be? The answer lies in two other facts: First, Covington was not only a witness to the shooting; he was also the victim. Second, he was already in dire physical condition when he made his accusations, and he died several hours later; taking his deposition would not have been humanely feasible, even if Bryant, the accused, could have been found in that time.

If Covington is to be believed, therefore, the reason he could not testify live at a trial or deposition concerning the shooting is that Bryant shot him, causing his death within an interval too short to allow even for a deposition. That is the real impetus for admissibility of his accusations.

Bryant, in other words, should have been decided as an ordinary forfeiture case. If the trial court concluded that Bryant committed serious wrongdoing that foreseeably caused Covington to be unavailable to testify subject to cross-examination, then Bryant should be deemed to have forfeited his confrontation right with respect to Covington. Such a ruling would have allowed for an equitable result—ensuring that murder of the witness by the accused would not prevent the jury from learning of the accusation—without need to deny what should have been obvious: that the accusation was testimonial.

But Giles simply forecloses this result. Even if Bryant murdered Covington, the murder was not designed to render Covington unavailable as a witness; rather, the altercation appears to have been rooted in a mundane dispute over a drug transaction. And so, as a direct consequence of Giles, the Supreme Court has adopted a mushy understanding of what it means for a statement to be testimonial, an understanding that will likely continue to impair Confrontation Clause jurisprudence in contexts far broader than the immediate one involved in Bryant itself.

Because it is often difficult to prove the accused’s intent to render a witness unavailable, Giles has also made prosecution of some homicide cases more difficult than it should be.  Sometimes, though, where the prosecution has failed to satisfy the Giles standard, the harmless-error doctrine has come to its rescue. Also, Giles’s effect has been somewhat mitigated by language in the majority opinion, and in Justice Souter’s concurrence, indicating that in the domestic violence context the necessary intent can be inferred from a pattern of abuse. I hope that over time the effect of this qualifying language becomes more significant than that of the holding itself.

III. Williams and Extraordinary Coincidences

I suspect that the overreaching of Giles—which may have given the impression of unbridled doctrinal zealotry—may also be responsible in part for Justice Breyer’s joining Justice Kennedy’s aggressive dissents in Melendez-Diaz and Bullcoming and Justice Alito’s equally aggressive opinion in Williams. The reports in Melendez-Diaz identified a substance as cocaine, and the one in Bullcoming revealed an inflated blood-alcohol level. That these reports were testimonial should have been rather obvious: they were statements made in clear contemplation of prosecutorial use. The dissents seem motivated by fear of the expense caused by requiring in-court or deposition testimony by a lab analyst (and, as Bullcoming requires, the analyst who made the report, not a surrogate)—a concern that Justice Scalia rightly brushed aside in Melendez-Diaz, on the twin bases that it is constitutionally irrelevant and factually unfounded.

Williams was more complex. That was a “cold-hit” DNA case. A private lab, Cellmark, analyzed a swab taken from a rape victim and generated a male DNA profile. The state police later matched this profile, by searching a large database, to that of Williams, who had not previously been a suspect in the case. An expert from the state police testified at Williams’s trial to the match, but Cellmark’s report was not introduced and nobody from Cellmark appeared.

Five justices concluded, though on different grounds, that the Cellmark report was not testimonial. Justice Thomas alone thought it was not sufficiently formal to be characterized as testimonial. I believe that his emphasis on formality is misguided, because a statement’s lack of proper formality does not mean it should be regarded as nontestimonial in nature; rather, it means that the statement should not be regarded as acceptable testimony at trial. Here I thought Justice Thomas made almost a parody of a formality test. The report was signed by two supervisors, it referred to “evidence” and a case number, and it was clearly written in anticipation of prosecutorial use. But it was not certified, and to Justice Thomas that made the difference.

The other four justices emphasized that the Cellmark report was not directed at a “targeted individual.” That assertion is debatable—the report was directed at the person who had a given DNA profile, which is a much more precise targeting than a name or a facial description. But in any event, so what? If only the makers of targeted statements had to confront the accused, criminal trials would look very different. For example, at least before identification of a suspect, a police officer, or anyone else, could make a statement describing the crime scene and never come to court. Similarly, a witness who observed an apparent crime but was unable to provide identifying information would never have to confront the accused.

No more persuasive is the four justices’ assertion that the Cellmark report was not used for its truth because it underlay an expert’s opinion that there was a match: the report supported that opinion only on the assumption that the report was accurate. I do believe, though, that a factor emphasized by Justice Alito’s opinion gets closer to the mark: it would have been an extraordinary coincidence for Cellmark to come up with this particular profile if it were not accurate. That consideration might have supported admissibility of the Cellmark report, without anybody testifying from Cellmark, but only if the prosecution had not relied—as it did in fact—on Cellmark’s proficiency in doing DNA testing. In other words, the essence of the reasoning would have been this: “We sent Cellmark the vaginal swab, and they sent back numbers that happened to match the profile of an actual male, Sandy Williams, who happens to live near the crime scene and against whom there happens to be other evidence linking him to the crime.  Whatever you might think of Cellmark’s proficiency, that would be far too unlikely a coincidence unless it so happens that Cellmark came up with those numbers by accurately recognizing the profile of Williams’s DNA in the swab.”

Ultimately, though, I suspect that the four justices were motivated largely by another concern, highlighted in a separate concurrence by Justice Breyer: multiple people usually participate in performing a DNA test, so if the report is testimonial, does that mean they will all have to testify? The answer is no, as shown by the experience of Michigan, which did not need the Supreme Court to tell it that, absent stipulation, the authors of forensic lab reports must testify at trial. I supervised a study of Michigan criminal sexual-conduct trials and found that, when DNA results were introduced, the presentation took an average of 1.24 live witnesses.[14]  Several factors explain why this number is so low. Often stipulations ease the prosecution’s burden. Michigan’s police lab integrates vertically to a considerable extent, minimizing the number of people involved in performing a DNA test and still operating with suitable efficiency.  (And note: lab practice should conform to the Constitution, not the other way around.) Perhaps most significantly, it is just not so that everyone who participates in a lab test must testify at trial. The only ones who must are those who make testimonial statements that the prosecution wants to use as proof. And, within limits, it is up to the prosecution to decide whose statements it needs.

A Way Forward

So the state of confrontation doctrine a decade after Crawford is disappointing. I believe the Supreme Court can do better over the next decade or two.

A crucial first step would be, one way or another, to render Giles a dead letter. That would allow for more equitable results when an accused murders a witness. It would also free the Court to implement an optimal standard for determining whether a statement is testimonial: whether a reasonable person in the position of the speaker would anticipate prosecutorial use for the statement. And conscientious application of that standard would encourage the Court to adhere to the rule of Melendez-Diaz in the full range of forensic laboratory settings (including autopsies); a few years of experience might persuade the full Court that the rule is not impractical and so end attempts to find ways around it.

Removing the obstacle created by Giles would also give the Court an incentive to develop a doctrine requiring the state, on a type of last-clear-chance theory, to take reasonable steps to mitigate potential forfeiture. If the state claims that an accused intimidated a witness, it might be required to take reasonable measures to protect her or to create conditions under which she will feel comfortable testifying. Or suppose, as in Bryant, the accused fatally strikes the victim, but, before dying, the victim makes a testimonial statement. If the victim dies soon afterward, then presumably the accused should be held to have forfeited his confrontation right (without any need for a separate dying-declaration doctrine). But if the victim remains alive for a considerable period, it may be appropriate to require the prosecution, as a condition of using testimonial statements by the victim, to offer the accused the opportunity to be confronted with the victim at a deposition; this was the practice at about the time the Confrontation Clause was adopted.[15]

Freeing forfeiture doctrine of the unfortunate stricture created by Giles, but shaping it by a sensible mitigation requirement, might therefore lead to the development of a body of Confrontation Clause law that is reasonably clear and simple, practical, and in keeping with the traditional understanding of the confrontation right. I have the pleasant hope that the Court will move in this direction, that Justice Breyer will be lured back onto the Crawford boat, and that those two developments will each reinforce the other.

I have left aside one conundrum that the Court has not yet addressed but will very soon—how to handle statements by very young children. I believe that some very young children should be considered beyond the bounds of the clause because such children are not capable of being witnesses at all; nevertheless, they are still sources of potentially valuable evidence, and the accused should have an opportunity to examine them out of court through a qualified expert, presumably a psychologist. But that is the subject of another essay,[16] as it will be the subject of another case.[17]


*         Alene and Allan F. Smith Professor of Law, University of Michigan Law School.

[1].     527 U.S. 116 (1999).

[2].     448 U.S. 56 (1980).

[3].     541 U.S. 36 (2004).

[4].     Transcript of Oral Argument at 9–10, Crawford, 541 U.S. 36 (No. 02-9410).

[5].     547 U.S. 813 (2006).

[6].     131 S. Ct. 1143 (2011).

[7].     557 U.S. 305 (2009).

[8].     131 S. Ct. 2705 (2011).

[9].     132 S. Ct. 2221 (2012).

[10].     554 U.S. 353 (2008).

[11].     Transcript of Oral Argument at 13, Giles, 554 U.S. 353 (No. 07-6053).

[12].     Fed. R. Evid. 801(d)(2)(E).

[13].     Jeffrey Toobin, The Nine: Inside the Secret World of the Supreme Court 316 (2007).

[14].     Richard D. Friedman, Is There a Multi-Witness Problem with Respect to Forensic Lab Tests?, The Confrontation Blog (Dec. 7, 2010, 10:11 AM), http://confrontationright.blogspot.com/2010/12/is-there-multi-witness-problem-with.html [http://perma.cc/VWP4-9L2R].

[15].     See, e.g., R v. Forbes, Holt 599, 171 E.R. 354 (1814).

[16].     Richard D. Friedman & Stephen J. Ceci, The Child Quasi-Witness, 82 U. Chi. L. Rev. (forthcoming 2015).

[17].     Ohio v. Clark, 999 N.E.2d 592 (Ohio 2013), cert. granted, 135 S. Ct. 43 (2014) (mem.).


Download as PDF.