Amicus Curiae Brief — Adam Samia, aka Sal, aka Adam Samic, Petitioner v. United States

No. 22-196

IN THE

ADAM SAMIA, PETITIONER

v.

UNITED STATES OF AMERICA

_______________

On Writ of Certiorari to the United States Court of

Appeals for the Second Circuit

_______________

BRIEF OF AMICI CURIAE NATIONAL

ASSOCIATION OF CRIMINAL DEFENSE

LAWYERS AND AMERICAN CIVIL LIBERTIES

UNION IN SUPPORT OF PETITIONER

_______________

Jeffrey L. Fisher

Co-Chair, NACDL AMICUS

COMMITTEE

559 Nathan Abbott Way

Stanford, CA 94305

David D. Cole

AMERICAN CIVIL LIBERTIES

UNION FOUNDATION

915 15th St. N.W.

Washington, D.C. 20005

Jonathan Y. Ellis

Counsel of Record

APPELLATE JUSTICE

INITIATIVE AT

MCGUIREWOODS LLP

888 16th St. N.W.

Suite 500

Washington, D.C. 20006

(202) 828-2887

jellis@mcguirewoods.com

Gregory J. DuBoff

APPELLATE JUSTICE

INITIATIVE AT

MCGUIREWOODS LLP

800 East Canal St.

Richmond, VA 23219

i

TABLE OF CONTENTS

TABLE OF AUTHORITIES ...................................... ii

INTEREST OF AMICI CURIAE............................... 1

INTRODUCTION AND SUMMARY OF

ARGUMENT ................................................... 3

ARGUMENT .............................................................. 7

A.

The Bruton Rule Rests on the

Unique Harms Presented By

Codefendant Confessions that

Shift or Spread Blame.......................... 8

B.

The Bruton Rule Should Apply to

All Directly Accusatory

Codefendant Confessions

Regardless of Other Context.............. 13

C.

In All Events, the Second

Circuit’s Approach Should Be

Rejected .............................................. 21

CONCLUSION ........................................................ 25

ii

TABLE OF AUTHORITIES

Cases

Arizona v. Fulminante,

499 U.S. 279 (1991) ........................................... 10

Bruton v. United States,

391 U.S. 123 (1968) ..........3-6, 8, 10-14, 19, 21-24

California v. Green,

399 U.S. 149 (1970) ............................................. 3

Colorado v. Connelly,

479 U.S. 157 (1986) ........................................... 10

Commonwealth v. Bosworth,

39 Mass. 397 (1839)......................................... 8, 9

Crawford v. United States,

212 U.S. 183 (1909) ............................................. 9

Crawford v. Washington,

541 U.S. 36 (2004) ......................................... 3, 12

Cruz v. New York,

481 U.S. 186 (1987) ..................... 3, 7-8, 14-18, 22

Foster v. People,

18 Mich. 266 (1869) ............................................. 9

Gray v. Maryland,

523 U.S. 185 (1998) ....4-5, 7, 10, 13, 15-20, 22-23

Hanifa v. State,

505 S.E.2d 731 (Ga. 1998) ................................. 20

iii

Harris v. New York,

401 U.S. 222 (1971) ........................................... 12

Lee v. Illinois,

476 U.S. 530 (1986) ....................................... 9, 23

Lilly v. Virginia,

527 U.S. 116 (1999) ........................................... 11

Melendez-Diaz v. Massachusetts,

557 U.S. 305 (2009) ........................................... 18

Miranda v. Arizona,

384 U.S. 436 (1966) ........................................... 12

People v. Williams,

18 Cal. 187 (1861)................................................ 9

Richardson v. Marsh,

481 U.S. 200 (1987) ......................... 5-6, 16-18, 21

Shepard v. United States,

290 U.S. 96 (1933) ............................................. 11

State v. Tucker,

414 S.E.2d 548 (N.C. 1992) ............................... 20

United States v. Jass,

569 F.3d 47 (2d Cir. 2009)....................... 6, 22, 23

United States v. Lyle,

919 F.3d 716 (2d Cir. 2019)..................... 7, 22, 23

United States v. Simmons,

374 F.3d 313 (5th Cir. 2004) ............................. 21

Williamson v. United States,

512 U.S. 594 (1994) ........................................... 10

iv

Constitutional Provisions

U.S. Const. amend. VI ........................................ 8, 14

Other Authorities

E. Cleary,

McCormick on Evidence (2d ed. 1972) .............. 10

William D. Evans,

On the Law of Evidence (1806) ......................... 10

H. Fielding,

An Enquiry Into the Causes of the Late

Increase of Robbers (London 1751) ..................... 9

John H. Langbein,

Shaping the Eighteenth-Century

Criminal Trial,

50 U. Chi. L. Rev. 1 (1983) .................................. 9

Andrew D. Leipold & Hossein A. Abbasi,

The Impact of Joinder and Severance on

Federal Criminal Cases: An Empirical

Study,

59 Vand. L. Rev. 349 (2006) ........................ 20, 21

7 J. Wigmore, Evidence § 2056 (3d ed. 1940) .......... 8

1

INTEREST OF AMICI CURIAE 1

The National Association of Criminal Defense

Lawyers (NACDL) is a nonprofit voluntary professional bar association that works on behalf of criminal

defense attorneys to ensure justice and due process for

those accused of crime or misconduct. NACDL was

founded in 1958. Its nationwide membership of many

thousands of direct members—and up to 40,000 with

affiliates—includes private criminal defense lawyers,

public defenders, military defense counsel, law professors, and judges. It is the only nationwide professional bar association for public defenders and private

criminal defense lawyers.

NACDL is dedicated to advancing the proper, efficient, and just administration of justice. To that end,

NACDL files numerous amicus briefs each year in this

Court and other federal and state courts, seeking to

provide amicus assistance in cases that present issues

of broad importance to criminal defendants, criminal

defense lawyers, and the criminal justice system as a

whole.

The American Civil Liberties Union (ACLU) is a

nationwide, nonprofit, nonpartisan organization with

approximately two million members and supporters

dedicated to defending the principles embodied in the

Bill of Rights and our nation’s civil rights laws. Since

its founding in 1920, the ACLU has frequently appeared before this Court, the lower federal courts, and

1 Pursuant to Supreme Court Rule 37.6, no counsel for any party

has authored this brief in whole or in part, and no person or entity, other than amici and their counsel, made a monetary contribution intended to fund the preparation or submission of this

brief.

2

state courts in cases defending the fair trial rights of

individuals accused of crime.

Amici submit this brief in support of the petitioner

because the issue presented in this case is of paramount importance to criminal defense attorneys

throughout the country and the fair trial rights of the

individuals they represent.

3

INTRODUCTION AND

SUMMARY OF ARGUMENT

Accomplice confessions that shift or spread blame

to others present a special threat to the justice system.

They are notoriously unreliable, yet wildly prejudicial. Fortunately, in many cases, a defendant can

combat such prejudice with “the crucible of crossexamination,” Crawford v. Washington, 541 U.S. 36,

61 (2004), the “greatest legal engine ever invented for

the discovery of truth,” California v. Green, 399 U.S.

149, 158 (1970). But at a joint trial, “[t]his prejudice

cannot be dispelled . . . if the co-defendant does not

take the stand.” Bruton v. United States, 391 U.S.

123, 132 (1968).

In Bruton, the Court held that the introduction of

an out-of-court confession that implicates another defendant by name violates the Confrontation Clause,

even if the jury is instructed not to consider the confession against the nonconfessing defendant. In such

cases, the Court explained that “the risk that the jury

will not, or cannot, follow [that] instruction[]” is too

great, and the threat to liberty too severe. Id. at 135.

Bruton thus excludes “the entire category of codefendant confessions that implicate the defendant in the

crime.” Cruz v. New York, 481 U.S. 186, 191 (1987).

The question presented in this case is whether the

same rule should apply where the codefendant’s confession is redacted or altered to refer to the nonconfessing defendant only by neutral terms—e.g., “another person”—instead of by name. Here, for example, a codefendant’s confession accused petitioner of

shooting the murder victim at point-blank range. It

also provided details about petitioner, like his travel

history, living arrangements, and the weapons he

4

owned. But, presented through the testimony of a

DEA agent at trial, the redacted confession did not refer to petitioner by name. Instead, the DEA agent replaced petitioner’s name with phrases like “someone”

or “the other person.”

In the Second Circuit’s view, the introduction of

the redacted confession did not violate Bruton because

it did not identify petitioner by name. That ruling is

wrong. For the reasons that follow, the Court should

reverse the Second Circuit’s decision and hold that, as

a class, directly accusatory confessions—that is, confessions that “refer directly to” a nonconfessing codefendant’s role in the crime, Gray v. Maryland, 523

U.S. 185, 196 (1998)—violate the Confrontation

Clause. Because the confession in this case, even as

redacted, directly accused another person of committing the crime, and that other person was petitioner,

the Court should hold that its introduction violated

the Confrontation Clause.

A. The Court’s decision in Bruton rested on four

notable attributes of codefendant confessions. First,

codefendant confessions that shift or spread blame are

notoriously unreliable and “inevitably suspect.” Bruton, 391 U.S. at 136. Second, codefendant confessions

are—despite their unreliability—singularly prejudicial. They are “so damaging” because juries believe

them rather than “give such evidence the minimal

weight it logically deserves.” Id. at 138 (Stewart, J.,

concurring). Third, the prejudice caused by codefendant confessions is uniquely impervious to limiting instructions. In no other context are juries asked to consider a piece of evidence as direct, substantive, and

valid proof of one defendant’s guilt, but to completely

ignore that same evidence as it pertains to the guilt of

5

another. And, fourth, the risk of prejudice from a

codefendant’s confession is inherently avoidable by

holding separate trials. That procedure gives the government “the benefit of the confession to prove the

confessor’s guilt . . . without at the same time infringing the nonconfessor’s right of confrontation.” Id. at

133-34.

B. The Court should hold that the Bruton rule categorically applies to any directly accusatory codefendant confession, regardless of any other context, for

four reasons.

First and foremost, each of the attributes of codefendant confessions that led this Court to adopt the

Bruton rule applies to all directly accusatory codefendant confessions. Indeed, codefendant confessions

are so dangerous precisely because they are “directly

accusatory.” Barring all directly accusatory confessions would align the Bruton rule with its supporting

rationale.

Second, excluding all directly accusatory codefendant confessions comports with the Court’s decisions in

this area. The Court held that the Confrontation

Clause was violated in Bruton, Cruz, and Gray, all of

which involved directly accusatory statements. In

Richardson v. Marsh, 481 U.S. 200, 211 (1987), by contrast, the Court found no Bruton violation from the

admission of a confession “redacted to eliminate not

only the defendant’s name, but any reference to his or

her existence.” With those redactions, the confession

was not directly accusatory because it “did not refer

directly to the defendant.” Gray, 523 U.S. at 196.

Third, a rule excluding all directly accusatory codefendant confessions is readily administrable. The

6

Court has previously emphasized the importance of a

trial court’s ability to “predict the admissibility of a

confession in advance of trial.” Richardson, 481 U.S.

at 209. Petitioner has explained why his contextbased rule achieves that goal. But should the Court

have any concerns about that rule’s administrability,

Amici submit that the answer is not to adopt a standard that wagers constitutional rights on the notion

that a jury may not connect the dots linking a defendant with the “someone” implicated in his codefendant’s confession. Instead, the proper solution would

be to adopt a rule recognizing that those dots will

practically always be connected—and thus all directly

accusatory confessions should be excluded.

Fourth, prohibiting all directly accusatory confessions from being admitted at joint trials presents no

serious obstacle to the proper functioning of our criminal justice system. To the contrary, it would preserve

an important role for separate trials. Given this “viable alternative[],” Bruton, 391 U.S. at 126, several

States have long prohibited all codefendant confessions that refer to the defendant’s existence. Federal

defendants and those in other States should likewise

not have to bear the inherently avoidable risk of juries

“look[ing] to the incriminating extrajudicial statements in determining . . . guilt.” Id. at 126.

C. In all events, the Court should emphatically reject the Second Circuit’s approach. Under that circuit’s precedent, “whether a jury might infer” that “a

confederate . . . referenced the defendant” is not the

“critical inquiry.” United States v. Jass, 569 F.3d 47,

61 (2d Cir. 2009). Instead, the court asks only

whether redactions “sufficiently conceal[] the fact of

explicit identification.” Id. (emphasis added). And the

7

court restricts its analysis to “the redacted statement

in isolation.” Id. at 62. This approach leads to the

wooden rule that “introduction of a co-defendant’s confession with the defendant’s name replaced by a neutral noun or pronoun does not violate Bruton.” United

States v. Lyle, 919 F.3d 716, 733 (2d Cir. 2019).

That rule is irreconcilable with the Court’s precedents. Indeed, the Second Circuit’s approach bypasses the central inquiry under Bruton: whether the

jury will, in fact, understand that the confessor has

implicated a defendant in the crime. And it ignores

Gray’s holding that “directly accusatory” confessions,

as a class, “closely resemble Bruton’s unredacted

statements” because “the jury will often realize that

the confession refers specifically to the defendant.”

Gray, 523 U.S. at 192-94.

Amici urge the Court to clarify that the Confrontation Clause bars unconfronted confessions that are

“directly accusatory.” Gray, 523 U.S. at 194. This

standard protects an indispensable constitutional

right, is easily administrable before trial, and preserves an important role for separate trials in the few

cases ill-suited to Bruton-compliant redactions.

ARGUMENT

In Bruton, the Court “held that a defendant is deprived of his rights under the Confrontation Clause

when his codefendant’s incriminating confession is introduced at their joint trial, even if the jury is instructed to consider that confession only against the

codefendant.” Cruz, 481 U.S. at 187-88. Under such

circumstances, the practical reality is that juries will

virtually always still consider the nontestifying code-

8

fendant to be, in effect, a “witness against” the defendant. U.S. Const. amend. VI. Bruton thus establishes

a constitutional rule that trumps “the general rule” of

evidence “that jury instructions suffice to exclude improper testimony.” Cruz, 481 U.S. at 191.

The Bruton rule rests on four notable attributes of

codefendant confessions. Those same attributes support extending Bruton to all directly accusatory codefendant confessions—that is, confessions that implicate a nonconfessing defendant in the crime. This rule

follows from this Court’s precedents on such confessions, provides a principled and easily administrable

standard, and poses no fundamental obstacle to the

efficient and just operation of the criminal justice system. The Court should adopt that rule and reject the

Second Circuit’s flawed approach to codefendant confessions.

A.

The Bruton Rule Rests on the Unique

Harms Presented By Codefendant Confessions that Shift or Spread Blame.

The Court has explained the need for the Bruton

rule by focusing on four notable attributes of codefendant confessions that shift or spread blame.

1. First, a codefendant’s confession is “presumptively unreliable” and “inevitably suspect.” Bruton,

391 U.S. at 136 & n.12. Indeed, accomplice confessions are among the lowest forms of evidence known

to the law. “[T]he source of this evidence is so corrupt,

that it is always looked upon with suspicion and jealousy.” Commonwealth v. Bosworth, 39 Mass. 397, 399

(1839). English common law once disqualified all accomplices as incompetent witnesses. See 7 J. Wigmore, Evidence § 2056 (3d ed. 1940). And even when

9

that restriction loosened, the corroboration rule

emerged in its place. Under that rule, “the evidence

of the accomplice . . . unless it be corroborated by some

other evidence, it is not sufficient.” H. Fielding, An

Enquiry Into the Causes of the Late Increase of Robbers 172-73 (London 1751); see also John H. Langbein,

Shaping the Eighteenth-Century Criminal Trial, 50 U.

Chi. L. Rev. 1, 96 (1983) (describing the corroboration

rule as “The First Rule of Evidence”).

American practice has displayed a similar distrust

of accomplice confessions. See Crawford v. United

States, 212 U.S. 183, 204 (1909). In the past, some

judges “consider[ed] it their duty to advise a jury to

acquit, where there [was] no evidence other than the

uncorroborated testimony of an accomplice.” Bosworth, 39 Mass. at 399. Other courts went further and

entered a directed verdict, explaining that the law

“will not permit any citizen to be convicted solely by

the testimony of the accomplice.” People v. Williams,

18 Cal. 187, 191 (1861). In the same vein, this Court

has “spoken with one voice in declaring presumptively

unreliable accomplices’ confessions that incriminate

defendants.” Lee v. Illinois, 476 U.S. 530, 541 (1986).

Past jurists also agreed that the best way to counter the unreliability of accomplice confessions was

cross-examination. Courts observed that “[w]hen accomplices are allowed to testify . . . . there is every reason to compel them to submit to the fullest and most

searching inquiry.” Foster v. People, 18 Mich. 266, 276

(1869). And they explained that while exposing flaws

in a witness’s testimony “is the general office of a

cross-examination, . . . this is more especially important in respect to [accomplices].” Williams, 18 Cal.

10

at 191; see also William D. Evans, On the Law of Evidence 260 (1806) (accomplice testimony “very properly

occasions a great degree of caution,” requiring “a minute examination of circumstances” through crossexamination). As this Court observed more recently,

an accomplice confession “creates a special, and vital,

need for cross-examination.” Gray, 523 U.S. at 194.

2. Second, experience teaches that accomplice

confessions are uniquely prejudicial for the simple fact

that juries believe them, despite their unreliability.

“A confession is like no other evidence.” Arizona v.

Fulminante, 499 U.S. 279, 296 (1991) (plurality opinion). As even the Bruton dissent recognized, it comes

“from the actor himself, the most knowledgeable and

unimpeachable source of information about his past

conduct.” Bruton, 391 U.S. at 140 (White, J., dissenting); cf. Williamson v. United States, 512 U.S. 594,

599-600 (1994) (“One of the most effective ways to lie

is to mix falsehood with truth, especially truth that

seems particularly persuasive because of its selfinculpatory nature.”). For that reason, confessions

“have profound impact on the jury.” Fulminante, 499

U.S. at 296. In fact, “[t]riers of fact accord confessions

such heavy weight in their determinations that ‘the

introduction of a confession makes the other aspects

of a trial in court superfluous.’” Colorado v. Connelly,

479 U.S. 157, 182 (1986) (Brennan, J., dissenting)

(quoting E. Cleary, McCormick on Evidence 316 (2d

ed. 1972)).

It is because accomplice confessions that shift or

spread blame are “so damaging” that juries cannot always “be trusted to give such evidence the minimal

weight it logically deserves.” Bruton, 391 U.S. at 138

(Stewart, J., concurring). And Bruton held that this

11

prejudice is “intolerably compounded when the alleged accomplice . . . cannot be tested by crossexamination.” Id. at 136 (majority opinion).

3. Third, codefendant confessions are uniquely illsuited to limiting instructions. Such instructions ask

the jury “to perform the overwhelming task of considering [the confession] in determining the guilt or innocence of the declarant and then of ignoring it in determining the guilt or innocence of any codefendants.”

Id. at 131. And Bruton determined that the Confrontation Clause cannot tolerate the fiction that a jury

could be expected to “segregate [such] evidence into

separate intellectual boxes.” Bruton, 391 U.S. at 131;

cf. Shepard v. United States, 290 U.S. 96, 104 (1933)

(“It is for ordinary minds, and not for psychoanalysts,

that our rules of evidence are framed.”).

Put simply, a jury that credits A’s confession that

he committed the crime with B when determining A’s

guilt is highly unlikely to ignore the “inevitable conclusion” that B committed the crime with A when determining B’s guilt. Bruton, 391 U.S. at 131, 132 n.8.

Instead, the jury will almost inevitably “look[] to the

incriminating extrajudicial statements in determining” the nonconfessing defendant’s guilt. Id. at 126.

That is so even though the accusation against the nonconfessing defendant will be the least reliable portion

of the codefendant’s confession. See Lilly v. Virginia,

527 U.S. 116, 133 (1999) (Stevens, J., concurring) (noting the “presumptive unreliability of the ‘non-selfinculpatory’ portions” of a confession). And it is so,

even though—or perhaps because—the nonconfessing

defendant will have no opportunity to subject his accuser to cross-examination.

12

4. Fourth, the risk of prejudice from a codefendant’s unconfronted confession is uniquely avoidable.

In most contexts, when evidence is admitted for a

limited purpose, there is no way to present the probative aspects apart from the unfairly prejudicial ones.

For example, statements obtained in violation of Miranda v. Arizona, 384 U.S. 436 (1966), cannot be used

to impeach a defendant’s trial testimony without risking that the jury will rely on those statements as substantive evidence of a defendant’s guilt. After all, the

same jury charged with assessing the defendant’s

guilt must also evaluate his credibility. Because that

risk is unavoidable, the Court has reasonably determined that a limiting instruction strikes the right balance between protecting the defendant’s constitutional rights and allowing the prosecution to employ

un-Mirandized statements for a valid purpose. See

Harris v. New York, 401 U.S. 222, 224 (1971) (reasoning that a contrary rule would let a defendant “resort

to perjurious testimony in reliance on the Government’s disability to challenge his credibility”).

A codefendant’s confession, however, does not raise

the same dilemma because the government always

has the option of holding separate trials for the confessing and nonconfessing defendants. In the confessing defendant’s trial, no limits need to be imposed on

the jury’s consideration of the confession. See Crawford, 541 U.S. at 45. But in the trial of the nonconfessing defendant (or defendants), the confession has

no legitimate evidentiary value. It is a “presumptively unreliable out-of-court statement of a nonparty

who was not a witness subject to cross-examination.”

Bruton, 391 U.S. at 136 n.12. Separate trials thus afford the government “the benefit of the confession . . .

13

without at the same time infringing the nonconfessor’s right of confrontation.” Id. at 133-34. For that

reason, Bruton held that any convenience from joint

trials gained at the expense of “fundamental principles of constitutional liberty” was too high a price. Id.

at 134-35.

B.

The Bruton Rule Should Apply to All Directly Accusatory Codefendant Confessions Regardless of Other Context.

Petitioner persuasively argues that a Confrontation Clause violation occurs not only when a defendant is identified by name in a codefendant’s confession, but whenever the jury is likely to ascertain that

the codefendant implicated a nonconfessing defendant

as an accomplice based on various contextual clues

concerning the trial.

For several reasons, however, Amici propose a

somewhat broader rule. We urge the Court to hold

that Bruton prohibits all directly accusatory confessions, regardless of other context. Such confessions,

as a class, implicate every concern that animated Bruton. Excluding all directly accusatory codefendant

confessions, moreover, aligns with and harmonizes

the Court’s precedents. Such a rule is easily administrable. And it would pose no serious obstacle to the

efficient operation of the criminal justice system, but

rather preserves an important role for separate trials.

1. The origin of Bruton-type prejudice is a confession’s directly accusatory nature. All confessions that

shift or spread blame “function the same way grammatically” in that they “obviously refer directly to

someone.” Gray, 523 U.S. at 194, 196. And with any

14

such confession, the concerns animating Bruton inevitably follow. That is why Bruton excluded “the entire

category of codefendant confessions that implicate the

defendant in the crime.” Cruz, 481 U.S. at 191 (emphasis added).

The exclusion of all directly accusatory confessions

flows directly from Bruton. All directly accusatory

confessions are inherently unreliable given the declarant’s “recognized motivation to shift blame onto others.” Bruton, 391 U.S. at 136. All are uniquely prejudicial because juries view the declarant, having admitted his own guilt, as a “knowledgeable and unimpeachable source of information.” Id. at 140 (White,

J., dissenting). All are uniquely impervious to jury instructions that ask the jury to simultaneously consider and not consider the same evidence as direct,

substantive evidence of guilt. And the prejudice that

flows from all directly accusatory confessions is

uniquely avoidable through separate trials.

In Bruton, the Court identified the source of prejudice as the jury’s “[in]ability to disregard a codefendant’s confession implicating another defendant.” 391

U.S. at 130. Thus, a Confrontation Clause violation

occurs whenever the jury believes that a codefendant

implicated another defendant in his confession—and

thus functionally understands the confessor to be a

“witness against” the defendant. U.S. Const. amend.

VI. Under the Confrontation Clause, how the jury

draws that connection is immaterial. Once the jury

understands that a confession “implicate[s] the defendant in the crime,” how that connection was made

“cannot conceivably be relevant to whether . . . the

jury is likely to obey the instruction to disregard it,”

15

or whether “the jury’s failure to obey is likely to be inconsequential.” Cruz, 481 U.S. at 193.

Redactions that replace a defendant’s name with a

phrase like “someone else” do not meaningfully

change the calculus. The Court has already explained

why efforts to obscure the “someone” referenced in a

directly accusatory confession are ill conceived: Too

many clues will inevitably lead jurors to “work out the

reference.” Gray, 523 U.S. at 193. To start, the government’s case will necessarily depend on the implicit

assertion that the defendant alone matches the identity of this “someone.” Were it otherwise, the prosecutor could not “argue the confession is reliable.” Id.

Moreover, even if the fact of redaction is effectively

concealed, a mildly attentive jury will notice the prosecutor’s strangely incurious attitude toward the identity of this unnamed accomplice. And if the jury harbored any doubts that the “someone” was the other defendant, his attorney’s inaction would erase them.

Nothing could be more helpful to a defendant’s case

than a confessor’s admission that he committed the

crime with someone other than the defendant. If a confession did not implicate a defendant, jurors would expect his lawyer to hammer that point home. And they

will inevitably infer from a lawyer’s silence that his

client was implicated. Finally, those practical realities will be reinforced—not dissipated—when the jury

“hears the judge’s instruction not to consider the confession as evidence against” the other defendant. Id.

These considerations all support the Court’s conclusion that “directly accusatory” confessions, “considered as a class, so closely resemble Bruton’s unredacted statements that . . . the law must require the

same result.” Id. at 192, 194. As Gray explains, what

16

makes a confession “difficult to thrust out of mind” is

that it “obviously refer[s] directly to someone.” Id. at

196 (emphasis added). Nothing in Bruton or later

cases suggests that a jury’s ability to disregard the

confession will be enhanced if it has to connect a few

dots before concluding that “someone” is the defendant.

2. Excluding directly accusatory codefendant confessions, as a class, comports with this Court’s precedents. The statements in Bruton and Cruz were “directly accusatory” because they “use[d] a proper name

to point explicitly to an accused defendant.” Id. at

194; see also Cruz, 481 U.S. at 188-89. But the statement in Richardson was not “directly accusatory” because it “d[id] not point directly to a defendant at all.”

Gray, 523 U.S. at 194. That statement was “redacted

to eliminate not only the defendant’s name, but any

reference to his or her existence.” Richardson, 481

U.S. at 211.

a. The Court explicitly relied on the directly accusatory standard to decide Gray. That case resolved

the issue Richardson left open: “[T]he admissibility of

a confession in which the defendant’s name has been

replaced with a symbol or neutral pronoun.” Richardson, 481 U.S. at 211 n.5. Gray held that such redactions, “considered as a class, so closely resemble Bruton’s unredacted statements that . . . the law must require the same result.” 523 U.S. at 192. The Court

explained that the confessions “function the same way

grammatically” in that they are “directly accusatory.”

Id. at 194. And even when a directly accusatory confession is redacted to remove proper names, it still “obviously refer[s] directly to someone, often obviously

the defendant.” Id. at 196.

17

Gray acknowledged that the “someone” referenced

in a confession may be more or less transparent depending on the method of redaction and other casespecific factors. See id. at 194-95. But consistent with

its earlier decisions in this area, the Court chose to

exclude “directly accusatory” confessions “as a class,”

rather than invite case-by-case determinations. Id.;

see also Cruz, 481 U.S. at 191 (noting Bruton’s rejection of the view that a confession’s prejudicial impact

“should be assessed on a case-by-case basis”).

Of course, trial judges cannot know for certain

whether a jury will inevitably make the connection

that Bruton guards against. See Cruz, 481 U.S. at 192

(describing “[t]he infinite variability of inculpatory

statements . . . and of their likely effect on juries”).

Gray thus drew a principled line at “directly accusatory” confessions as a class. 523 U.S. at 194. Such

confessions—even when redacted with symbols or pronouns—“obviously refer directly to someone” and “involve inferences” about the identity of the unnamed

individual “that a jury ordinarily could make immediately.” Id. at 196.

b. By contrast, the Court recognized in Richardson that confessions “redacted to eliminate not only

the defendant’s name, but any reference to his or her

existence,” are different in kind, not just degree. Richardson, 481 U.S. at 211. Such confessions might prejudice a nonconfessing codefendant by disclosing unhelpful facts, but they are not directly accusatory. In

Richardson, for example, Marsh’s codefendant confessed that, while driving to the crime scene, he told a

third person “that he would have to kill the victims.”

481 U.S. at 203. Marsh later testified that she was in

18

the same car, but denied knowledge of her codefendant’s murderous intent. See id. at 204. Plainly, if the

jury had impermissibly relied on her codefendant’s

confession about the car conversation in determining

Marsh’s knowledge, it would have unfairly prejudiced

Marsh.

In finding the confession admissible with a limiting instruction, the Court contrasted the prejudice

from such “inferential incrimination” with that from a

confession implicating a defendant “as [an] accomplice.” Id. at 208. And while it acknowledged that “it

may not always be simple for the members of a jury to

obey the instruction that they disregard” the former,

it concluded that “there does not exist the overwhelming probability of their inability to do so that is the

foundation of Bruton[].” Id.

Whatever prejudice might have flowed from the

confession in Richardson, it was not Bruton prejudice

because Marsh’s codefendant had not directly “implicate[d] [her] in the crime.” Cruz, 481 U.S. at 191. The

redacted confession did not “refer directly to the defendant” in any way. Gray, 523 U.S. at 196. It merely

disclosed a non-accusatory fact that, when connected

with other evidence, could have unfairly prejudiced

the defendant. As Richardson held, mitigating risk of

that kind of prejudice is the traditional office of limiting instructions. 2

2 The focus here on whether the codefendant’s confession is “ac-

cusatory” is unique, for the reasons just elaborated, to the Bruton

context. In general, a statement by a nontestifying witness need

not be accusatory to implicate the Confrontation Clause. See,

e.g., Melendez-Diaz v. Massachusetts, 557 U.S. 305, 313-14

(2009).

19

3. Excluding all directly accusatory codefendant

confessions is also readily administrable. Foremost,

it aligns with the Court’s goal of enabling trial judges

to “predict the admissibility of a confession in advance

of trial.” Richardson, 481 U.S. at 209. Whether a redacted confession still “refer[s] directly to someone”

else involved in the crime is easily discernable, even

when reviewing the confession in isolation. Gray, 523

U.S. at 186.

Moreover, barring all directly accusatory confessions would not only offer predictability before trial,

but it would also benefit the conduct of trial itself. It

is a rare claim of Bruton error that does not focus, at

least in part, on a prosecutor’s comments during opening statements or closing arguments. And that makes

sense. A prosecutor’s focus at those times is on persuading the jury to convict, not maintaining the dubious pretension that the “someone” referenced in a confession may not be the other defendant. So it is no

surprise that prosecutors often “undo the effect of [a]

limiting instruction by urging the jury to use [a] confession” against the nonconfessing defendant, Richardson, 481 U.S. at 211, particularly given the obvious

“windfall” to the government of the jury drawing that

very connection, Bruton, 391 U.S. at 129. Faithful application of our rule would reduce the occasions for

such errors. If nothing in a codefendant’s redacted

confession implicates the other defendant, then prosecutors will be unlikely to wager their own credibility

with the jury by implying the opposite.

4. Finally, excluding all directly accusatory confessions is consistent with the efficient operation of

the criminal justice system. This approach would present prosecutors with a clear choice. They will know

20

long before trial whether a confession can be redacted

to remove its directly accusatory elements. Where

that is possible, the prosecutor may still decide that

the efficiencies of a joint trial are outweighed by the

advantage of using the unredacted confession in a separate trial of the confessor. Where not, the government is always free to pursue separate trials without

losing any legitimate evidentiary benefit from the

codefendant’s confession.

To be sure, adoption of our rule might still lead to

a marginal increase in the number of separate trials.

But the increase would likely be modest. Relatively

few trials involve multiple defendants—much less a

confession by one defendant implicating another. See

Andrew D. Leipold & Hossein A. Abbasi, The Impact

of Joinder and Severance on Federal Criminal Cases:

An Empirical Study, 59 Vand. L. Rev. 349, 366 (2006).

For the small number that do, the confession can usually be redacted to remove its directly accusatory elements, particularly when, as is typical, the confession

is introduced through the testimony of the officer who

received it.

Some States have long prohibited all codefendant

confessions that “refer[] to the defendant’s name or existence,” with no apparent breakdown in their ability

to prosecute crimes. Hanifa v. State, 505 S.E.2d 731,

738 (Ga. 1998); see also State v. Tucker, 414 S.E.2d

548, 554 (N.C. 1992) (“[I]n joint trials of defendants it

is necessary to exclude extrajudicial confessions unless all portions which implicate defendants other

than the declarant can be deleted without prejudice

either to the State or the declarant.”); cf. Gray, 523

U.S. at 197 (“[S]everal Circuits have interpreted Bruton similarly for many years, yet no one has told us of

21

any significant practical difficulties arising out of

their administration of that rule.” (citations omitted)).

There is no sound reason to expect a different result

in the federal system or other States.

It is of course true that joint trials serve an important function in the criminal justice system. See

Richardson, 481 U.S. at 209. But courts should be

“keenly aware that the claimed ‘efficiency’ of a joint

trial can be a surrogate for the reality that a joint trial

of multiple defendants is simply to the advantage of

the government.” United States v. Simmons, 374 F.3d

313, 318 (5th Cir. 2004). By contrast, “[d]efendants

gain virtually nothing from a separate trial.” Leipold,

supra, at 389. “They are not entitled to keep out evidence in a separate proceeding that would otherwise

be admissible.” Id. The only real benefit is that a separate trial “avoids the prejudice of joinder.” Id.

Excluding all directly accusatory confessions accounts for these disparate incentives and helps guard

against joint trials serving as “windfall[s]” for the government, at the expense of a defendant’s fundamental

rights. Bruton, 391 U.S. at 129. In a joint trial, the

only valid use of a codefendant’s confession is to establish one simple fact: The confessor admitted his guilt.

Thus, if a prosecutor objects to removing a confession’s

directly accusatory elements, it is fair to question

whether that resistance is based, at least in part, on

the possibility of the redacted content being used

against other defendants.

C.

In All Events, the Second Circuit’s Approach Should Be Rejected.

Whatever standard this Court ultimately adopts,

it must reject the Second Circuit’s approach. Bruton

22

explained the ineffectiveness of instructing a jury “to

disregard a codefendant’s confession implicating another defendant when it is determining that defendant’s guilt or innocence.” 391 U.S. at 130. And Gray

held that the confession in that case raised the same

concern because, “despite redaction, [it] obviously refer[red] directly to someone.” 523 U.S. at 196.

The standard developed by the Second Circuit asks

the wrong question and predictably produces wrong

answers. It does not even consider whether, as a practical matter, a confession “implicate[s] the defendant

in the crime.” Cruz, 481 U.S. at 191. To the contrary,

circuit precedent holds that “whether a jury might infer” that “a confederate . . . referenced the defendant”

is not “[t]he critical inquiry.” Jass, 569 F.3d at 61.

The Second Circuit instead asks whether replacing a

defendant’s name with a neutral term “sufficiently

conceals the fact of explicit identification.” Id. And it

cabins this analysis to “the redacted statement in isolation.” Id. at 62. This leads to the wooden rule that

the “introduction of a co-defendant’s confession with

the defendant’s name replaced by a neutral noun or

pronoun does not violate Bruton.” Lyle, 919 F.3d at

733.

The root flaw in the Second Circuit’s approach is

its mischaracterization of the prejudice Bruton guards

against. The Court has identified a codefendant’s implication of another defendant as the “context[] in

which the risk that the jury will not, or cannot follow

instructions is so great . . . that the practical and human limitations of the jury system cannot be ignored.”

Bruton, 391 U.S. at 135. But the Second Circuit recasts this risk as arising only from “explicit identifica-

23

tion.” Jass, 569 F.3d at 61. And because explicit identification can only come from the confession itself, the

court of appeals “view[s] the redacted statement in

isolation.” Id. at 62.

Redacting confessions to “conceal[] the fact of explicit identification” does not “eliminate the overwhelming probability that a jury hearing the confession at a joint trial will not be able to follow an appropriate limiting instruction.” Id. at 61. If the jury “realize[s] that the confession refers specifically to the

defendant,” Gray, 523 U.S. at 193, then Bruton prejudice applies with full force. And Gray explains why

“directly accusatory” confessions, as a class, “so closely

resemble Bruton’s unredacted statements that . . . the

law must require the same result.” Id. at 192, 194.

Finally, the Second Circuit’s misguided approach

is further reflected in the various justifications the

court has offered for its rule. For example, the court

sometimes indulges dubious theories to explain why a

jury might not reach the unavoidable conclusion that

the “other person” mentioned throughout a redacted

confession is the defendant. Perhaps, one theory goes,

the jury will think that the confessor “admitt[ed] his

own culpability . . . while shielding the specific identity of his confederate.” Lyle, 919 F.3d at 734. Even

assuming the jury might think that possible, but see

Bruton, 391 U.S. at 136 (recognizing a confessor’s

strong “motivation to shift blame onto others”), this

Court’s precedents do not allow a defendant’s constitutional rights to hang on such a doubtful proposition.

See Lee, 476 U.S. at 545 (noting a confessor’s “desire

to shift or spread blame, curry favor, avenge himself,

or divert attention to another”). In other cases, like

the decision below, the court merely observes that “a

24

juror . . . could have concluded” that the confession referred to someone other than petitioner. Pet. App. 11a

(emphasis added). That reasoning turns Bruton on its

head by asking whether it is possible that petitioner’s

Confrontation Clause rights were not violated by his

jury “look[ing] to the incriminating extrajudicial

statements in determining [his] guilt.” Bruton, 391

U.S. at 126. That is the precise supposition that Bruton forbids. This Court should not countenance it.

25

CONCLUSION

For the foregoing reasons, Amici urge this Court to

reverse the decision of the court of appeals and hold

that admission of a nontestifying codefendant’s directly accusatory confession violates the Confrontation Clause.

Respectfully submitted,

Jeffrey L. Fisher

Co-Chair, NACDL AMICUS

COMMITTEE

559 Nathan Abbott Way

Stanford, CA 94305

David D. Cole

AMERICAN CIVIL LIBERTIES

UNION FOUNDATION

915 15th St. N.W.

Washington, D.C. 20005

February 1, 2023

Jonathan Y. Ellis

Counsel of Record

APPELLATE JUSTICE

INITIATIVE AT

MCGUIREWOODS LLP

888 16th St. N.W.

Suite 500

Washington, D.C. 20006

(202) 828-2887

jellis@mcguirewoods.com

Gregory J. DuBoff

APPELLATE JUSTICE

INITIATIVE AT

MCGUIREWOODS LLP

800 East Canal St.

Richmond, VA 23219