Amicus Curiae Brief — David Asa Villarreal, Petitioner v. Texas

No. 24-557

IN THE

Supreme Court of the United States

_____________

DAVID ASA VILLARREAL,

v.

TEXAS,

_____________

Petitioner,

Respondent.

On Writ of Certiorari

to the Court of Criminal Appeals of Texas

_____________

BRIEF OF NATIONAL ASSOCIATION OF

CRIMINAL DEFENSE ATTORNEYS AS

AMICUS CURIAE ON BEHALF OF

PETITIONER

_____________

BARBARA E. BERGMAN

Co-Chair, AMICUS COMMITTEE

NATIONAL ASSOCIATION OF

CRIMINAL DEFENSE LAWYERS

1201 E. Speedway Blvd.

Tucson, AZ 85721

(520) 621-3984

bbergman@arizona.edu

ANDREA ROTH

Counsel of Record

UC BERKELEY SCHOOL OF LAW

347 Law Building

Berkeley, CA 94720

(510) 643-6092

aroth@law.berkeley.edu

Counsel for Amicus Curiae

QUESTION PRESENTED

Whether a trial court abridges the defendant's Sixth

Amendment right to counsel by prohibiting the defendant and his counsel from discussing the defendant's

testimony during an overnight recess.

(i)

ii

TABLE OF CONTENTS

Page

QUESTION PRESENTED .......................................... i

TABLE OF CONTENTS ............................................. ii

TABLE OF AUTHORITIES ......................................iii

INTEREST OF AMICUS CURIAE ............................ 1

SUMMARY OF ARGUMENT .................................... 2

ARGUMENT ............................................................... 4

I. An order prohibiting counsel from discussing

a defendant’s testimony during a significant

break in trial is fundamentally incompatible

with the Sixth Amendment right to counsel. ....... 4

A. Counsel cannot meaningfully assist a

criminal defendant without being able

to discuss the defendant’s testimony

during an overnight recess in that

testimony. .................................................... 5

B. Any attempt to distinguish discussions

of testimony from discussions of the

case in general, in an order ostensibly

prohibiting only the former, is

inherently unworkable.. ........................... 10

II. A Geders violation during a significant break

in the defendant’s testimony is structural

error ...................................................................... 14

CONCLUSION .......................................................... 26

iii

TABLE OF AUTHORITIES

Page

CASES

Arizona v. Fulminante, 499 U.S. 279 (1991)............. 15

Brooks v. Tennessee, 406 U.S. 605 (1972) ............. 9, 10

Chapman v. California, 386 U.S. 18 (1967) .............. 15

Clark v. State, 301 A.3d 241 (Md. 2023) ................... 22

Coleman v. Alabama, 399 U.S. 1 (1970) ................... 19

Commonwealth v. Johnson, 828 A.2d 1009 (Pa. 2003)

................................................................................. 22

French v. Jones, 332 F.3d 430 (6th Cir. 2003) .......... 22

Geders v. United States, 425 U.S. 80 (1976)

............................................................... 2, 5, 9, 13, 17

Gideon v. Wainwright, 372 U.S. 335 (1963).............. 26

Gilbert v. California, 388 U.S. 263 (1967) ................ 19

Hamilton v. Alabama, 368 U.S. 52 (1961) ................ 18

Jackson v. United States, 422 A.2d 1202 (D.C. 1979)

(en banc) .................................................................. 22

Martin v. United States, 991 A.2d 791 (D.C. 2010)

....................................................................... 8, 14, 22

Massiah v. United States, 377 U.S. 201 (1964) ........ 20

Mickens v. Taylor, 535 U.S. 162 (2002) ............... 18, 20

Milton v. Wainwright, 407 U.S. 371 (1972) .............. 20

Moore v. Commonwealth, 771 S.W.2d 34 (Ky. 1988) ....

................................................................................. 23

Moore v. Illinois, 434 U.S. 220 (1977). ...................... 19

Moore v. Purkett, 275 F.3d 685 (8th Cir. 2001) ........ 22

Mudd v. United States, 798 F.2d 1509 (D.C. Cir.

1986) .................................................... 2, 8, 14, 22, 25

Neder v. United States, 527 U.S. 1 (1999) ................. 15

People v. Hernandez, 273 P.3d 1113 (Cal. 2012) ...... 24

Perry v. Leeke, 488 U.S. 272 (1989) ... 2, 5, 9, 13, 14, 17

iv

Petty v. United States, 317 A.3d 351 (D.C. 2024) . 8, 14

Portuondo v. Agard, 529 U.S. 61 (2000)................ 9, 10

Rose v. State, 304 P.3d 387 (Mont. 2013) .................. 23

Sanders v. Lane, 861 F.2d 1033 (7th Cir. 1988) ....... 24

Satterwhite v. Texas, 486 U.S. 249 (1988), ... 20, 21, 24

State v. Smith, 375 So.3d 654 (La. App. 2023) ......... 24

Sullivan v. Louisiana, 508 U.S. 275 (1993) .............. 15

United States v Roy, 855 F.3d 1133 (11th Cir. 2017)

................................................................................. 23

United States v. Cobb, 905 F.2d 784 (4th Cir.1990) ... 8

United States v. Cronic, 466 U.S. 648 (1984)............ 18

United States v. Gonzalez-Lopez, 548 U.S. 140 (2006)

..................................................................... 15, 16, 18

United States v. Johnson, 267 F.3d 376 (5th Cir.

2001) ........................................................................ 21

United States v. Miguel, 111 F.3d 666 (9th Cir. 1997)

................................................................................. 21

United States v. Nelson, 884 F.3d 1103 (11th Cir.

2018) ........................................................................ 23

United States v. Russell, 205 F.3d 768 (5th Cir. 2000)

................................................................................. 22

United States v. Sandoval–Mendoza, 472 F.3d 645

(9th Cir. 2006) ........................................................... 7

United States v. Santos, 201 F.3d 953 (7th Cir. 2000)

............................................................................. 8, 13

United States v. Torrez, 997 F.3d 624 (5th Cir. 2021)

................................................................................. 21

United States v. Triumph Cap. Grp., Inc., 487 F.3d

124 (2d Cir. 2007) ................................................... 21

United States v. Wade, 388 U.S. 218 (1967) ............. 19

Wallace v. State, 851 So.2d 216 (Fla. 2003) .............. 23

White v. Maryland, 373 U.S. 59 (1963) ..................... 18

v

OTHER AUTHORITY

Brief for Petitioner, Coleman v. Alabama (1970), (No.

68-72) ....................................................................... 20

Brief for Respondent, Coleman v. Alabama, 399 U.S.

1 (1970) (No. 68-72). ............................................... 20

Brief for United States, Geders v. United States

(1976) (No. 74-5968). .............................................. 17

Oral Argument, Coleman v. Alabama, Nov. 18, 1969,

available at https://www.oyez.org/cases/1969/72 .. 20

INTEREST OF AMICUS CURIAE

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 and has a nationwide membership of

many thousands of direct members and up to 40,000

attorneys in affiliate organizations. NACDL is dedicated to advancing the proper, efficient, and fair administration of justice. NACDL files many amicus

briefs each year in this Court and other federal and

state courts, seeking to assist in cases presenting issues important to criminal defendants, criminal defense lawyers, and the criminal system.1

NACDL also has a particular interest in this case,

as it directly affects the ability of criminal defense lawyers to advise their clients at trial, to make informed

strategic decisions about the defense, and to protect

privileged attorney-client communications.

1

Pursuant to this Court’s Rule 37.6, counsel for amicus curiae state that this

brief was not authored in whole or in part by counsel for any party, and that

no person or entity other than amicus and its counsel made a monetary contribution to the preparation or submission of this brief.

2

SUMMARY OF ARGUMENT

A trial court’s order prohibiting an accused and his

counsel from “discussing” the accused’s testimony during an overnight recess is equivalent to a Geders violation and requires automatic reversal.

First, any meaningful overnight consultation between an accused and his lawyer during the accused’s

testimony will include discussion of that testimony. Indeed, this Court in Geders v. United States, 425 U.S.

80 (1976), and Perry v. Leeke, 488 U.S. 272 (1989), explicitly acknowledged that an overnight recess in testimony will inevitably involve discussion of that testimony. The defendant’s testimony is relevant to a host

of critical trial issues, such as whether to plead guilty,

whether to pursue additional investigation, decisions

about the rest of the defense case, and how to litigate

jury instructions and handle upcoming closing arguments. Accordingly, several lower courts have recognized prohibitions on discussing testimony during an

overnight recess as a Geders violation. See, e.g., Mudd

v. United States, 798 F.2d 1509, 1515 (D.C. Cir. 1986));

id. (Scalia, J., concurring) (“I therefore join in the majority's holdings that a prohibition on attorney-defendant discussion during substantial recesses, even if limited to discussion of testimony, violates the sixth

amendment and that . . . it constitutes per se reversible error.”).

Moreover, even if a prohibition narrowly targeted at

“discussion of testimony” were theoretically both possible and compatible with Geders, such a prohibition is

impossible in practice and therefore the equivalent of

the order in Geders itself. No lawyer in the throes of

trial can do his constitutional duty if burdened with

distinguishing between discussions of “testimony” and

“non-testimonial matters” for purposes of complying

3

with such an order. Of course, lawyers are already expected to avoid brazen attempts to improperly influence a witness, which are banned by ethical rules. But

nearly every conversation a lawyer might have in an

overnight recess during the defendant’s testimony

might arguably relate, even if indirectly, to that testimony. If preoccupied by trying to comply with such a

confusing mandate, at the risk of his bar card, a lawyer

will be chilled from having the type of robust discussions with his client needed to satisfy the Sixth

Amendment.

To the extent a trial court wishes to further control

counsel’s ability to “coach” a defendant, it can – and

must – do so without violating the Sixth Amendment.

The trial court is always free to limit breaks in a defendant’s testimony to shorter periods; to ensure that

cross-examination follows immediately after direct;

and to allow the prosecution latitude to cross-examine

the defendant with any evidence of coaching.

In turn, a Geders violation interfering with attorneyclient communications during a significant break in

the defendant’s testimony is structural error. This

Court said as much implicitly in Geders, and explicitly

in Perry. But even if this Court had never spoken on

the matter, no other approach would be possible. The

inability to meaningfully strategize with one’s attorney about key issues on the eve of what may be the last

and most important day of trial has a prejudicial effect

that is exceedingly hard to quantify. It affects numerous aspects of the trial, and potentially even whether

the trial will continue at all. And a prejudice analysis

in this context would involve invasive appellate probing of privileged attorney-client communications. Tellingly, the few times this Court or lower courts have

applied harmless error analysis to a claim of denial of

counsel at a critical stage have been (1) where the

4

court found no error to begin with because the stage

was not “critical”; (2) where a would-be Geders error

involved a short break with an immediate subsequent

chance for consultation; or (3) where the error related

exclusively to one identifiable piece of evidence at trial.

Here, as in Geders itself, the error is not capable of being cabined in such a way; rather, it went to the heart

of counsel’s ability to assist the accused at trial.

ARGUMENT

I. AN ORDER PROHIBITING COUNSEL

FROM DISCUSSING A DEFENDANT’S TESTIMONY DURING A SIGNIFICANT BREAK

IN TRIAL IS FUNDAMENTALLY INCOMPATIBLE WITH THE SIXTH AMENDMENT

RIGHT TO COUNSEL.

The trial court’s order in this case is incompatible

with the Sixth Amendment in two respects. First,

counsel cannot meaningfully advise a criminal defendant about the case during a significant break in trial

without being able to discuss the defendant’s testimony. Second, because of the inherent difficulty in distinguishing between matters related and unrelated to

the defendant’s testimony, such an order – even if its

purportedly narrow focus on testimony were somehow

justified in the abstract – is unworkable and has an

impermissible chilling effect on attorney-client communications during a critical stage in trial.

5

A. Counsel cannot meaningfully assist a

criminal defendant without being able to

discuss the defendant’s testimony during

an overnight recess in that testimony.

This Court in Geders v. United States, 425 U.S. 80

(1976), recognized that a defendant must be able to

speak with counsel about the case during an overnight

recess, even about matters related to the defendant’s

testimony. While the unconstitutional order in Geders

broadly prohibited all attorney-client communications

during the recess, this Court noted that those communications will often involve discussing the testimony

itself. For example, a lawyer may need “to obtain from

his client information made relevant by the day's testimony, or he may need to pursue inquiry along lines

not fully explored earlier.” 425 U.S. at 88. The Court

also mentioned the “important ethical distinction between discussing testimony and seeking improperly to

influence it,” 425 U.S. at 90 n.3 (emphasis added), indicating that the former is entirely proper.

In Perry v. Leeke, 488 U.S. 272 (1989), this Court

again acknowledged the inevitability that attorney-client communications during a significant break in a defendant’s testimony will involve discussing a client’s

testimony. The Court reaffirmed an accused’s “right to

unrestricted access to his lawyer for advice on a variety of trial-related matters” during an overnight recess

in his testimony even while noting that these “discussions will inevitably include some consideration of the

defendant’s ongoing testimony.” Id. at 284. In contrast,

as the Perry Court eventually held, a mere fifteen-minute break in testimony is not a critical time for counsel and client to discuss the case as a whole or how the

defendant’s testimony affects it. Id. at 285.

6

Beyond the examples given in Geders and Perry,

there are a host of other ways in which a criminal defendant’s ability to meaningfully discuss the case in

general with his lawyer during an overnight recess in

his testimony will require some direct or indirect discussion of that testimony, including:

• Facilitating compliance with court orders by reminding the defendant not to discuss excluded or otherwise inadmissible evidence;

• Discussing strategy related to requesting jury

instructions made newly relevant by the defendant’s

testimony and likely to be given shortly after the defendant’s testimony;

• Warning the defendant about potential questions that could raise self-incrimination concerns;

• Considering whether to seek a mistrial based on

something that happened in the courtroom related to

the testimony (e.g. an elicited remark, a jury note related to the testimony);

• Discussing what to argue or emphasize during

closing argument, based on the content and strength

of the defendant’s testimony;

• Contemplating the need to subpoena new witnesses or obtain physical or documentary evidence

based on new leads from a defendant’s direct or crossexamination;

• Evaluating the strength or weakness of the testimony in the context of deciding whether to plead

guilty before the end of trial;

• Assessing whether the defendant’s testimony

established a strong alibi defense and, thus, whether

to abandon plans to call an alternative alibi witness

whose credibility might be attacked;

7

• Correcting what would otherwise be inaccurate

testimony (whether inadvertent or otherwise) so the

attorney does not suborn false testimony;

• Determining the pros and cons of raising competence, if the defendant’s testimony newly suggests

he may not be competent to proceed with the rest of

trial;

• Offering basic information about what will happen the next day in terms of the completion of direct

examination and the nature and purpose of cross- and

redirect-examination;

• Discussing the client’s concerns about whether

his testimony will subject him to being targeted or

killed and what to do about it;

• Explaining legal concepts that arose during testimony, such as a hearsay objection, if the client asks;

• Informing the defendant, based on new investigation, that an alleged felony on his record is actually

a misdemeanor that cannot be used to impeach him; or

• Deciding whether the defendant’s testimony

went sufficiently well before the jury that the defense

should decline to call other witnesses who might be

discredited.

None of these critically important topics would be permissible attorney-client communications under the

trial court’s order in this case.

Not surprisingly, then, several lower courts have

recognized that limitations on discussing testimony violate Geders. See, e.g., United States v. Sandoval–Mendoza, 472 F.3d 645, 651 (9th Cir. 2006) (‘‘[I]t is hard to

see how a defendant’s lawyer could ask him for the

name of a witness who could corroborate his testimony

8

or advise him to change his plea after disastrous testimony . . . without discussing the testimony itself.’’);

United States v. Santos, 201 F.3d 953, 965 (7th Cir.

2000) (noting that a prohibition on discussion of testimony ‘‘would as a practical matter preclude the assistance of counsel across a range of legitimate legal and

tactical questions, such as warning the defendant not

to mention excluded evidence’’); United States v. Cobb,

905 F.2d 784, 792 (4th Cir.1990) (‘‘To remove from [an

accused] the ability to discuss with his attorney any

aspect of his ongoing testimony [would] effectively

eviscerate[] his ability to discuss and plan trial strategy.’’); Mudd, 798 F.2d at 1512 (“The only logical implication [of Geders] is that the Court also meant to

forbid prohibitions on attorney/defendant discussions

of the defendant's testimony during a substantial recess, not just blanket prohibitions during such a recess.”); id. at 1515 (Scalia., J., concurring) (agreeing

with Court’s holding on this point); Petty v. United

States, 317 A.3d 351, 352-53 (D.C. 2024) (holding that

prohibition on an attorney-defendant discussion of testimony during overnight recess required reversal); id.

at 356 (quoting Martin v. United States, 991 A.2d 791,

795 (D.C. 2010)) (noting that the “majority of the federal circuits agree that ‘under Perry and Geders’ a trial

court ‘may not order a defendant to refrain from discussing his ongoing testimony with counsel during an

overnight recess, even if all other communication is allowed’”).

Nor will the prosecution be prejudiced by allowing

unfettered communications between the accused and

his counsel during an overnight recess in the accused’s

testimony. While it is always possible that an attorney

will attempt to “coach” or improperly influence a defendant during a recess in testimony, the ethical rules

themselves already prohibit such witness coaching,

9

without the need for a court order that broadly prohibits discussions of testimony between an attorney and

client. See Geders, 425 U.S. at 90 n.3 (citing ethical

rules barring an attorney’s improper influencing of a

client’s testimony).

Moreover, because the possibility of coaching witnesses exists from the very beginning of trial, there is

little additional truth-enhancing value in an overbroad prohibition on discussing testimony during a

particular overnight recess. See Geders, 425 U.S. at 88

(noting that a rule prohibiting discussion of testimony

“accomplishes less when it is applied to the defendant,” given that he “has the opportunity to discuss his

testimony with his attorney up to the time he takes the

witness stand.”). Cf. Perry, 488 U.S. at 282 (noting that

the “poise and sense of strategy” that a witness regains

after being able to take a break and consult with their

attorney would happen during an overnight recess

whether or not the attorney “coached” the witness).

Instead of hampering attorney-client communications during one of the most critical moments in trial,

trial courts should instead use “other ways to deal with

the problem” of client coaching. Geders, 425 U.S. at 89.

First, the prosecution “may cross-examine a defendant

as to the extent of any ‘coaching’ during a recess,” and

“exploit” any inculpatory answers “in closing argument.” Id. at 89-90. This remedy is similar to that approved by the Court to account for the defendant’s ability to sit through other witnesses’ testimony before testifying himself. See Brooks v. Tennessee, 406 U.S. 605,

613 (1972) (holding that the rule on sequestering witnesses cannot be enforced against a criminal defendant); Portuondo v. Agard, 529 U.S. 61, 75 (2000) (holding that the prosecution may comment on the defendant’s ability to avoid sequestration). Second, the trial

judge, “if he doubts that defense counsel will observe

10

the ethical limits on guiding witnesses, may direct that

the examination of the witness continue without interruption until completed.” Id. at 90. Third, the trial

judge is always free to “arrange the sequence of testimony so that direct- and cross-examination of a witness will be completed without interruption.” Id.

Ultimately, of course, “to the extent that conflict remains between” the right to consult one’s attorney and

the prosecution’s interest in minimizing “the risk of

improper ‘coaching,’” “the conflict must, under the

Sixth Amendment, be resolved in favor of the right to

the assistance and guidance of counsel.” Id. at 91 (citing Brooks, 406 U.S. 605). But in any event, the resolution of any perceived concerns about improper influence cannot be to prohibit an accused from discussing

his testimony with his attorney during an overnight

break in that testimony.

B. Any attempt to distinguish discussions of

testimony from discussions of the case in

general, in an order ostensibly prohibiting only the former, is inherently unworkable.

Even if the Sixth Amendment were theoretically

compatible with an order prohibiting only discussions

of a client’s testimony and not discussions about the

case more broadly, any attempt to tailor a prohibition

to apply only to the former and not the latter is unworkable in practice. Because of the inherent difficulty

in determining which communications will be deemed

“testimony” related, a lawyer cannot zealously advocate for the accused if preoccupied with navigating

such a confusing prohibition.

Faced with potential sanctions for violating an order

like the one here, an attorney may well be discouraged

11

from having critical conversations with the accused

where the testimonial/non-testimonial line is unclear.

For example, nearly all the examples listed above of

communications involving “discussion of testimony”

could also be deemed discussions of trial strategy. But

if they are both, do they violate the order? More precisely, will a given trial judge decide that they violate

the order? While even an appellate court in the light of

day might have difficulty determining whether some

attorney-client communications are related to the defendant’s testimony, an attorney in the throes of trial

would have even more difficulty deciding what does

and does not comply with an order like the one here.

To further visualize how difficult it would be for a

defendant and his counsel to navigate such an order,

one could imagine the following types of back-andforth during an overnight recess:

Defendant: “I’m really nervous and afraid. Testifying is terrifying. Do you think I’m going to be okay?”

Attorney: “I’m sorry; I’m not sure if we can discuss

that.”

Defendant: “How much longer will my testimony

last?”

Attorney: “I’m sorry; I’m not sure if we can discuss

that.”

Defendant: “Well, can you at least remind me what

a cross-examination and redirect examination are, and

what I can expect?”

Attorney: “I’m sorry; I’m not sure if we can discuss

that.”

Defendant: “Do you think my testimony today might

lead me to be targeted or even killed? Can I seek protection?”

12

Attorney: “I’m sorry; I’m not sure if we can discuss

that.”

Defendant: “What did the prosecutor mean by ‘hearsay’ when he objected to me telling the jury what I told

the police that night?”

Attorney: “I’m sorry; I’m not sure if we can discuss

that.”

Defendant: “That supposed felony conviction you

said they might bring up – have you figured out

whether it’s actually a misdemeanor?”

Attorney: “I’m sorry; I’m not sure if we can discuss

that.”

Defendant: “Do you think the jury will believe my

testimony over the informant? If not, do you think we

need to put on more of a defense case tomorrow?”

Attorney: “I’m sorry; I’m not sure if we can discuss

that.”

Defendant: “Can you track down that other alibi witness I mentioned today? His address is –"

Attorney: “I’m sorry; I’m not sure if we can discuss

that.”

Similarly, one could imagine a defense attorney under

the cloud of such an order having to hesitate before

comforting a weeping client who is worried the jury

will not believe his truthful testimony; before explaining to a juvenile client that his mother will be able to

watch the rest of his testimony in the morning, to calm

him down; or before preparing the client to hear the

judge give the standard jury instruction on a defendant’s testimony. There are as many other examples as

there are criminally accused clients, each of whom

have different needs, issues, and questions.

13

By forcing both counsel and the accused to secondguess their every utterance during what may be one of

the most important conversations in the client’s entire

life, such an order jeopardizes counsel’s basic ability to

do his constitutional duty. Awkwardly hamstrung in

this fashion, counsel cannot meaningfully assist a

criminal defendant with his defense.

This Court recognized this reality in Geders itself.

From the beginning, the trial court in Geders “expressed some doubt” that Geders’ attorney “would be

able to” confine his discussion to matters other than

the imminent cross-examination. 425 U.S. at 82. This

Court nonetheless determined that the attorney required unfettered access to the accused during the

overnight recess and that any concerns about coaching

be resolved in favor of access to counsel. Id. at 91. If

the Geders Court had deemed a more narrowly tailored

prohibition on discussing testimony to be workable

and appropriate, it could have easily suggested that as

an obvious solution. Instead, the Court appears to

have understood that the only tenable option is to allow unfettered attorney-client communication and

deal with “coaching” concerns through the tools of the

adversarial system.2

Courts construing Geders have also acknowledged

the difficulty in determining whether discussions are

related to testimony, and the chilling effect this uncertainty could have on lawyers trying to be effective advocates. See, e.g., Santos, 201 F.3d 953, 966 (7th Cir.

2

In contrast, as the Perry Court noted, there is a “virtual certainty” that any

attorney-client communication in a brief fifteen-minute recess during a client’s testimony will relate only to the client’s testimony, rather than the case

as a whole. 488 U.S. at 283. Thus, the prohibition in Perry did not present

the same unworkable dilemma as the one here.

14

2000) (describing a similar order as “confusing marching orders”); Petty, 317 A.3d at 353 (striking down a

similar prohibition on discussing testimony during an

overnight recess and noting that counsel said he was

“chilled as to what [he] could discuss” with his client

under the order); Martin, 991 A.2d at 794 n.13 (citing

cases explaining why “[c]onsultations between lawyers

and clients cannot be neatly divided into discussions

about ‘testimony’ and those about ‘other’ matters” (internal quotation marks omitted)); Perry, 488 U.S. at

295 n.8 (Marshall, J., dissenting) (noting that if “direct

examination of the defendant inadvertently elicits

damaging information that can effectively be neutralized on redirect only if the defendant has the opportunity to explain his direct testimony to counsel,”

counsel might wonder whether “the ensuing attorneydefendant discussion” is “about trial strategy” or

“about upcoming testimony”); id. (predicting “a

chilling effect on cautious attorneys, who might avoid

giving advice on nontestimonial matters for fear of violating [an order barring discussion of testimony]”)

(quoting Mudd, 798 F.2d at 1512).

II. A GEDERS VIOLATION DURING A SIGNIFICANT BREAK IN THE DEFENDANT’S TESTIMONY IS STRUCTURAL ERROR.

This Court, and most lower courts to have addressed the issue, have repeatedly stated that Geders

violations during an overnight recess require an automatic retrial. Indeed, no other approach would make

sense, given the inherently diffuse and indeterminate

effects of such a violation, as well as the privileged nature of attorney-client communications. Moreover, unlike trial errors caused by counsel’s mistakes, a prosecutor’s overreach, or a trial court’s attempt to resolve

difficult novel legal issues as best it can, a Geders error

15

like the one here involve an affirmative step by a judge

to interfere with attorney-client communications and

is entirely within the trial court’s control to avoid. In

this sense, Geders errors are analogous to faulty reasonable doubt instructions, which – not coincidentally

– are also structural errors.

This Court has divided constitutional errors into

two classes: (1) “trial errors” that happen “during

presentation of the case to the jury” and whose effect

can “be quantitatively assessed” under the harmlessness standard of Chapman v. California, 386 U.S. 18

(1967); and (2) “structural defects,” which “defy analysis” under Chapman because they involve “consequences that are necessarily unquantifiable and indeterminate.” United States v. Gonzalez-Lopez, 548 U.S.

140, 149-50 (2006) (quoting Sullivan v. Louisiana, 508

U.S. 275, 282 (1993)). Trial errors thus involve constitutional violations that at most affect a particular witness, piece of evidence, or element. See, e.g., Arizona v.

Fulminante, 499 U.S. 279, 295 (1991) (subjecting erroneous admission of coerced confession to Chapman

test); Neder v. United States, 527 U.S. 1, 7-9 (1999) (applying Chapman to a failure to instruct on a single element supported by overwhelming evidence of guilt).

In contrast, this Court has treated as structural those

errors affecting multiple aspects of the case: the denial

of the right to counsel, self-representation, or public

trial, as well as denial of the jury right through a faulty

reasonable doubt instruction. Gonzalez-Lopez, 548

U.S. at 150 (citations omitted).

Like other structural errors, a Geders violation during an overnight recess in a defendant’s testimony is

not specific to a particular witness, piece of evidence,

or element. Rather, a prohibition on attorney-client

discussions of a defendant’s testimony during that testimony potentially affects not only the presentation of

16

the defendant’s testimony but a host of other aspects

of trial. As explained in Part I, these aspects include

the decision whether to plead guilty, given how the testimony came out; the need for protection from retaliation; whether to request a mistrial; discussions of

strategy related to jury instructions and closing argument; the decision whether to subpoena additional

documents or objects or call additional defense witnesses; and the raising of competence as an issue.

In turn, the prejudicial effect of denying the accused a chance to discuss these matters with his lawyer at a critical point in trial is too hard to quantify.

As this Court explained in holding that denial of retained counsel of one’s choice is structural error:

Different attorneys will pursue different strategies with regard to investigation and discovery, development of the theory of defense, selection of the jury, presentation of the witnesses,

and style of witness examination and jury argument. And the choice of attorney will affect

whether and on what terms the defendant cooperates with the prosecution, plea bargains, or

decides instead to go to trial. In light of these

myriad aspects of representation, the erroneous

denial of counsel bears directly on the “framework within which the trial proceeds,” Fulminante, supra, at 310—or indeed on whether

it proceeds at all.

Gonzalez-Lopez, 548 U.S. at 150. In the same respect,

a prohibition on discussing issues that go to the heart

of the defense strategy at a key point in trial also implicates “myriad aspects of representation.” Indeed,

being able to discuss a defendant’s testimony can be

critical to deciding whether to plead guilty, i.e.,

“whether [the trial] proceeds at all.” Id. Thus, applying

17

Chapman “in such a context would be a speculative inquiry into what might have occurred in an alternate

universe.” Id.

Accordingly, this Court in both Geders and Perry

treated a Geders violation involving an overnight recess during a defendant’s testimony as requiring automatic reversal. The Geders Court reversed based solely

on the finding of a right-to-counsel violation, without

any inquiry into prejudice, 425 U.S. at 91, even though

the government’s brief urged the Court to affirm on

this alternative ground. See Brief for United States at

12-24, Geders v. United States (1976) (No. 74-5968).

Nor did the Geders majority take issue with or otherwise comment on Justices Marshall’s characterization

in his concurrence that “the Court holds” that a defendant establishing a Geders violation “need not

make a preliminary showing of prejudice.” Id. at 92

(Marshall, J., concurring).

This Court in Perry went even further, explicitly rejecting a prejudice analysis for the type of error in

Geders. It first noted that the Geders Court’s automatic reversal was both proper and “consistent with”

the Court’s right-to-counsel precedents. 488 U.S. at

279. It next explained the difference between actual

ineffectiveness claims, which are subject to prejudice

analysis, and actual or constructive denials of counsel

at critical stages of the proceedings, which are not. Id.

at 280 (noting that a Geders violation is the latter, not

the former). The Perry Court ultimately concluded that

it “c[ould not] accept the rationale” of the lower court

that a Geders violation can be harmless beyond a reasonable doubt under Chapman. Id.

Beyond the Geders context, this Court has repeatedly reaffirmed that no prejudice showing is necessary

when a Sixth Amendment violation stems from denial

18

of counsel at a critical stage. “We have spared the defendant the need of showing probable effect upon the

outcome, and have simply presumed such effect, where

assistance of counsel has been denied . . . during a critical stage of the proceeding.” Mickens v. Taylor, 535

U.S. 162, 166 (2002) (holding, in contrast to the rule in

critical-stage cases, that a petitioner must show prejudice to merit a new trial on ineffective-assistance-ofcounsel grounds based on a conflict of interest of which

the trial court was unaware). See also United States v.

Cronic, 466 U.S. 648, 659 n.25 (1984) (citing Geders for

the proposition that this Court has “uniformly found

constitutional error without any showing of prejudice

when counsel was ... prevented from assisting the accused during a critical stage of the proceeding”); id. at

659 (explaining that “[t]he presumption that counsel's

assistance is essential requires us to conclude that a

trial is unfair if the accused is denied counsel at a critical stage of his trial”); White v. Maryland, 373 U.S. 59

(1963) (holding that admission at trial of guilty plea

made at preliminary hearing without counsel required

a new trial, and noting “we do not stop to determine

whether prejudice resulted”); Hamilton v. Alabama,

368 U.S. 52, 55 (1961) (holding that where counsel is

denied at arraignment in a capital case, “the degree of

prejudice can never be known” because “only the presence of counsel could have enabled this accused to

know all the defenses available to him and to plead intelligently”); Cf. Gonzalez-Lopez, 548 U.S. at 150 (distinguishing denial-of-counsel claims from ineffectiveness claims on grounds that the “requirement of showing prejudice in ineffectiveness claims stems from the

very definition of the right”).

Nonetheless, there appear to be a small handful of

cases in which this Court has indicated that Chapman

applies to a particular claim of denial of counsel during

19

a critical stage. But none is analogous to a case like

this one, involving interference with attorney-client

discussions during an overnight recess or involving a

defendant’s testimony. And all but one of these cases

involve errors that undisputedly affected only a single

piece of evidence offered at trial.

In the first of these cases, United States v. Wade, 388

U.S. 218 (1967), the petitioner argued he was unconstitutionally denied counsel at a post-arraignment

lineup identification procedure and that elicitation of

a subsequent in-court identification was the “fruit” of

that error. Id. at 235, 244.3 Thus, the crux of Wade’s

claim was that a particular piece of evidence – the incourt identification – was erroneously admitted. While

this Court agreed with petitioner that a lineup should

be considered a “critical stage,” id. at 237, it remanded

for the trial court to determine whether the in-court

identification had an “independent source” (and was

thus not a fruit) and whether its admission was harmless under Chapman. Id. at 242-43. See also Gilbert v.

California, 388 U.S. 263, (1967) (remanding for determination of harmlessness of Wade error); Moore v. Illinois, 434 U.S. 220, 232 (1977) (same). But the error

in Wade was qualitatively different from a Geders error in that it involved counsel’s absence at a stage relevant only to rebutting a single item of evidence. Thus,

the ultimate issue in Wade – the erroneous admission

of an item of evidence that is a fruit of a right-to-counsel violation – lends itself to Chapman review in a way

that a Geders error does not.

In the second of these cases, Coleman v. Alabama,

399 U.S. 1 (1970), the Court held that a preliminary

3

The lineup identification was not introduced by the government at trial

and was only elicited by defense counsel on cross-examination. Wade, 388

U.S. at 220.

20

hearing was a “critical stage” for right-to-counsel purposes but remanded for a determination of whether denial of counsel was harmless under Chapman. Id. at

10-11. The parties in Coleman did not cite Chapman

or otherwise brief or mention at oral argument the issue of harmlessness,4 other than a conclusory statement without citation in the state’s brief that the absence of counsel did not “substantially affect the rights

of accused on trial.”5 The Court’s reasons for invoking

Chapman are therefore obscure, especially in light of

the Court’s later insistence that a defendant is

“spared” the “need of showing probable effect upon the

outcome” if “counsel has been denied . . . during a critical stage of the proceeding.” Mickens, 535 U.S. 166.

In the third case, Milton v. Wainwright, 407 U.S. 371

(1972), this Court held that admission of a confession

obtained in violation of Massiah v. United States, 377

U.S. 201 (1964) (holding that a post-arraignment interrogation of a defendant by government agents was

a “critical stage” at which the right to counsel applied)

was harmless beyond a reasonable doubt under Chapman. Again, the denial of counsel related only to one

identifiable piece of evidence, allowing the harm analysis to focus on the effect of that piece of evidence (the

confession).

And in Satterwhite v. Texas, 486 U.S. 249 (1988),

this Court declined to presume prejudice where a defendant was unconstitutionally denied counsel during

a pretrial psychiatric evaluation by a state expert. Id.

4

See Brief for Petitioner, Coleman v. Alabama (1970), (No. 68-72), 1969

WL 119859 (May 20, 1969); Oral Argument, Coleman v. Alabama, Nov.

18, 1969, available at https://www.oyez.org/cases/1969/72.

5

Brief for Respondent at 6, Coleman v. Alabama, 399 U.S. 1 (1970) (No.

68-72).

21

at 257-58. This Court relied on the fact that the error

was “limited to the admission into evidence of [the expert’s] testimony” and noted that “[w]e have permitted

harmless error analysis . . . where the evil caused by a

Sixth Amendment violation is limited to the erroneous

admission of particular evidence at trial.” Id. Again,

such an error is qualitatively different from a Geders

error disallowing key conversations about multiple aspects of trial.

The majority of lower courts also appear to agree

that a Geders violation interfering with attorney-defendant communications during an overnight recess is

structural error requiring automatic reversal. See, e.g.,

United States v. Torrez, 997 F.3d 624, 627 (5th Cir.

2021) (describing Geders error, when objected to, as

“presumptively prejudicial,” and reversing in this case

for plain error); United States v. Triumph Cap. Grp.,

Inc., 487 F.3d 124, 131 (2d Cir. 2007) (explaining that

it is “well settled that, in the Geders context, a violation of a defendant's Sixth Amendment right to counsel ... constitutes a structural defect which defies

harmless error analysis and requires automatic reversal” (internal quotation marks omitted)); United States

v. Johnson, 267 F.3d 376, 379 (5th Cir. 2001) (explaining that, in Perry, the Supreme Court “held that a

showing of prejudice is not an essential component of

establishing a violation of the Geders rule” and expressing support for that view as “the constitutional

right to counsel warrants the most zealous protection”); Moore v. Purkett, 275 F.3d 685, 689 (8th Cir.

2001) (explaining that a Geders violation is “reversible

without a showing of prejudice”); United States v. Miguel, 111 F.3d 666, 673 (9th Cir. 1997) (noting that,

where a Geders violation occurs, the violation constitutes “a denial of the assistance of counsel ‘altogether’

22

so as to require reversal with no showing of prejudice”); Mudd, 798 F.2d at 1513 (adopting a per se rule

that “reversal is required” following a Geders violation

as the approach that “best vindicates the right to the

effective assistance of counsel”); id. at 1515 (Scalia, J.,

concurring) (agreeing with majority on this point);

Clark v. State, 301 A.3d 241, 257, 271 (2023) (explaining that Geders violations require automatic reversal

even when raised in post-conviction proceedings and

reviewed for plain error); Martin, 991 A.2d at 793 (citing Geders, 425 U.S. at 91, and Perry, 488 U.S. at 27880) (“[A]n order prohibiting a defendant from conferring with his counsel during an overnight (or other significant) interruption of his testimony is a denial of the

defendant's Sixth Amendment right to counsel that requires reversal without any showing of prejudice.”);

Jackson v. United States, 422 A.2d 1202, 1203 (D.C.

1979) (en banc) (holding that an order directing a defendant not to discuss his testimony during a

lunchtime recess was reversible error without regard

to prejudice because “the degree of prejudice suffered

by the accused, and the impact on jury deliberations

often cannot be assessed on the record”). Cf. United

States v. Russell, 205 F.3d 768, 769-70 (5th Cir. 2000)

(absence of counsel for two days was structural error);

French v. Jones, 332 F.3d 430, 438 (6th Cir. 2003) (presuming prejudice from denial of counsel during response to jury note, given the “uncertainty of the prejudice” from the absence); Commonwealth v. Johnson,

828 A.2d 1009, 1015 (Pa. 2003) (presuming prejudice

from “temporary” denial of counsel “during reiterative

jury instructions”).

Even those courts declining to treat an alleged denial of counsel at a critical stage as structural have

typically done so only after they decline to find error to

begin with, because either the stage is not “critical” or

23

a break was too short to be deemed analogous to

Geders. See, e.g., United States v. Nelson, 884 F.3d

1103, 1107, 1108 (11th Cir. 2018) (noting that an order

prohibiting discussion of testimony may be structural

error, and that the Eleventh Circuit had earlier come

“pretty close” to explicitly saying so, but that the order

here was at the defense’s own request);6 United States

v Roy, 855 F.3d 1133 1147-48, 1150 (11th Cir. 2017)

(declining to apply structural error rule because the

seven-minute break involved a “momentary absence”

of counsel trivial enough not to be a “critical stage”);

Rose v. State, 304 P.3d 387, 394 (Mont. 2013) (declining to find Geders error where order required counsel

to leave jail by 10:30 and counsel could have come at

5:30 but chose not to come until 9:30); Wallace v. State,

851 So.2d 216, 220 (2003) (declining to find limit on

consultation during lunch break to be error, but acknowledging that it would otherwise presume prejudice from the error); Moore v. Commonwealth, 771

S.W.2d 34, 40-41 (Ky. 1988) (declining to presume

prejudice or even find error where order related to a

”brief” lunch recess during defendant’s testimony), ab-

6

The Nelson Court noted that the United States appeared to concede that a

prohibition on discussing testimony during an overnight break is structural

error if objected to:

Does it violate the Sixth Amendment to prevent a criminal defendant from discussing his testimony, but not other topics, during

a single overnight recess? Although no existing precedent resolves that precise question, even the Government seems to concede that the answer, at least as a general matter, is probably yes.

See Br. of Appellee at 52 (“[T]he district court’s limitation here

impermissibly constrained Skillern’s ability to consult with his attorney during the first overnight recess.”).

884 F.3d at 1106.

24

rogated on other grounds by McGuire v. Commonwealth, 885 S.W.2d 931 (Ky. 1994).7 As the California

Supreme Court explained, a defendant’s inability to

consult his lawyer about his testimony is different in

kind even from other Geders-type errors:

The harm from preventing an accused from

speaking with his or her attorney about his or

her own testimony extends further than that

attending court-imposed limitations on communications about a nondefendant witness, and is

also more difficult to quantify. . . [T]o remove

from the accused “the ability to discuss with his

attorney any aspect of his ongoing testimony effectively eviscerated his ability to discuss and

plan trial strategy. To hold otherwise would

defy reason. How can competent counsel not

take into consideration the testimony of his client in deciding how to try the rest of the case?”

People v. Hernandez, 273 P.3d 1113, 1121-22 (Cal.

2012) (declining to presume prejudice where limit on

discussions related only to a cooperator’s declaration).

Cf. State v. Smith, 375 So.3d 654, 668-69 (La. App.

2023) (declining to find error or presume prejudice

from limits on consultation with counsel during short

recess, but noting this “is different from” “instances

where the defendants were not allowed to consult with

their attorneys during their testimonies”).

7

One exception appears to be Sanders v. Lane, 861 F.2d 1033 (7th Cir.

1988), in which the court deemed the absence of counsel in a “brief routine

recess during the trial day” a Sixth Amendment violation but nonetheless

harmless under Chapman. Id. at 1035, 1040. The court noted that nearly all

other courts took a per se approach, id. at 1039 (citing cases), but cited this

Court's then-recent opinion in Satterwhite (subjecting Massiah error

to Chapman) in concluding that the deprivation during a mere "lunch recess" was analogous to the easily cabined error in Satterwhite. Sanders, 861

F.2d at 1040.

25

Treating Geders violations as structural errors also

avoids awkwardly requiring judicial probing of otherwise privileged attorney-client communications to determine the effect of the violation. Cf. Mudd, 798 F.2d

at 1513 (noting that a harm analysis in a Geders situation “would create an unacceptable risk of infringing

on the attorney-client privilege”). A reviewing court

would presumably have to ask of the defendant, for example, “what he and counsel discussed, what they

were prevented from discussing, and how the order altered the preparation of his defense.” See id. “Presumably the government would then be free to question defendant and counsel about the discussion that did take

place, to see if defendant nevertheless received adequate assistance.” Id.

Moreover, determining prejudice when it comes to

prohibitions on discussing a defendant’s testimony

during an overnight recess is uniquely difficult, given

the centrality of that evidence and everything it affects. In any criminal trial where a defendant testifies,

the defense risks inadvertently shifting the burden in

the minds of the jurors. Instead of asking whether the

prosecution has proved its case beyond a reasonable

doubt, the jurors might simply compare the competing

narratives of prosecution and defense, determining

which is most persuasive. As a result, the defendant’s

choice to testify leads to a host of strategic dilemmas

that must be resolved on the fly, related to whether

additional corroborating witnesses are worth the attendant risks, whether to pursue further investigation, whether to plead guilty, whether to seek a mistrial, which jury instructions to request or modify, and

what to say in closing arguments.

Given this Court’s precedents treating Geders errors

as structural, the growing consensus of lower courts on

the matter, and the ill-advised nature of a prejudice

26

inquiry in the context of attorney-client discussions of

a criminal defendant’s testimony, this Court should reaffirm that Geders violations are structural errors. To

change course now would impose an impossible burden

on reviewing courts and denigrate the one procedural

right – the right to assistance of counsel – through

which “all other rights of the accused are protected.”

Gideon v. Wainwright, 372 U.S. 335, 344 (1963).

CONCLUSION

For the foregoing reasons, the judgment of the court

of criminal appeals should be reversed.

Respectfully submitted,

BARBARA BERGMAN

Co-Chair,

AMICUS COMMITTEE

NATIONAL ASSOCIATION OF

CRIMINAL DEFENSE

LAWYERS

1201 E. Speedway Blvd.

Tucson, AZ 85721

(520) 621-3984

bbergman@arizona.edu

ANDREA ROTH

Counsel of Record

UC BERKELEY

SCHOOL OF LAW

347 Law Building

Berkeley, CA 94720

(510) 643-6092

aroth@law.berkeley.edu

Counsel for Amicus Curiae

June 10, 2025