United States v. Hohn

123 F.4th 1084

Opinion

FILED

United States Court of Appeals PUBLISH Tenth Circuit

UNITED STATES COURT OF APPEALS December 16, 2024

Christopher M. Wolpert

FOR THE TENTH CIRCUIT Clerk of Court

________________________________

UNITED STATES OF AMERICA, Plaintiff - Appellee, v. No. 22-3009 STEVEN M. HOHN, Defendant - Appellant.

---------------------------------

FEDERAL PUBLIC DEFENDERS FOR THE DISTRICTS OF COLORADO & WYOMING, UTAH, NEW MEXICO, AND OKLAHOMA (NORTHERN, WESTERN AND EASTERN); NATIONAL ASSOCIATION OF CRIMINAL DEFENSE LAWYERS,

Amici Curiae.

_________________________________

Appeal from the United States District Court for the District of Kansas (D.C. Nos. 2:19-CV-02082-JAR-JPO & 2:12-CR-20003-JAR-3 & 2:19-CV-02491-JAR-JPO)

_________________________________

Kannon K. Shanmugam, of Paul, Weiss, Rifkind, Wharton & Garrison LLP, Washington, DC (Abigail Frisch Vice, of Paul, Weiss, Rifkind, Wharton & Garrison LLP, Washington, DC, and Melody Brannon, Daniel T. Hansmeier, Paige A. Nichols, and Lydia Krebs Albert, Kansas Federal Public Defenders Office, Kansas City, Kansas, with him on the brief), for Defendant - Appellant.

James A. Brown, Assistant United States Attorney (Kate E. Brubacher, United States Attorney, with him on the brief), Topeka, Kansas, for Plaintiff - Appellee.

Virginia L. Grady, Federal Public Defender, Districts of Colorado and Wyoming, Denver, Colorado, Julia O’Connell, Federal Public Defender, Northern District of Oklahoma, Tulsa, Oklahoma, Margaret Katze, Federal Public Defender, District of New Mexico, Albuquerque, New Mexico, Scott Graham, Federal Public Defender, Eastern District of Oklahoma, Muskogee, Oklahoma, Jeffrey M. Byers, Federal Public Defender, Western District of Oklahoma, Oklahoma City, Oklahoma, Scott Wilson, Federal Public Defender, District of Utah, Salt Lake City, Utah, filed an amicus curiae brief in support of Defendant - Appellant.

Randall L. Hodgkinson and Norman Mueller of the National Association of Criminal Defense Lawyers, Washington, DC, with Jon M. Sands, Federal Public Defender and Daniel L. Kaplan, Assistant Federal Public Defender, District of Arizona, Phoenix, Arizona, filed an amicus curiae brief on behalf of the National Association of Criminal Defense Lawyers in support of Defendant - Appellant.

_________________________________

Before HOLMES, Chief Judge, HARTZ, TYMKOVICH, MATHESON, BACHARACH, PHILLIPS, McHUGH, EID, CARSON, and ROSSMAN, Circuit Judges. *

_________________________________

PHILLIPS, Circuit Judge, joined by HOLMES, Chief Judge, HARTZ, TYMKOVICH, MATHESON, McHUGH (except as to Part II(C)(2)), EID, and CARSON, Circuit Judges.*

_________________________________

Steven M. Hohn appeals the denial of a 28 U.S.C. § 2255 petition to vacate his judgment and discharge his case with prejudice or, alternatively, to reduce his sentence by half. CCA Recordings 2255 Litigation v. United States,

____________________

* The Honorable Nancy L. Moritz and the Honorable Richard E. N. Federico are recused in this matter.

3

Nos. 19-CV-2491, 12-CR-20003-03, 19-CV-2082, 2021 WL 5833911, at *1, *25 (D. Kan. Dec. 9, 2021), see United States v. Carter, 429 F. Supp. 3d 788, 798 (D. Kan. 2019). This claim arises out of the Kansas U.S. Attorney’s Office’s (Kansas USAO) since-discovered practice of obtaining and listening to attorney-client phone calls from detainees at CoreCivic. 1 Id. We have dealt with batches of similar appeals from CoreCivic detainees that emanate from the Kansas USAO’s mishandling of attorney-client communications. See generally United States v. Spaeth, 69 F.4th 1190 (10th Cir. 2023) (defendants with cases resolved by guilty pleas); United States v. Orduno-Ramirez, 61 F.4th 1263 (10th Cir. 2023) (defendants challenging post-plea, presentence intrusions). So Hohn’s appeal does not come to us in a vacuum.

Yet Hohn’s case is different because it implicates one of this court’s precedents, Shillinger v. Haworth, particularly Shillinger’s structural-error rule that presumes prejudice to a defendant when the government intentionally intrudes into the attorney-client relationship without a legitimate law- enforcement purpose. 70 F.3d 1132, 1142 (10th Cir. 1995). Under Shillinger, Hohn argues he had no burden to show that the government’s intrusion into his attorney-client relationship prejudiced him at trial. By removing prejudice from the Sixth Amendment equation, Hohn contends, Shillinger compels us to grant

1 The opinion in United States v. Carter establishes the district court’s factual findings on the Kansas USAO’s practice of intercepting and listening to attorney-client communications between CoreCivic detainees and their counsel. See generally 429 F. Supp. 3d 788.

4

him a remedy for the government’s intentional, unjustified intrusion into his confidential communication with his attorney.

Without Shillinger, Hohn’s argument collapses. Hohn concedes that he suffered no prejudice by the prosecution’s obtaining and listening to his six- minute call with his attorney—the communication at the heart of this case—and so he relies solely on Shillinger’s structural-error rule to sustain a Sixth Amendment violation. But Shillinger is a twenty-nine-year-old case, and we conclude that Shillinger is out of step with the Supreme Court’s cases on structural error and the “very limited class of cases” to which structural error extends. Greer v. United States, 593 U.S. 503, 513 (2021) (citation omitted). Recognizing that Shillinger is tenuous and yet critical to resolving Hohn’s appeal, this court voted to hear Hohn’s case en banc to decide whether we should retain Shillinger’s structural-error rule or reverse it.

After reconsidering Shillinger en banc, we conclude that the case—and its structural-error rule—is untenable under Supreme Court law. So for the reasons below, we now overrule Shillinger and hold instead that a Sixth Amendment violation of the right to confidential communication with an attorney requires the defendant to show prejudice. Here, Hohn concedes that he suffered no prejudice, so his claim automatically fails. On that ground, we exercise our jurisdiction under 28 U.S.C. §§ 1291 and 2253 to affirm the district court’s decision denying Hohn’s § 2255 petition.

5

BACKGROUND

I. Hohn’s Prosecution In January 2012, Hohn and several codefendants were indicted on one count of conspiring to possess with the intent to distribute and to distribute fifty grams or more of methamphetamine, in violation of 21 U.S.C. §§ 846, 841(a)(1), (b)(1)(A) (Count 1), as well as two counts of possession of a firearm by an unlawful user of a controlled substance, in violation of 18 U.S.C. §§ 922(g)(3), 924(a)(2) (Counts 13 and 14). 2 Months later, Hohn was charged in a multi-defendant second superseding indictment that raised additional gun- and-drug related charges. 3 The lead prosecutor in Hohn’s case was Kansas Assistant United States Attorney (AUSA) Terra Morehead. The primary investigators were Deputy Perry Williams of the Johnson County Sheriff’s Office, and Drug Enforcement Administration (DEA) Task Force Officer Christopher Farkes. The prosecution team offered plea deals to Hohn’s codefendants, many of whom accepted and

In June 2022 Congress amended 18 U.S.C. § 924(a) by striking § 922(g)

2

violations from paragraph (2) and moving them to a new paragraph (8). Bipartisan Safer Communities Act, Pub. L. No. 117-159, 136 Stat. 1313 (2022). So though § 922(g) violations are now found under § 924(a)(8), we reference the statutes as listed in the indictment.

The second superseding indictment charged two additional counts: one 3

count of possession of a firearm by an unlawful user of a controlled substance, in violation of 18 U.S.C. § 922(g)(3) (Count 15); and one count of possession of an unregistered short-barreled shotgun, in violation of 26 U.S.C. § 5861(d) (Count 16).

6

agreed to testify against Hohn. 4 At trial, these cooperating codefendants and several law-enforcement agents testified to Hohn’s participation in the alleged conspiracy; to his possession, use, and distribution of methamphetamine; and to his possession of the charged firearms. The government also introduced as evidence the illegally possessed firearms seized during a lawful search of Hohn’s truck.

After a twelve-day trial, the jury found Hohn guilty of all counts charged in the second superseding indictment, and the district court later sentenced Hohn to 360 months’ imprisonment to be followed by five years of supervised release. 5 Hohn appealed his judgment and sentence directly to this court, and we affirmed. United States v. Hohn, 606 F. App’x 902, 911 (10th Cir. 2015) (unpublished). II. Hohn’s Attorney-Client Call While he awaited trial, Hohn was detained at CoreCivic. He remained there from January 27, 2012, to March 28, 2014. During this time—the district

Hohn was tried alongside a codefendant, Michael Redifer. Redifer 4

separately appealed his conviction to this court, which we affirmed. United States v. Redifer, 631 F. App’x 548, 552 (10th Cir. 2015) (unpublished).

Hohn’s case was originally assigned to Judge Carlos Murguia, who 5

presided over all pre- and post-trial proceedings; the case was later reassigned to Chief Judge Julie Robinson on February 21, 2020, who issued the order denying Hohn’s § 2255 petition, which we review in this appeal. Chief Judge Robinson authored the opinion in United States v. Carter and is presiding over the consolidated CoreCivic § 2255 litigation, see infra n.10.

7

court would later come to discover—the Kansas USAO had been obtaining and listening to recorded attorney-client jail calls between CoreCivic detainees and their attorneys for “a wide variety of criminal cases,” Carter, 429 F. Supp. 3d at 847, including a large-scale drug investigation (the Black Investigation), id. at 798, 801, 848. 6 When the government charged the indictment in the Black Investigation, its practice of obtaining recorded attorney-client calls and meetings at CoreCivic came to light. Concerned about the constitutional ramifications of this practice, the district court ordered a Special Master to investigate the extent of the government’s intrusions into CoreCivic detainees’ attorney-client communications. As part of that investigation, in January 2019, the government produced some of the improperly obtained recordings to the Federal Public Defender (FPD), including the April 23, 2012 call that Hohn placed from CoreCivic to his newly appointed attorney. 7

6 The Black Investigation began in 2016 under the government’s suspicion that CoreCivic employees and detainees were working together to smuggle drugs into the facility. Carter, 429 F. Supp. 3d at 798, 801. The investigation was initially named after the lead defendant, but after he pleaded guilty the case was renamed for another defendant and became United States v. Carter. See id. at 801 n.10.

7 James Campbell was appointed as Hohn’s counsel on April 23, 2012;

Hohn’s previous appointed counsel, Assistant Federal Public Defender Tim Burdick, withdrew as counsel the next day.

8

Hohn’s call with his new attorney lasted six minutes. 8 During this introductory call, they discussed legal advice and trial strategy, “including: Hohn’s desire to have a trial in the matter, his criminal history, what he believed the evidence against him to be and problems with that evidence, concern about his truck being impounded, and the general way that they would proceed to meet and discuss the case going forward.” CCA Recordings 2255 Litigation, 2021 WL 5833911, at *12.

The government subpoenaed three batches of Hohn’s calls during his detention at CoreCivic. The April 23, 2012 call was extracted from the second batch, which the government had subpoenaed in connection with an investigation into the death of Gregory Price. The government issued that subpoena after one of Hohn’s codefendants informed Deputy Williams that Hohn had stuffed Price’s body into a refrigerator and transported the body to a property in De Soto, Kansas, where it was buried and later recovered by investigators. The government subpoenaed Hohn’s calls from CoreCivic placed between April 19, 2012, and April 23, 2012. Of the four total calls Hohn made during that time, one was recorded—the one to Hohn’s newly appointed attorney. The other three calls (to the FPD’s Office) were not recorded because

8 These facts draw from the district court’s findings after reviewing the audio recording of the phone call in camera. Hohn has not included the six- minute call in the record, and the district court denied the government access to the recorded call, so the government could not include it in the record. We operate solely from the findings of the district court as to its contents.

9

the toll-free numbers to that office had been privatized, according to CoreCivic’s procedures.

CoreCivic maintained procedures that allowed detainees to privatize their attorney-client calls from the prison. 9 Hohn signed a CoreCivic handbook, issued to him upon his arrival, which detailed the process for requesting to remove attorney calls from the prison’s recording system. The handbook advised Hohn that if he failed to abide this process, then his calls would be monitored and recorded for security purposes. Hohn admitted that he knew how to privatize attorney-client calls, yet he did not follow that protocol for the call he placed to his new attorney on April 23, 2012. In addition to the handbook, Hohn signed a “Monitoring of Inmate/Detainee Telephone Calls form,” which alerted him that CoreCivic retained the right to monitor his phone calls from the facility, that use of CoreCivic phones constituted consent to such monitoring, and that certain steps must be taken to exclude phone calls from CoreCivic’s recording system, including and especially calls to attorneys. Id. at *9. The area next to the phones at CoreCivic displayed signs that read, “ALL CALLS MAY BE RECORDED/MONITORED,” and/or “CALLS ARE SUBJECT TO MONITORING AND RECORDING.” Id. From all this, the

9 CoreCivic used a third-party company, Securus Technologies, to record all outgoing telephone calls from the prison. Any outgoing phone call to a successfully privatized number should have been excluded from Securus’s automatic recording system. But the record shows that, even when attorneys had correctly followed the procedures to privatize their numbers, Securus sometimes inexplicably recorded their client calls anyway.

10

district court made a finding that Hohn understood his attorney calls would be recorded, but that Hohn did not understand those recordings could be procured by the prosecution. Sure enough, AUSA Morehead later obtained Hohn’s calls, and the district court found that she had “possessed” and “listened to” Hohn’s six-minute attorney call from April 23, 2012, despite the AUSA’s sworn denials that she had never heard them. Id. at *22–23. III. Hohn’s Postconviction Proceedings In early 2019, upon learning that the government had obtained a confidential call with his attorney, Hohn sought habeas relief under § 2255. 10 In his § 2255 petition, Hohn argued that the government’s interception of the six- minute attorney-client call violated his Sixth Amendment right to communicate in confidence with his attorney and therefore warranted either a vacation of his judgment with prejudice or a fifty-percent reduction of his sentence.

The district court granted Hohn an evidentiary hearing on his § 2255 petition because the record did not conclusively show that Hohn was not

10 Hohn’s petition was part of an expansive effort by the FPD’s office to seek habeas relief under § 2255 for over 100 petitioners affected by the Kansas USAO’s “routine and systematic collection of all recorded phone calls from [CoreCivic] with no exception for attorney-client calls or any other precautionary measures.” Carter, 429 F. Supp. 3d at 900. The district court consolidated these § 2255 petitions under Federal Rule of Civil Procedure 42(a) to establish consistent “legal standards and threshold procedures” and to ensure “consistent relief” for all petitioners. Id. at 904; see id. at 902. But the result that the court reached in CCA Recordings 2255 Litigation was “limited to [the] facts before it with respect to Hohn.” 2021 WL 5833911, at *19.

11

entitled to relief. Before the hearing, Hohn stipulated that the six-minute attorney-client call was not introduced at trial, did not affect his trial, and did not affect his sentencing. After holding an evidentiary hearing, the district court denied Hohn’s motion to supplement his § 2255 petition. The court then issued its order resolving a handful of motions from the government and, most relevant here, denying Hohn’s § 2255 petition and his request for a certificate of appealability (COA).

The district court denied Hohn’s § 2255 petition largely based on its interpretation of the attorney-client privilege. See CCA Recordings 2255 Litigation, 2021 WL 5833911, at *15–18. For two reasons, the court did not reach a direct Sixth Amendment analysis, concluding that the Amendment’s protections had never attached to Hohn’s April 23, 2012 call: (1) because the call was not covered by the attorney-client privilege and (2) because, in the alternative, Hohn had waived the attorney-client privilege by placing the call despite knowing that the call would be recorded because he hadn’t followed the proper steps to privatize it. See id. at *17 (finding Hohn’s conduct “inconsistent with an objectively reasonable expectation of confidentiality in the attorney- client communications” and therefore outside the scope of the privilege). Above all, the court considered the attorney-client privilege to be “a necessary

12

underpinning of Hohn’s Sixth Amendment right,” and so, absent the privilege, his Sixth Amendment claim failed. 11 Id. at *18.

After the district court’s denial, Hohn timely appealed and applied to this court for a COA, which we granted on two questions:

(1) Whether the district court erred in ruling that Mr. Hohn failed to prove the elements of his Sixth Amendment claim.

(2) Whether the district court erred in ruling that the government proved Mr. Hohn waived his Sixth Amendment right.

Order Granting Certificate of Appealability, United States v. Hohn, No. 22- 3009 (10th Cir. Oct. 25, 2022), ECF No. 54.

A panel of this court heard oral argument on these questions in September 2023. 12 After argument, the panel called sua sponte for an en banc poll, 13 asking the full court to reconsider our holding in Shillinger based on

11 The court then discussed the Sixth Amendment’s parameters for two elements of Hohn’s claimed violation under Shillinger: (1) the purposeful intrusion into the attorney-client relationship and (2) the legitimate law- enforcement purpose for the government’s intrusion. CCA Recordings 2255 Litigation, 2021 WL 5833911, at *19–24. Obviously, we do not fault the district court for expounding on Tenth Circuit law as it existed at the time, but because this opinion supplants Shillinger, we do not recount this portion of the district court’s order.

Hohn’s pending motion to file an oversized reply brief not to exceed 12

7,500 words is granted.

13 In the panel briefing, the government did argue that the district court erred by relying on “Shillinger’s presumption of prejudice, because applying that presumption runs contrary to the rule and rationale of Weatherford [,] . . . and . . . Morrison . . . , and other Supreme Court cases.” Appellee’s Br. at 24. Obviously, the Shillinger decision bound the panel.

13

Supreme Court precedent that the panel asserted casts Shillinger’s structural- error rule in doubt. The poll carried, and this court agreed to hear Hohn’s case en banc. With our en banc order, we directed the parties to file supplemental briefs on two questions:

(1) Did Shillinger v. Haworth, 70 F.3d 1132 (10th Cir. 1995)

correctly hold that it is structural error for the government to purposefully intrude without legitimate justification into the attorney-client relationship and that prejudice must be presumed?

(2) When, if ever, does the government unjustifiably intrude into the attorney-client relationship by intentionally obtaining attorney-client communications that are not privileged?

United States v. Hohn, 91 F.4th 1060 (10th Cir. 2024) (mem.).

An en banc court heard oral arguments in May 2024. Having received the parties’ supplemental briefs and heard their arguments, we issue this decision.

LEGAL BACKGROUND

The Sixth Amendment guarantees a right to the effective assistance of counsel. Strickland v. Washington, 466 U.S. 668, 686 (1984); see McMann v. Richardson, 397 U.S. 759, 771 n.14 (1970) (“[T]he right to counsel is the right to the effective assistance of counsel.”). The Constitution recognizes this right because the guidance of an attorney helps ensure that the defendant receives a fair trial. Orduno-Ramirez, 61 F.4th at 1267; see Strickland, 466 U.S. at 685 (explaining that the Sixth Amendment confers a right to effective counsel because of the “critical” role attorneys play in “the ability of the adversarial system to produce just results”). Part and parcel of the right to effective

14

assistance is the right to communicate confidentially with an attorney. See Weatherford v. Bursey, 429 U.S. 545, 554 n.4 (1977). This is so because the government’s intrusion into the attorney-client relationship “inhibit[s] [the] free exchanges between defendant and counsel” and therefore constrains an attorney’s ability to effectively represent a defendant. Id.; see Mohawk Indus., Inc. v. Carpenter, 558 U.S. 100, 108 (2009) (“By assuring confidentiality, the [attorney-client] privilege encourages clients to make ‘full and frank’ disclosures to their attorneys, who are then better able to provide candid advice and effective representation.”).

Yet the Supreme Court has never held that the Sixth Amendment right to attorney-client confidentiality “subsumes a right to be free from intrusion” by government agents into the attorney-client relationship. Weatherford, 429 U.S. at 553 (discussing Hoffa v. United States, 385 U.S. 293 (1966)); see also Maine v. Moulton, 474 U.S. 159, 176 (1985) (“Thus, the Sixth Amendment is not violated whenever—by luck or happenstance—the State obtains incriminating statements from the accused after the right to counsel has attached.”). Rather, to establish a Sixth Amendment violation, the defendant must show (1) that the government intentionally intruded into the defense camp and (2) that the intrusion caused prejudice. See United States v. Morrison, 449 U.S. 361, 365 (1981) (stating that a successful Sixth Amendment claim must identify a “constitutional infringement” which “has produced some . . . prejudice to the defense”); Sanborn v. Parker, 629 F.3d 554, 571–72 (6th Cir. 2010) (noting

15

that “purposeful intrusion alone is not a Sixth Amendment violation” because, “[e]ven if [the government’s intrusion] [we]re . . . purposely designed to elicit information” from the defense, the defendant must also show “actual[] prejudice[]”). So while evidence of an intentional intrusion into the attorney- client relationship creates a “colorable” Sixth Amendment claim, United States v. Chandler, 56 F.4th 27, 38 (2d Cir. 2022), the violation is not complete until the defendant establishes prejudice, see Mickens v. Taylor, 535 U.S. 162, 166 (2002) (clarifying that “defects in assistance that have no probable effect upon the trial’s outcome do not establish a constitutional violation” under the Sixth Amendment). Prejudice in this context means a realistic possibility of injury to the defendant or benefit to the government. Weatherford, 429 U.S. at 558.

DISCUSSION

Hohn argues that the district court erred in denying his § 2255 petition because his petition advances a per se Sixth Amendment violation that entitles him to relief. Hohn asserts that because “[t]he Sixth Amendment’s constitutional protection is not limited by the scope of the attorney-client privilege,” the district court mistakenly required Hohn to have premised his Sixth Amendment claim on a privileged attorney-client communication. Suppl. Br. at 24. Hohn also insists that he was not required to show prejudice from AUSA Morehead’s intrusion because Shillinger’s structural-error rule presumes prejudice in his situation: where the government intruded into an attorney- client conversation purposefully and absent any law-enforcement interest. In

16

considering these claims, we accept the district court’s factual findings unless they are clearly erroneous and review its legal conclusions de novo. United States v. Orange, 447 F.3d 792, 796 (10th Cir. 2006).

We begin with the district court’s interpretation of Shillinger—

specifically, whether the district court misread Shillinger as requiring that attorney-client communications protected by the Sixth Amendment also be covered by the attorney-client privilege. We then consider the prejudice component of Hohn’s Sixth Amendment claim. In doing so, we review, reverse, and replace Shillinger’s structural-error rule. I. Attorney-Client Privilege The district court began its analysis by stating the four elements of a per se Sixth Amendment violation, as set by Shillinger: (1) a protected attorney- client communication; (2) purposeful intrusion into the attorney-client relationship; (3) the prosecutor’s becoming privy to the attorney-client communication due to the intrusion; and (4) no legitimate law-enforcement justification for the intrusion. CCA Recordings 2255 Litigation, 2021 WL 5833911, at *14. The first element—a protected attorney-client communication—led the district court to consider the attorney-client privilege. It noted that, though “the attorney-client privilege is not a right guaranteed by the Sixth Amendment,” id. at *16, the privilege nevertheless “relate[s] to” the constitutional right “to speak candidly and confidentially with counsel,” id. at *15. So in the court’s view, the privilege provides “an appropriate framework

17

for showing that the recordings between [Hohn] and counsel [were] protected communications under the Sixth Amendment.” Id. at *16.

The district court concluded that “to establish the protected-

communication element” of a Sixth Amendment violation, Hohn had to “show that he had a reasonable expectation of confidentiality in his attorney-client call and that he did not otherwise waive the attorney-client privilege.” Id. As to the former, the district court found that Hohn’s calling his attorney on a phone that Hohn knew was monitored and recorded by CoreCivic was “inconsistent with an objectively reasonable expectation of confidentiality.” Id. at *17. As to the latter, the district court found that Hohn waived the attorney-client privilege by calling his attorney from CoreCivic despite his knowledge and understanding that the communication would be exposed to third parties. Taking these findings together, the district court reasoned that Sixth Amendment protections never attached to Hohn’s attorney-client call because the call was nonprivileged. The court thus declined to reach Sixth Amendment waiver.

By conditioning Hohn’s Sixth Amendment claim on a showing that the attorney-client privilege had attached, the district court equated confidential communications protected by Shillinger with those covered by the privilege. See id. at *18 (“[T]he attorney-client privilege [was] a necessary underpinning of Hohn’s Sixth Amendment right.”). This determination premised Hohn’s initial appeal and COA application to this court, in which he argued that the

18

district court had erroneously injected the evidentiary privilege into Shillinger’s elemental test.

In granting an en banc hearing, we asked the parties to consider “[w]hen, if ever, does the government unjustifiably intrude into the attorney-client relationship by intentionally obtaining attorney-client communications that are not privileged?” Hohn, 91 F.4th at 1060. Put differently, is the Sixth Amendment right to attorney-client confidentiality coextensive with the attorney-client privilege? Hohn insists that “the Sixth Amendment’s constitutional protection is not limited by the scope of the attorney-client privilege” and that a violation of the constitutional right does not depend on “whether such attorney-client communications are privileged.” Suppl. Br. at 23. Hohn argues that the prosecution violated the Sixth Amendment by “intentionally and unjustifiably bec[oming] privy to the contents of confidential (even if nonprivileged) attorney-client communications.” Id. at 25. In response, the government focuses on the defendant’s unreasonable expectation that his six-minute call from CoreCivic would remain confidential. The government also endorses the district court’s view that “nonprivileged communications are generally not protected by the Sixth Amendment.” Suppl. Resp. Br. at 20.

Hohn is correct to the extent he argues that the Sixth Amendment right to attorney-client confidentiality and the attorney-client privilege furnish separate protections over the attorney-client relationship—one flowing from the Constitution’s text and the other flowing from evidentiary principles that

19

predate the Constitution. See U.S. Const. amend. VI (“In all criminal prosecutions, the accused shall enjoy the right . . . to have the Assistance of Counsel for his defence.”); 1 McCormick on Evidence § 87 (8th ed. July 2022 update) (describing the attorney-client privilege’s beginnings in Roman law); cf. Howell v. Trammell, 728 F.3d 1202, 1222 (10th Cir. 2013) (“[S]tanding alone, the attorney-client privilege is merely a rule of evidence; it has not yet been held a constitutional right.” (citation omitted)). In this vein, the district court’s order correctly acknowledges that “the Sixth Amendment subsumes the attorney-client privilege,” but is “not limited to . . . [its] scope.” CCA Recordings 2255 Litigation, 2021 WL 5833911, at *15 (cleaned up).

We agree that Sixth Amendment attorney-client confidentiality is distinct from and broader than the attorney-client privilege. 14 See Weatherford, 429 U.S. at 554 (disagreeing that “the defendant assumes the risk” and relinquishes the right to “complain if [a] third party turns out to be an informer” any time the “defendant converses with his counsel in the presence of a third party thought to be a confederate and ally”). Because the Sixth Amendment is

14 The Supreme Court’s interpretation of the attorney-client privilege and the Sixth Amendment right to counsel support that the privilege has a shorter runway than its constitutional counterpart. Compare United States v. Zolin, 491 U.S. 554, 562 (1989) (caveating that because “the privilege has the effect of withholding relevant information from the factfinder, it applies only” so far as it is “necessary to achieve its purpose” of enabling open discussions between clients and counsel (quoting United States v. Fisher, 425 U.S. 391, 403 (1976)), with Montejo v. Louisiana, 556 U.S. 778, 786 (2009) (affirming that a defendant’s Sixth Amendment right to counsel remains intact unless “the right is voluntar[ily], knowing[ly], and intelligent[ly]” waived).

20

broader, it is at least conceivable that Hohn’s six-minute call from CoreCivic was constitutionally protected, even if nonprivileged. Under that heuristic, we assume without deciding that Sixth Amendment protections attached to Hohn’s attorney-client call from April 23, 2012. Meaning, we assume that Hohn has satisfied the first component of his Sixth Amendment claim: an intentional intrusion into the attorney-client relationship. II. Prejudice Even when the government intentionally intrudes into the defense camp, the Sixth Amendment is not violated unless the intrusion prejudiced the defendant during the criminal proceedings. See Weatherford, 429 U.S. at 553– 54; cf. United States v. Ginsberg, 758 F.2d 823, 833 (2d Cir. 1985) (“[T]he mere presence of a government agent, informant, or cooperating witness at conferences between defendant and counsel does not violate the sixth amendment.”). As such, “[a]bsent some effect of challenged conduct on the reliability of the trial process,” a defendant cannot usually prevail on a Sixth Amendment right-to-counsel claim. United States v. Cronic, 466 U.S. 648, 658 (1984).

But there are exceptions to this general rule. The prejudice component of a Sixth Amendment violation is presumed when “the cost of litigating [its] effect in a particular case is unjustified.” Id.; accord Strickland, 466 U.S. at 692. This type of violation amounts to structural error—an error so egregious it “def[ies] analysis” under our typical harmless-error rubric. United States v.

21

Gonzalez-Lopez, 548 U.S. 140, 148 (2006) (quoting Arizona v. Fulminante, 499 U.S. 279, 310 (1991)). Structural errors “affect[] the framework within which the trial proceeds,” meaning that the trial’s “reliab[ility] [in] serv[ing] its function as a vehicle for determination of guilt or innocence” has been irreparably compromised. Fulminante, 499 U.S. at 310; accord Greer, 593 U.S. at 513 (affirming that a structural error “affect[s] the entire . . . proceeding from beginning to end” as opposed to a “discrete defect[]” (cleaned up)). For this reason, criminal defendants subjected to structural error are entitled to a remedy even without having shown prejudice. See Fulminante, 499 U.S. at 294; Gonzalez-Lopez, 548 U.S. at 148.

In this appeal, we must decide whether intentional, unjustified intrusions into the attorney-client relationship belong among the “limited class of cases” that presumes prejudice under the Sixth Amendment or whether, like ineffective-assistance-of-counsel claims, the defendant must show prejudice. Johnson v. United States, 520 U.S. 461, 468–69 (1997); 15 cf. Spaeth, 69 F.4th at

15 Judge Rossman’s dissent asserts that the majority opinion errs by its “comprehensive reliance on Strickland’s prejudice prong—which applies to Sixth Amendment claims based on defense counsel’s performance. . . .” Rossman, J., dissenting, at 17. The dissent contends that “the Sixth Amendment violation at issue here is based on ‘direct governmental interference with the right to counsel,’ which the Supreme Court has ‘expressly noted . . . is a different matter.’” Id. (quoting Perry v. Leeke, 488 U.S. 272, 279 (1989)). It further maintains that “the majority misunderstands the nature of the Sixth Amendment right at issue,” as seen by its requiring “the defendant to show prejudice here, because Strickland did. . . .” Id. This misses the mark. In fact, it is the dissent that errs by its repeated use of the overbroad term “governmental (footnote continued)

22

1211–12 (elaborating on the difference between Strickland ineffective- assistance claims and “other” Sixth Amendment claims, which “d[o] not require any showing” by the defendant). 16 The Supreme Court generally classifies an error as structural (1) “if the right at issue is not designed to protect the defendant from erroneous conviction but instead protects some other interest”; (2) “if the effects of the error are simply too hard to measure”; and (3) “if the error always results in fundamental unfairness.” Weaver v. Massachusetts, 582 U.S. 286, 295–96 (2017). Within these categories, the Supreme Court has determined the following to constitute structural errors: the admission of a defendant’s guilt over his objection, McCoy v. Louisiana, 584 U.S. 414, 427–28 (2018); the deprivation of a

interference,” by which the dissent merges prosecutorial interference with judicial interference. See, e.g., id. at 17, 26, 30. The majority opinion acknowledges the legion of Supreme Court cases finding structural error from judicial interference with counsel’s ability to provide effective assistance. See, e.g., Majority Op. at 22–24, 44–47, 53. But if that line of cases requires structural error for prosecutorial intrusions, Weatherford and Morrison failed to notice so and blundered by repeatedly discussing the need for prejudice.

16 Spaeth differs from Hohn’s case in material ways. A jury convicted Hohn, but Spaeth entered an unconditional guilty plea. So Spaeth was left to pursue a § 2255 petition alleging ineffective assistance of counsel—deficient performance and prejudice—under Strickland. We rejected Hohn’s reliance on the line of cases that instead applied “lack-of-effective-assistance” principles— that is, mostly cases involving judicially placed impediments to counsel’s effectiveness. We rejected Hohn’s doing so, characterizing his effort as one trying to “shoehorn Shillinger” outside Strickland’s ineffective-assistance-of- counsel realm. Spaeth, 69 F.4th at 1211. So the majority opinion does not abandon the court’s “distinct understanding of these two kinds of Sixth Amendment violations,” as Judge Rossman charges, but hews to this distinction. See Rossman, J., dissenting, at 20 n.12.

23

defendant’s right to counsel of his choice, Gonzalez-Lopez, 548 U.S. at 150; the provision of an incorrect reasonable-doubt instruction, Sullivan v. Louisiana, 508 U.S. 275, 279–81 (1993); the exclusion of grand jurors of a defendant’s same race, Vasquez v. Hillery, 474 U.S. 254, 263–64 (1986); the denial of the right to a public trial, Waller v. Georgia, 467 U.S. 39, 47 (1984); the deprivation of a defendant’s right to proceed pro se, McKaskle v. Wiggins, 465 U.S. 168, 187–88 (1984); and the denial of an impartial judge, Tumey v. Ohio, 273 U.S. 510, 531, 535 (1927). Specific to the Sixth Amendment right to effective assistance of counsel, the Court recognizes structural error for the complete denial of counsel, Gideon v. Wainwright, 372 U.S. 335, 344 (1963), the constructive denial of counsel, Davis v. Alaska, 415 U.S. 308, 320 (1974) (denial of right to effective cross-examination), and the utter inability of any attorney to be effective under the circumstances, Powell v. Alabama, 287 U.S. 45, 59 (1932) (counsel appointed the morning of the trial). See Cronic, 466 U.S. at 659–60 (collecting cases).

Shillinger established a structural-error rule for certain kinds of Sixth Amendment intrusion claims. The Shillinger court held that when the government intentionally intrudes into the attorney-client relationship “absent a countervailing state interest,” prejudice becomes “so likely” that it must be presumed, and therefore the intrusion causes a structural error in the trial’s framework. 70 F.3d at 1142 (quoting Strickland, 466 U.S. at 692). But nearly twenty years earlier, in Weatherford, the Supreme Court condemned structural-

24

error rules for Sixth Amendment intrusion claims. See 429 U.S. at 557–58. So in fashioning its new structural-error rule, Shillinger distinguished Weatherford on its facts—which involved an undercover informant sitting in on a defendant’s pretrial meetings with his counsel—and determined that Weatherford was not binding. Shillinger, 70 F.3d at 1139–40. Shillinger erred by departing from Weatherford, which was and remains binding authority on this court. Shillinger’s holding contradicts those pronounced in Weatherford and its progeny because those cases affirm that, even when the prosecution becomes privy to attorney-client communications without a legitimate law- enforcement purpose, the defendant still must demonstrate a prejudicial use of the overheard information at trial. See 429 U.S. at 553–54. Under these authorities, we find Shillinger’s application of structural error unsound.

A. Weatherford and its Progeny Weatherford was a 42 U.S.C. § 1983 case. 429 U.S. at 547. The plaintiff, Bursey, sued defendant Weatherford (an undercover government informant) for constitutional violations that Weatherford allegedly committed after the two men vandalized a county building, a crime for which both were arrested and charged. Id. To maintain his cover, Weatherford accompanied Bursey—at Bursey’s invitation—to two pretrial-preparation meetings with Bursey’s defense counsel. Id. at 547–48. Weatherford did not elicit information from Bursey or his counsel during these meetings, he did not ask to attend these meetings, nor did he communicate anything that he learned during these

25

meetings to the prosecution team. Id. at 548. At Bursey’s trial, the prosecution called Weatherford as a witness, at which point Weatherford testified to his “undercover activities.” Id. at 549. Bursey was convicted. Id.

Later, Bursey brought a civil action under § 1983 against Weatherford, claiming in part that Weatherford’s presence during pretrial defense-strategy meetings infringed on Bursey’s Sixth Amendment right to effective assistance of counsel. Id. Reviewing a judgment for the defendants on appeal, the Fourth Circuit reversed in Bursey’s favor. Id. The Fourth Circuit held that a per se Sixth Amendment violation occurs “whenever the prosecution knowingly arranges and permits intrusion into the attorney-client relationship.” Id. (quoting Bursey v. Weatherford, 528 F.2d 483, 486 (4th Cir. 1975)). The Supreme Court granted certiorari and reversed. Id. at 550.

The Supreme Court disapproved of the Fourth Circuit’s strict approach, which assumed a constitutional violation regardless of “the purpose of the agent in attending the meeting,” regardless of “whether or not [the agent] reported on the meeting to [the prosecution],” and regardless of “whether or not any specific prejudice to the defendant’s preparation for or conduct of the trial [wa]s demonstrated or otherwise threatened.” Id. The Court concluded that the Fourth Circuit’s rule “cut[] much too broadly.” Id. at 557; see id. at 558 (imagining various scenarios that would violate the Fourth Circuit’s per se rule and yet cause no prejudice to Bursey, for example, if “the entire conversation had consisted of [defense counsel’s] questions and Weatherford’s answers

26

about Weatherford’s own defense plans”). And the Court rejected the Fourth Circuit’s conclusion that the holdings in Black v. United States, 385 U.S. 26 (1966), O’Brien v. United States, 386 U.S. 345 (1967), and Hoffa v. United States, 385 U.S. 293 (1966), supported the Fourth Circuit’s conclusion to adopt a per se rule. Id. at 550–51. The Supreme Court clarified that these precedents neither “require” nor “suggest” a per se Sixth Amendment violation for the government’s intentional intrusions into the defense camp. Id. at 551.

From Black and O’Brien—two cases involving the government’s illegal electronic surveillance of defendants’ conversations with counsel before trial— the Court extrapolated that the constitutionality of the government’s overhearing confidential attorney-client conversations “depends on whether the . . . conversations have produced, directly or indirectly, any of the evidence offered at trial.” Id. at 551–52. The Court noted that, even though the Black and O’Brien Courts both ordered new trials, neither Court did so because the government’s intrusion was per se prejudicial. See id. Rather, the Black Court determined that a new trial was appropriate based on the “particular facts” of the case, and the O’Brien Court merely cited the Black per curiam decision with no additional reasoning, giving Weatherford (and us) little to draw from. Id. Contrasting these two cases with Weatherford, the Supreme Court criticized the Fourth Circuit’s rule for leveraging its precedents to render “trial prejudice . . . irrelevant.” Id. at 552.

27

The Court next addressed its decision in Hoffa, which the Court also said did not countenance the Fourth Circuit’s strict rule. Weatherford rejected that Hoffa supported per se Sixth Amendment intrusion rules because the Hoffa Court had merely “assumed without deciding” that the prosecution’s becoming privy to attorney-client communications in a separate case would have violated the Sixth Amendment. Id. at 553. But even so, the Court determined such violation would not have impacted Hoffa’s rights in a different criminal trial for jury tampering—the conviction under review. Id. Because Hoffa had merely assumed without deciding a Sixth Amendment violation for the particular (and unique) factual circumstances of that case, the Weatherford Court was unconvinced that Hoffa justified the Fourth Circuit’s sweeping conclusion that a per se Sixth Amendment violation occurs whenever the government intentionally intrudes into the attorney-client relationship. See id. at 554.

Next, the Weatherford Court posited that a Sixth Amendment violation might have occurred if Weatherford had testified to the conversations that took place during the pretrial meetings, if the state’s evidence had originated from Weatherford’s inside information, if the state had “used in any other way” the substance of the conversations against Bursey, or if the state had “learned from Weatherford” the trial-preparation details of the attorney-client conversations. Id. But even then, the Court continued, Bursey would have had only “a much stronger case” in proving a Sixth Amendment violation, so obviously not a guaranteed one. Id. For even in the worst cases, where the informant

28

purposefully intrudes into confidential attorney-client conversations or where the informant relates those conversations to the prosecution, Weatherford still advises against assuming that the confidential information “has the potential for detriment to the defendant or benefit to the prosecutor’s case.” Id. at 557. So we should not presume that the information intercepted from an attorney-client conversation is prejudicial. See id. And if prejudice is not to be presumed, then it follows that it must be shown. Weatherford thus established a prejudice requirement for intrusion-based Sixth Amendment claims.

Other Supreme Court decisions reinforce Weatherford’s prejudice requirement. In Morrison—another case decided pre-Shillinger—the Supreme Court assumed without deciding that the two DEA agents visiting the defendant in jail and advising her to seek different defense counsel had intruded into her attorney-client relationship and so violated her Sixth Amendment rights. 449 U.S. at 364. Having assumed without deciding that a Sixth Amendment intrusion occurred, the Court proceeded to consider the Third Circuit’s remedy of dismissal with prejudice. See id. The remedy, the Court surmised, “should be tailored to the injury suffered from the constitutional violation.” Id.

But the defendant never alleged that the DEA agents’ interference “prejudiced the quality or effectiveness of [her] legal representation,” “induced her to plead guilty,” “resulted in the prosecution having a stronger case against her,” or “had any other adverse impact on her legal position.” Id. at 363. That is, the defendant made no showing or even an allegation of prejudice. Id. at

29

366. Without a showing or allegation of prejudice, the Court found there was “no effect of a constitutional dimension” that “need[ed] to be purged” and therefore “no justification for interfering with the criminal proceedings . . . much less the [dismissal of the indictment] granted by the [circuit court].” 17 Id. at 366–67. The Court considered that, once a “constitutional infringement [has been] identified,” there must be some “threat[]” of an “adverse effect upon the effectiveness of counsel’s representation” or “some other prejudice to the defense” to have a remediable Sixth Amendment claim. Id. at 365. An evidentiary hearing would give the defendant an opportunity to tease out the effects of such a “threat,” to prove that the government’s interference had indeed been prejudicial. Id. (explaining that a threat of prejudice may “impact . . . the criminal proceeding” and create a “basis for imposing a remedy”). Yet without threatened or demonstrable prejudice, the Court found no reason to impose any remedy, including an evidentiary hearing. Id. at 366. The Morrison Court was unequivocal that, even assuming the government has infringed a defendant’s Sixth Amendment right to counsel, no relief should inure absent some demonstration or threat of prejudice. See id. at 365. Even in Cronic, the Court declared that the “Sixth Amendment guarantee is generally not implicated” “[a]bsent some effect of the challenged conduct on the reliability

17 We understand Morrison’s reference to “interfering with the criminal proceedings” as including the court’s granting the defendant an evidentiary hearing to demonstrate prejudice. 449 U.S. at 366–67.

30

of the trial process.” 466 U.S. at 658; see Gonzalez-Lopez, 548 U.S. at 147 (“[A] violation of the Sixth Amendment right to effective representation is not ‘complete’ until the defendant is prejudiced.”). The Cronic Court noted that the only cases that do not require “some . . . effect” on the trial process are those where prejudice may be presumed because the “circumstances . . . are so likely to prejudice the accused that the cost of litigating their effect in a particular case is unjustified.” 466 U.S. at 658. Cronic divided these prejudice-per-se effective-assistance claims into three buckets: (1) when “the accused is denied counsel at a critical stage of his trial”; (2) when “counsel entirely fails to subject the prosecution’s case to meaningful adversarial testing”; and (3) when “the likelihood that any lawyer, even a fully competent one, could provide effective assistance is so small that a presumption of prejudice is appropriate.” Id. at 659–61. These limited categories signal the Supreme Court’s desire to reserve Sixth Amendment structural errors for “extreme situations.” Orduno- Ramirez, 61 F.4th at 1268.

Weatherford and the decisions that followed formed the backdrop for Shillinger’s structural-error rule. And since Shillinger, the Supreme Court has only entrenched its view that a “very limited class of cases” warrant structural error. Greer, 593 U.S. at 513 (citation omitted); see, e.g., id. (rejecting as structural an error where an element of the offense was omitted from a guilty plea colloquy); United States v. Davila, 569 U.S. 597, 610–11 (2013) (rejecting the defendant’s argument that judicial interference in plea discussions required

31

automatic vacatur because conduct violative of Rule 11 “does not belong in th[e] highly exceptional category” of structural error); Neder v. United States, 527 U.S. 1, 7–10 (1999) (rejecting that “a jury instruction that omits an element of the offense” amounts to structural error). Having surveyed this legal landscape, we return to Shillinger.

B. Shillinger

1. Shillinger held that the government’s intentional, unjustified intrusion into the attorney-client relationship is structural error.

In Shillinger, this court faced a habeas petition under § 2254 from Wyoming state prisoner Steven Haworth—a man convicted of aggravated assault and battery for wielding a pocketknife against another man outside a bar. 70 F.3d at 1134, 1136. Haworth sought habeas relief on the ground that the state prosecutor intruded into his attorney-client relationship and used their communications against Haworth at trial. Id. at 1136. The district court granted Haworth’s petition. Id. We affirmed the district court but remanded for an evidentiary hearing to determine the proper remedy. Id. at 1143.

Before his state trial, Haworth’s defense counsel arranged to bring Haworth into the courtroom on weekends to prepare for his testimony. Id. at 1134. Because Haworth was being detained pretrial, a deputy sheriff accompanied him and his defense counsel during these preparatory sessions. Id. Somewhat unconventionally, the defense counsel invited the deputy sheriff into the defense camp by paying him overtime wages and instructing him to

32

“consider himself an employee of defense counsel,” with the understanding that nothing the deputy heard should leave the courtroom. 18 Id. The deputy sheriff acceded to this odd arrangement. Id. But as it turned out, the prosecutor got wind of it and questioned the deputy about what he heard and observed during the defense’s preparatory sessions: defense counsel advising Haworth to use attenuating terms during his testimony to describe his criminal conduct (i.e., to say that Haworth “cut” rather than “stabbed” the victim) and advising Haworth about courtroom deportment, such as sitting up straight during testimony and looking at the jury. Id. at 1134–35, 1137.

During an in-chambers conference on Haworth’s motion to suppress evidence of these preparatory sessions, it became evident that the prosecution had become privy to the substance of the sessions. Id. at 1134–35. In another in-chambers conference, the defense counsel emphasized his concern that the prosecutor’s knowledge about the defense’s preparatory sessions violated the attorney-client privilege and therefore that the court should disallow the prosecution from using any of this information in cross-examining Haworth at trial. Id. at 1135. The court agreed with the defense counsel’s concerns, but nevertheless gave the prosecution some room to cross-examine Haworth using information that it had gathered from the preparatory sessions. Id. Haworth was

18 The state trial judge stated that, “had [defense counsel] called [him]” to ask about this unconventional arrangement, he would have ordered “the deputy to step outside.” Haworth v. State, 840 P.2d 912, 914–15 (Wyo. 1992).

33

convicted. Id. at 1136. On direct appeal, the Wyoming Supreme Court upheld Haworth’s conviction because the court determined that his right to effective assistance of counsel had not been violated by the prosecutor’s conduct. Haworth v. State, 840 P.2d 912, 918 (Wyo. 1992).

Haworth then proceeded to federal court, filing a petition for habeas relief under 28 U.S.C. § 2254. Haworth v. Shillinger, 852 F. Supp. 961, 962 (D. Wyo. 1994). The district court found a Sixth Amendment violation and granted Haworth’s petition, id. at 969–70; on appeal, this court affirmed, Shillinger, 70 F.3d at 1143.

But in affirming Haworth’s § 2254 petition, this court went beyond the district court’s ruling, which concluded that the prosecution’s conduct violated the Sixth Amendment because the prosecutor had intentionally acquired information about the defense’s strategy and then “used the information that it learned at trial” against Haworth. Haworth, 852 F. Supp. at 969 (emphasis added). 19 Though the Shillinger court agreed that “the prejudice standard

19 The district court opined that “[i]t is clear from Weatherford that an intrusion, standing alone, is insufficient as a matter of law to constitute a violation and that an ‘intrusion plus’ standard is appropriate.” Haworth, 852 F. Supp. at 966. As to what this “plus” might include, the district court referenced “the four factors set out in Weatherford” derived from “what the petitioner in Weatherford had failed to show.” Id. at 965–66 (emphasis omitted). These factors are:

“(1) whether evidence used at trial was produced directly or indirectly by the intrusion; (2) whether the intrusion by the government was intentional; (3) whether the prosecution received otherwise confidential information about trial preparations or (footnote continued)

34

articulated in Weatherford ha[d] been met” on the facts, and though the court could have stopped there, it chose to expand its ruling by additionally presuming prejudice based on the nature of the prosecutor’s intrusion. Shillinger, 70 F.3d at 1139; see id. at 1142. Two aspects of the prosecutor’s intrusion—(1) that it was intentional and (2) that it lacked any legitimate law- enforcement purpose—led this court to hold that prejudice to Haworth could be presumed. Id. at 1142. And because the intrusion in Weatherford was unintentional and justified, the Shillinger court concluded that Weatherford’s holding didn’t bind its decision. Id. at 1140. After distinguishing Weatherford, the Shillinger court—as we now see it—felt free to adopt the sort of structural- error rule that Weatherford said “cuts much too broadly.” Weatherford, 429 U.S. at 557. Cutting a similarly wide swath, Shillinger held:

[W]hen the state becomes privy to confidential communications because of its purposeful intrusion into the attorney-client relationship and lacks a legitimate justification for doing so, a prejudicial effect on the reliability of the trial process must be presumed.

70 F.3d at 1142.

defense strategy as a result of the intrusion; and (4) whether the overheard conversations and other information were used in any way to the substantial detriment of the petitioner.”

Id. at 965.

At least two other circuits have applied these Weatherford “factors” in determining whether a criminal defendant “has shown the prejudice necessary to make out a sixth amendment violation.” United States v. Kelly, 790 F.2d 130, 137 (D.C. Cir. 1986); accord Sanborn, 629 F.3d at 571. We take no stance on them today. The Weatherford factors and their suitability as a barometer for prejudice goes beyond the issue presented in this appeal: whether the defendant must make a showing of prejudice at all.

35

2. In crafting its structural-error rule, Shillinger misinterpreted binding Supreme Court law.

We disagree with Shillinger’s interpretation of Weatherford and its prejudice requirement, along with Shillinger’s misapplication of other Supreme Court precedents that preclude the type of per se rule Shillinger announced.

Shillinger began its Sixth Amendment analysis with Weatherford, as the seminal case governing intrusions into the attorney-client relationship, but then proceeded to veer away from Weatherford’s “prejudice requirement.” Shillinger, 70 F.3d at 1140. In doing so, the Shillinger court noted that the facts then before the court differed from Weatherford in several key ways: in Shillinger, the prosecutor approached the deputy with the intent to gather information about the defense’s trial preparation, the deputy conveyed that information to the prosecutor, and no legitimate law-enforcement purpose justified the prosecutor’s behavior. Id. at 1137–39. In contrast, the Shillinger court noted that in Weatherford the Supreme Court had “emphasized” the unintentional and justified nature of the informant’s intrusion to conclude that Bursey’s Sixth Amendment rights hadn’t been violated. Id. at 1139. Shillinger quoted a passage from Weatherford to juxtapose the different, and determinative, facts in that case:

Moreover, this is not a situation where the State’s purpose was to learn what it could about the defendant’s defense plans and the informant was instructed to intrude on the lawyer-client relationship or where the informant has assumed for him that task and acted accordingly . . . .

36

....

We may disapprove an investigatory practice only if it violates the Constitution; and judged in this light, the Court of Appeals’ per se rule cuts much too broadly . . . . [U]nless Weatherford communicated the substance of the . . . conversations and thereby created at least a realistic possibility of injury to Bursey or benefit to the State, there can be no Sixth Amendment violation . . . .

There being no tainted evidence in this case, no communication of defense strategy to the prosecution, and no purposeful intrusion by Weatherford, there was no violation of the Sixth Amendment . . . .

Id. (quoting Weatherford, 429 U.S. at 557–58) (emphasis added in Shillinger).

The Shillinger court misconstrued this language as grounds to distinguish Weatherford and to circumvent Weatherford’s holding that denounced per se Sixth Amendment rules against government intrusions. First, Shillinger plucked the above-quoted passage from Weatherford out of context. In that section of the Weatherford opinion, the Supreme Court was elaborating on additional reasons the Fourth Circuit’s rule was too extreme and why Weatherford’s undercover work didn’t contravene the Sixth Amendment. See 429 U.S. at 557.

The Court in Weatherford suggested that Bursey might have presented a valid Sixth Amendment claim with proof that Bursey had been “an informer for the government who ha[d] reported on the conversations to the prosecution and who testifie[d] about them at the defendant’s trial.” Id. The Court further clarified that “Bursey would have a much stronger case” for a Sixth Amendment violation (which we read to say that even then his case might not

37

have been strong enough) had “Weatherford testified at Bursey’s trial as to the conversation between Bursey and [his counsel]; had any of the State’s evidence originated in these conversations; had those overheard conversations been used in any other way to the substantial detriment of Bursey; or even had the prosecution learned from Weatherford, an undercover agent, the details of the [attorney-client] conversations about trial preparations. . . .” Id. at 554 (emphasis added).

As seen, the thrust of the Court’s analysis focused on whether Bursey could show substantial detriment from the use of the confidential information at trial. See id. And because the information hadn’t been used, the Court deduced that there had been no potential for substantial detriment to Bursey and therefore no prejudice. See id. at 558 (“[U]nless Weatherford communicated the substance of the [attorney-client] conversations and thereby created at least a realistic possibility of injury to Bursey or benefit to the State, there can be no Sixth Amendment violation.”). The Court ruled that Bursey’s § 1983 claim failed for lack of a Sixth Amendment violation, because Bursey had failed to show “tainted evidence,” any “communication of defense strategy to the prosecution,” and “purposeful intrusion by Weatherford.” Id.

Second, Shillinger misunderstood Weatherford’s prejudice requirement as being conditioned on two facts that were part of that case: an unintentional intrusion by an informant that was justified by “the requirements of ‘effective law enforcement.’” Shillinger, 70 F.3d at 1139–40 (quoting Weatherford, 429

38

U.S. at 557)). Shillinger inferred that when these conditions are flipped—when the government intrudes intentionally and without a legitimate law-enforcement purpose—prejudice must be presumed. See id. at 1140 (“Weatherford may not dictate a rule that would require a showing of prejudice in cases where intentional prosecutorial intrusions lack a legitimate purpose.”). But Weatherford didn’t hinge the prejudice inquiry on the intentionality or legitimacy of the government’s intrusion. See 429 U.S. at 558. It merely explained that when these factors are missing—when the government wasn’t trying “to learn what it could about the defendant’s defense plans” or when the intrusion was merely an “unfortunate necessity of undercover work”—then there isn’t even “a realistic possibility of injury” on which the defendant can hang his Sixth Amendment hat. Id. at 557–58. And that possibility becomes only “much stronger,” the Supreme Court said, when confidential communications are “reported on . . . to the prosecution” or when an informant “testifies about [the conversations] at the defendant’s trial.” Id. at 554. For even these more egregious situations, Weatherford never stated or suggested that a showing of actual prejudice becomes unnecessary. See id. at 558 (affirming that “there can be no Sixth Amendment violation” without “at least a realistic possibility of injury”). Shillinger misunderstood Weatherford to say that the prejudice inquiry depends on the intentional or legitimate nature of the

39

government’s intrusion. But Weatherford affirms that, in any scenario, a defendant asserting a Sixth Amendment intrusion claim must show prejudice. 20

20 The authorities that Shillinger relies on to reach its contrary conclusion don’t support the proposition that Weatherford obviates the prejudice requirement for intentional, unjustified governmental intrusions. Shillinger leans on two cases for its “suggest[ion]” that Weatherford’s holding didn’t govern. 70 F.3d at 1140 (first citing Briggs v. Goodwin, 698 F.2d 486, 493 n.22 (D.C. Cir. 1983), reh’g granted, opinion vacated, and on reh’g, 712 F.2d 1444 (D.C. Cir.), cert. denied, 464 U.S. 1040 (1984); and then citing United States v. Morales, 635 F.2d 177, 179 (2d Cir. 1980)). First, we find Briggs wholly unpersuasive because the opinion makes several statements that either contravened Weatherford at the time or have since been contradicted by the D.C. Circuit’s own intervening authority. Compare Briggs, 698 F.2d at 495 (“[O]nce the investigatory arm of the government has obtained information, that information may reasonably be assumed to have been passed on to other governmental organs responsible for prosecution.”), with Weatherford, 429 U.S. at 556–57 (“Nor do we believe that federal or state prosecutors will be so prone to lie . . . that we must always assume . . . that an informant communicates what he learns . . . .”); and compare Briggs, 698 F.2d at 494 (“[T]he appellants need not prove that the prosecution actually used the information obtained.”), with ACLU Found. of S. Cal. v. Barr, 952 F.2d 457, 472 (D.C. Cir. 1991) (“[T]he constitutional right to counsel in a criminal case is violated only if the intercepted communications are somehow used against the defendant . . . .”).

And in Morales, the Second Circuit considered the appellants’ claim that, because a codefendant was later discovered to be a registered DEA informant, that meant there had been “a spy in the enemy camp” throughout the entire prosecution. 635 F.2d at 178. The Second Circuit rejected the appellants’ claim that “the Government’s conduct was so grossly improper that a showing of prejudice was not required” because the district court’s in camera review of the evidence revealed no “intentional, governmentally instigated intrusion” into confidential attorney-client communications. Id. at 178–79. Shillinger interprets Morales to say that prejudice would have been presumed if the government’s intrusion had been intentional. We dispute that logic. We do not assume that a court’s rejection of one theory automatically constitutes its affirmance of the opposite. In any case, since Morales, the Second Circuit has affirmed its view that a showing of prejudice is required to complete any Sixth Amendment right-to-confidentiality violation. See Chandler, 56 F.4th at 37 (reviewing Second Circuit precedents which state that “resulting prejudice” is (footnote continued)

40

Third, Shillinger’s conclusion that Weatherford might allow a structural-

error rule for certain types of governmental intrusions contradicts Weatherford’s general repudiation of per se rules to protect attorney-client confidentiality. The Supreme Court rejected the Fourth Circuit’s per se rule because the court had “deemed” “trial prejudice . . . irrelevant.” Id. at 553. The references to “prejudice” and “detriment” to the defendant peppered throughout the Weatherford opinion impress upon us that a Sixth Amendment claim cannot be “made out” without prejudice. Id. at 556; see id. at 550, 552, 556–57, 561. Despite these repeated references, Shillinger deduced that a structural-error rule is needed to vindicate a defendant’s Sixth Amendment rights when the government intrudes intentionally and unjustifiably into the attorney-client relationship. 70 F.3d at 1142. But this rule betrays Weatherford. In Weatherford, the Supreme Court stated that such per se Sixth Amendment rules “cut[] much too broadly” to safeguard the Amendment’s guarantees, 429 U.S. at 557, partly because these rules indiscriminately recognize constitutional violations “whether or not any specific prejudice to the defendant[] . . . is demonstrated or otherwise threatened,” id. at 550. We cannot see how Shillinger’s holding squares with Weatherford’s bottom line that a Sixth Amendment violation must include proof that the defendant was adversely

required to sustain a Sixth Amendment violation, even when “the government had intentionally invaded the attorney client relationship” (quoting United States v. Dien, 609 F.2d 1038, 1049 (2d Cir. 1979)).

41

affected by the government’s intrusion during trial. So for all these reasons, Shillinger’s interpretation of Weatherford misses the mark.

Shillinger also misconstrues Morrison. The issue in Morrison was whether dismissal of the defendant’s indictment was an appropriate remedy for the government’s intentional, unjustified intrusion into her relationship with her attorney, despite the fact that she “ha[d] demonstrated no prejudice” from the intrusion “of any kind.” 449 U.S. at 366. To decide that issue, the Supreme Court explained that it would first “identify . . . the taint”—which is to say the prejudice endured—“and then neutralize” it “by tailoring relief appropriate in the circumstances to assure the defendant the effective assistance of counsel and a fair trial.” Id. at 365. But because the defendant had “demonstrated no prejudice,” the Court concluded there was no injury of “a constitutional dimension” that “need[ed] to be purged” and “accordingly” the government’s nonprejudicial intrusion did not warrant any remedy, “much less the drastic relief” of dismissal. Id. at 366–67.

Shillinger surmised from this analysis that Morrison “declined to reach the issue” of whether “there could be [a] Sixth Amendment violation absent proof that the intrusion prejudiced the defendant.” Shillinger, 70 F.3d at 1140. In this, Shillinger is mistaken. Shillinger supposes that Morrison “declined to reach,” id., the prejudice issue because Morrison “assume[d], without deciding, that the Sixth Amendment was violated in the circumstances of th[e] case,” Morrison, 449 U.S. at 364. But Morrison assumed the intrusion component of

42

the defendant’s Sixth Amendment violation, not the prejudice component. See id. The Morrison Court “assume[d], without deciding, that the Sixth Amendment was violated,” id., by which the Court meant it would assume the DEA agents’ visiting the defendant in jail and advising her to fire her attorney constituted an impermissible intrusion into the attorney-client relationship, id. at 362. It did not assume that this intrusion caused prejudice. See id. at 365 (explaining that “[t]he premise of our prior cases is that the constitutional infringement identified has had or threatens some adverse effect . . . or has produced some other prejudice to the defense”). We suspect that Shillinger mistook Morrison’s use of the term “the violation” to mean the entire Sixth Amendment violation—an understandable misconception. See id. (“[A]bsent demonstrable prejudice, or substantial threat thereof, dismissal of the indictment is plainly inappropriate, even though the violation may have been deliberate.” (emphasis added)). But Morrison clearly intended “the violation” to mean only “the intrusion.” 21 This is the only sensible way to read the case: How could Morrison have presumed prejudice and then gone on to deny the defendant relief because she “demonstrated no prejudice”? 449 U.S. at 366.

21 The Fifth and Sixth Circuits have suggested this same interpretation of Morrison. See United States v. Melvin, 650 F.2d 641, 645 (5th Cir. 1981) (discussing whether “this type of Sixth Amendment violation” requires an intrusion into a confidential relationship (emphasis added)); Sanborn, 629 F.3d at 571 (using Morrison to support the court’s reasoning that “[e]ven if” the defendant had shown a “purposeful intrusion,” he would still need to show prejudice (citing Morrison, 449 U.S. at 365–66)).

43

Further, Morrison (like Shillinger) dealt with an intentional and unjustified intrusion into the defense camp, and yet the Court didn’t presume the defendant’s entitlement to any remedy. 22 See id. (stating that a deliberate intrusion did not justify dismissal of the indictment). The Court explained that a remedy “should be tailored to the injury suffered.” Id. at 364. By directing us to calibrate the appropriate remedy from a defendant’s injury, Morrison presupposes that by the remedies stage some demonstration of prejudice has already occurred. See id. at 365 (presuming under the remedy analysis that “the constitutional infringement identified . . . had or threatens some adverse effect upon the effectiveness of counsel’s representation or has produced some other prejudice to the defense”). Because without “some . . . prejudice,” the Court said, “there is no basis for imposing a remedy.” Id. In so stating, Morrison clarifies that the injury sustained from a Sixth Amendment attorney-client- confidentiality violation is not the government’s intrusion itself. See id. The injury is the “adverse effect” on defense counsel’s “ability . . . to provide adequate representation in the[] criminal proceedings” because this injury jeopardizes the fairness of the defendant’s trial, and so it “needs to be purged to

22 In his supplemental brief, Hohn attempts to distinguish Morrison from Shillinger on the ground that “Morrison involved an attempt to interfere with the overall attorney-client relationship,” rather than “attorney-client communications.” Suppl. Reply Br. at 5. Shillinger specifically used the term “attorney-client relationship” in its holding, see 70 F.3d at 1142, so we don’t find this distinction meaningful or convincing.

44

make certain that [the defendant] has . . . not [been] unfairly convicted.” Id. at 365–66.

Shillinger wrongly interpreted Morrison as further proof that “Weatherford—and the prejudice requirement articulated in that case—does not necessarily govern intentional intrusions by the prosecution that lack a legitimate purpose.” Shillinger, 70 F.3d at 1140. We disagree that Morrison supports that reading of Weatherford. Rather, Morrison bolsters Weatherford’s prejudice requirement by reiterating that a “constitutional infringement” under the Sixth Amendment requires “some adverse effect” to the defendant, whether it be to the effectiveness of counsel or “some other” “impact on the criminal proceeding.” Morrison, 449 U.S. at 365.

Shillinger also fails to grapple with Cronic’s limited categories for recognizing structural error in Sixth Amendment right-to-counsel claims. See Cronic, 466 U.S. at 659–61. Cronic conveys that structural error should apply to these claims in only extreme situations. See id.; Orduno-Ramirez, 61 F.4th at 1268. Shillinger appears inconsistent with Cronic’s narrow view of structural error. Though Shillinger acknowledges Cronic’s “discuss[ion]” of “circumstances justifying a presumption of prejudice,” it never states which of Cronic’s three structural-error classifications made a presumption of prejudice

45

justified in Haworth’s circumstances. 23 Shillinger, 70 F.3d at 1141. In this way, Shillinger treats the structural-error analysis too casually, when the upshot of Cronic (and its predecessors) was for courts to do so cautiously, particularly in the Sixth Amendment context. See Cronic, 466 U.S. at 658.

Finally, Shillinger cites several other Supreme Court cases that recognize the government’s interference into the attorney-client relationship as per se prejudicial, intimating that Shillinger’s structural-error rule is simply the latest addition to a well-established body of law. 70 F.3d at 1141 (citing Ferguson v. Georgia, 365 U.S. 570, 594–95 (1961) (defense counsel prevented from directly examining the defendant); Brooks v. Tennessee, 406 U.S. 605, 610 (1972) (defendant forced to testify before other defense witnesses); Herring v. New York, 422 U.S. 853, 863 (1975) (defense counsel denied opportunity to make closing argument in a bench trial); Geders v. United States, 425 U.S. 80, 88 (1976) (defendant prohibited from consulting with his attorney during overnight recess between his direct-examination and cross-examination)).

What Shillinger disregards is that almost all the cases it relies on address instances of judicial interference with either the defendant’s or the attorney’s

23 We have recently affirmed the three Cronic categories of structural error as such: “(1) the defendant suffers ‘the complete denial of counsel . . . at a critical stage’ of the criminal justice process; (2) ‘counsel entirely fails to subject the prosecution’s case to meaningful adversarial testing’; and (3) when ‘the likelihood that any lawyer, even a fully competent one, could provide effective assistance is so small that a presumption of prejudice is appropriate.’” Orduno-Ramirez, 61 F.4th at 1268 (quoting Cronic, 466 U.S. at 659–61).

46

fundamental ability to conduct a full-throated defense at trial. 24 Because the judicial interference in those cases jeopardized the integrity and fairness of the trial itself, the prejudicial impact was tangible. See Ferguson, 365 U.S. at 594– 95 (remarking that, without direct examination, the defendant “may fail properly to introduce or to introduce at all, what may be a perfect defense,” in which case “he faces the danger of conviction because he does not know how to establish his innocence” (citation omitted)); Brooks, 406 U.S. at 610 (discerning that by forcing the defendant to testify before his witnesses he “risk[s] the danger of taking the stand” without knowing “whether his own testimony will be necessary or even helpful to his cause”); Herring, 422 U.S. at 863 (“[T]he difference in any case between total denial of final argument and a concise but persuasive summation could spell the difference, for the defendant, between liberty and unjust imprisonment.”); Geders, 425 U.S. at 88 (explaining

24 Hohn does cite a prosecutorial-intrusion case, Maine v. Moulton, 474 U.S. 159 (1985), for the proposition that the government must honor the Sixth Amendment right to counsel once it has attached and that it thus has an “affirmative obligation not to act in a manner that circumvents” the right. Suppl. Br. at 8 (quoting Moulton, 474 U.S. at 176). In Moulton, prosecutors had a cooperating defendant wear a wire to a meeting with another defendant at which the two men had agreed to plan a defense strategy. Id. at 164–65. Though the prosecutors had instructed the cooperator not to question the codefendant at the meeting, the cooperator’s remarks caused the codefendant to make incriminating comments. Id. at 165–66. The Supreme Court ruled that the district court’s admitting the incriminatory comments at trial violated the codefendant’s Sixth Amendment rights. Id. at 180. So the Court suppressed the incriminatory statements, leaving available any evidence unrelated to crimes for which the Sixth Amendment right to counsel had not yet attached at the time the evidence was obtained. Id. The Court’s decision does not find or even mention structural error.

47

the impact of preventing a defendant from consulting with his attorney during an overnight recess because these periods are “often time of intensive work,” when “tactical decisions [are] to be made,” “strategies to be reviewed,” and when the lawyer might “obtain from his client information made relevant by the day’s testimony”).

The impact on fundamental fairness from a Shillinger-type intrusion is more inchoate. Shillinger doesn’t explain how the government’s obtaining confidential communications would so obviously prejudice a defendant absent any proof that the overheard information was actually used against the defendant at trial. Shillinger baldly concludes that the “sort of purposeful intrusion,” 70 F.3d at 1141, that the prosecution committed “constitutes a direct interference with the Sixth Amendment rights of a defendant,” id. at 1142, but without elucidating why such an intrusion necessarily undermines the fundamental fairness of the defendant’s trial in every case. In Ferguson, Brooks, Herring, and Geders, the Supreme Court explained how the government’s particular policy so fractured the adversarial process as to render any trial conducted under that policy an unreliable “vehicle for determin[ing] [the defendant’s] guilt or innocence.” Fulminante, 499 U.S. at 310. Shillinger insists that the prosecutor’s intentional intrusion into the attorney-client relationship, without any law-enforcement justification, causes a comparable fissure in the adversarial edifice. See 70 F.3d at 1142. Yet Shillinger’s structural-error rule “literally read” sweeps so broadly that it demands

48

prejudice be presumed for any intentional, unjustified intrusion into an attorney-client conversation, Weatherford, 429 U.S. at 558, perhaps even if all the attorney says to the defendant is, “Hello, how are you? When are you available to meet?” Though we make no comment as to whether these “harmless subjects” would qualify as “confidential communications” for Sixth Amendment purposes, 25 id., this example of trivial communications illustrates a

25 Hohn insists that Shillinger’s rule wouldn’t capture such mundane conversations because confidential attorney-client communications, for purposes of the Sixth Amendment, are limited to those that pertain to “legal advice or strategy that the defendant has not disclosed to the prosecution.” Suppl. Br. at. 1. This assertion, which Hohn supports with one of our cases discussing the attorney-client privilege (not Sixth Amendment confidentiality), see United States v. Johnston, 146 F.3d 785, 794 (10th Cir. 1998), even if true, doesn’t assuage our general concern that Shillinger’s rule is overbroad. In Weatherford, the Supreme Court expressed a similar concern that the Fourth Circuit’s per se rule, “literally read, would cloud Bursey’s subsequent conviction,” even if “the [attorney-client] conversation was confined to the weather or other harmless subjects.” 429 U.S. at 558. And partly for that reason, the Court reversed the Fourth Circuit’s rule, which would have granted Bursey relief “although there would have been no constitutional violation.” Id. Even though Hohn’s intercepted call contained trial strategy, we cannot assume that will be the case for every defendant. Yet Shillinger’s structural-error rule asks us to put on blinders regarding the contents of the overheard attorney- client communication. See 70 F.3d at 1142 (holding that “a prosecutor’s intentional intrusion into the attorney-client relationship . . . absent a countervailing state interest . . . must constitute a per se violation of the Sixth Amendment”). But we know that the contents matter greatly because, as Weatherford points out, not all attorney-client conversations contain information that would prejudice the defendant if used at trial. See 429 U.S. at 556–57 (“Nor do we believe that . . . we must always assume . . . that what [an informant] communicates has the potential for detriment to the defendant or benefit to the prosecutor’s case.”). The range of topics an attorney and client might discuss is precisely why Weatherford denounces per se rules and requires a showing of prejudice—to avoid defendants receiving relief without having suffered any real injury. See id. at 558.

49

larger point: the Sixth Amendment’s purpose is to “assure[] the fairness, and thus the legitimacy, of our adversary process,” Kimmelman v. Morrison, 477 U.S. 365, 374 (1986), not to guarantee that all attorney-client conversations will “be free from intrusion,” Weatherford, 429 U.S. at 553. This is why the Supreme Court instructs us that, notwithstanding an intentional prosecutorial intrusion, if the trial’s integrity is untarnished, then the show must go on. See Morrison, 449 U.S. at 366 (emphasizing that without “demonstrable prejudice” there is “no justification for interfering with the criminal proceedings” because there is “no effect of a constitutional dimension which needs to be purged”).

In sum, Supreme Court precedents predating Shillinger establish that the right to communicate confidentially with an attorney is not one that exists “for its own sake,” Cronic, 466 U.S. at 658, but rather one that exists because of its positive residual effect on the fairness of criminal proceedings, Strickland, 466 U.S. at 685. Shillinger’s structural-error rule loses sight of this purpose. Worse still, it directly contradicts Weatherford’s direction not to “always assume” that confidential information relayed to the prosecution “has the potential for detriment to the defendant or benefit to the prosecutor’s case.” 429 U.S. at 557. This language, along with the other caselaw discussed above, forecloses Shillinger’s holding and so we must overturn it. Instead, we hold that a Sixth Amendment violation of the right to confidential attorney-client communications requires the defendant to show trial prejudice resulting from

50

the government’s intrusion into the attorney-client relationship, even when the intrusion was purposeful and done without any law-enforcement justification.

3. None of Hohn’s other arguments convince us to uphold Shillinger’s structural-error rule.

Hohn makes several arguments in defense of Shillinger’s structural-error rule. First, Hohn contends Shillinger correctly held that purposeful, unjustified intrusions into the attorney-client relationship are “never harmless because they necessarily render a trial fundamentally unfair.” 70 F.3d at 1142 (cleaned up). But as we discussed above, see Discussion II.B.2, supra, Shillinger deficiently explains why these intrusions make every trial fundamentally unfair, even when the prosecution never uses the confidential information against the defendant any time during the proceedings. Shillinger and Hohn insist that an intentional, unjustified intrusion automatically renders a trial unfair, yet Hohn concedes that neither his trial nor his sentencing were made unfair by AUSA Morehead’s becoming privy to his six-minute call. Thus, Hohn’s warning about a specter of fundamental unfairness rings hollow, at least in his case. Certainly, we condemn the Kansas USAO’s practice. But Hohn’s case illustrates that the Kansas USAO’s misguided routines did not prejudice at least some of CoreCivic’s detainees, including Hohn. And as the government points out, many of the other defendants in the consolidated § 2255 litigation made “similar concessions” to Hohn’s regarding prejudice. Suppl. Resp. Br. at 17. So a case-by-case prejudice inquiry for each CoreCivic detainee is entirely

51

appropriate. We also reject Hohn’s assertion that the “systematic and pervasive” nature of the Kansas USAO’s recording scheme, alone, should compel us to keep Shillinger’s structural-error rule. Suppl. Br. at 19 (quoting Bank of Nova Scotia v. United States, 487 U.S. 250, 259 (1988)). Were that true, we would have applied Shillinger’s structural-error rule in Spaeth and Orduno-Ramirez—as well—cases spawned from the same scandal—and yet we did not. See Spaeth, 69 F.4th at 1211; Orduno-Ramirez, 61 F.4th at 1266.

Second, Hohn argues that Shillinger’s structural-error rule is sound because the narrow class of intrusions to which the rule applies make prejudice “so likely” that evaluating prejudice for each individual defendant is not “worth the cost.” Suppl. Br. at 19 (quoting Shillinger, 70 F.3d at 1142). Shillinger itself defeats this argument. In Shillinger, this court said that “under the facts of this case the prejudice standard articulated in Weatherford has been met,” 70 F.3d at 1139—meaning Shillinger would have found the prosecutor’s behavior prejudicial on the facts without applying structural error. 26 So evidently, prejudice was not so hard to measure there, and we fail to see why it would be here. Hohn’s Sixth Amendment challenge pertains to one recorded attorney- client call that lasted six minutes. This makes the analysis even simpler than

26 The government leverages this statement from Shillinger to argue that “[t]he rest of the opinion” was “not ‘essential to the determination of the case,’” and therefore Shillinger’s structural-error rule was dictum. Suppl. Resp. Br. at 18 (quoting United States v. Moore, 96 F.4th 1290, 1300 (10th Cir. 2024)). Because we overrule Shillinger on the merits, we do not reach this argument.

52

the one in Shillinger, where this court acknowledged “that it is . . . impossible to know what information obtained by the prosecution from the deputy was used at trial without knowing the extent of the information that was obtained.” 70 F.3d at 1138. In Hohn’s case, we know the extent of the information obtained: the recorded six-minute phone call. Though Hohn has not provided us with the call recording—an omission that speaks volumes—it seems that all a prejudice analysis would entail would be for Hohn to connect something he and his attorney discussed during those six minutes to anything used during the criminal proceedings that either disadvantaged him or advantaged the prosecution, and then for the district court to rule.

We also disagree that a defendant need not show prejudice or that prejudice becomes immeasurable when attorney-client communications contain trial strategy. See Hari, 67 F.4th at 912 (requiring the defendant to “show[] [that] the materials in question”—handwritten notes reflecting the defendant’s trial strategy—“were even referred to at trial, much less used as substantive evidence in the government’s case” to establish a Sixth Amendment violation); United States v. Costanzo, 740 F.2d 251, 256–57 (3d Cir. 1984) (rejecting the defendant’s Sixth Amendment claim that an informant’s disclosure “compromised [the defendant’s] confidential defense strategy” partly because the defendant “ha[d] not argued that actual prejudice resulted from the . . . disclosure”). Hohn argues that knowledge of trial strategy “allows the prosecutor to ‘anticipate and counter’ the defense,” giving the prosecution an

53

“upper hand” that is “beyond question harmful to any defendant.” Suppl. Br. at 19. Yet Hohn never argues that the prosecutor had an “upper hand” at his trial, nor does he make any claim that the prosecution used the call’s contents to “anticipate or counter” his trial defense. This takes the wind out of his sails. Without any demonstration that the overheard communications “produced, directly or indirectly, any of the evidence offered at trial,” Hohn cannot establish a Sixth Amendment violation. Weatherford, 429 U.S. at 552 (emphasis added).

Third, Hohn asserts that Shillinger’s structural-error rule falls within one of the Supreme Court’s designated structural-error rationales—“if the right at issue is not designed to protect the defendant from erroneous conviction but instead protects some other interest.” McCoy, 584 U.S. at 427 (emphasis added) (citation omitted). In making this argument, Hohn wrongly lumps the right to attorney-client confidentiality in with the autonomy rights established in McCoy (the right to maintain innocence at trial), McKaskle (the right to self- representation), and Gonzalez-Lopez (the right to counsel of choice). Hohn misses that this other bucket of rights exists for an entirely distinct and unique purpose: to guarantee that the defendant “ha[s] his voice heard.” McKaskle, 465 U.S. at 174. The autonomy rights safeguard the defendant’s power to steer the ship of his own defense. See United States v. Rosemond, 958 F.3d 111, 122 (2d Cir. 2020) (“[I]t is the defendant’s prerogative, not counsel’s, to decide on the objective of his defense,” while “trial management” and deciding “what

54

arguments to pursue” “is the lawyer’s province.” (cleaned up)). Our interest in honoring these rights predates the Constitution because doing so recognizes a longer-held “respect for the individual [as] the lifeblood of the law.” McCoy, 584 U.S. at 421 (citation omitted) (noting that the right to self-representation was recognized “[a]s the laws of England and the American colonies developed”). In this way, deprivation of an autonomy right offends a defendant’s liberty more profoundly and intrinsically than the violation of a right derived solely to promote adversarial fairness—like effective assistance of counsel. See Gonzalez-Lopez, 548 U.S. at 146–48 (distinguishing between “the right to the effective assistance of counsel” and “the right to counsel of choice” because the latter “has never been derived from the Sixth Amendment’s purpose of ensuring a fair trial” but rather “has been regarded as the root meaning of the constitutional guarantee”). The distinction between effective- assistance rights and autonomy rights explains why the first category requires a showing of prejudice and the second one does not. Because the defendant possesses a right to “effective (not mistake-free) representation,” “a violation of the Sixth Amendment right to effective representation is not ‘complete’ until the defendant is prejudiced.” Id. at 147.

The right to communicate confidentially with an attorney originates from the Sixth Amendment’s promise of effective assistance of counsel. See Weatherford, 429 U.S. at 554 n.4 (recognizing that “government interception of attorney-client communications” “threat[ens] . . . the effective assistance of

55

counsel”); United States v. Dyer, 821 F.2d 35, 37 (1st Cir. 1987) (“[T]he essence of the sixth amendment right [to effective assistance of counsel] is, indeed, privacy of communication with counsel” (quoting United States v. Rosner, 485 F.2d 1213, 1224 (2d Cir. 1973)). So in our view, the legal principles that govern effective-assistance claims apply equally to attorney- client confidentiality. 27 The Supreme Court instructs that, because we derive “the right to effective representation from the purpose of ensuring a fair trial,” we should “also derive[] the limits of that right from that same purpose.” Gonzalez-Lopez, 548 U.S. at 147; see Mickens, 535 U.S. at 166 (affirming that the right to effective assistance of counsel exists “not for its own sake” but to assure trial fairness (quoting Cronic, 466 U.S. at 658)); cf. Lafler v. Cooper, 566 U.S. 156, 178 (2012) (Scalia, J., dissenting) (emphasizing that the right to effective assistance “is not infringed unless” the incident under review “call[s] into question the basic justice of a defendant’s conviction or sentence”). As a derivative of the effective-assistance guarantee, it follows that the right to

27 In his original opening brief to this court, Hohn styled his claim as an “effective-assistance claim,” Op. Br. at 13, 25, and likewise Shillinger’s structural-error rule as an “effective assistance test,” id. at 37. Indeed, throughout Hohn’s opening brief, he seems to presuppose that effective- assistance-of-counsel principles inform his Sixth Amendment confidentiality right vis-à-vis the attorney-client privilege. But his position shifts in the supplemental briefing, in which he suggests that this court ought to treat effective-assistance and intrusion-based Sixth Amendment claims differently. See Suppl. Br. at 12–13 (“It is only if the defendant raises a Sixth Amendment claim that is rooted in defense counsel’s deficient performance that the defendant must prove prejudice . . . .”).

56

attorney-client confidentiality also exists solely for the purpose of ensuring a fair trial and not to “protect[] some other interest.” McCoy, 584 U.S. at 427. So contrary to Hohn’s argument, the right to confidential attorney-client communications does not vindicate “some other interest” that warrants structural error.

Fourth, Hohn suggests that the Supreme Court’s earlier precedents addressing Sixth Amendment attorney-client confidentiality—Black, O’Brien, Weatherford, and Hoffa—“impl[y]” that the government’s purposefully obtaining and becoming privy to confidential attorney-client communications without law-enforcement justification constitutes structural error. Suppl. Br. at 22. But the Weatherford Court’s interpretation of Black, O’Brien, and Hoffa defeats this argument. See Discussion II.A, supra. Weatherford clarified that Black, O’Brien, and Hoffa do not condone per se Sixth Amendment intrusion rules but rather emphasized that those cases support the defendant’s need to tether governmental intrusion to a realistic possibility of injury from the use of confidential communications at trial. See id. Other circuits have faced arguments identical to Hohn’s and accordingly rejected them under Weatherford. See, e.g., Kelly, 790 F.2d at 136–37 (“In Weatherford . . . the Court rejected Kelly’s reading of Black and Hoffa as creating a per se rule of presumed prejudice from any governmental intrusion.”). Hohn doesn’t address Weatherford’s unfavorable discussions of Black, O’Brien, and Hoffa, so we need go no further to dismiss this argument. See United States v. Walker, 918

57

F.3d 1134, 1151 (10th Cir. 2019) (arguments “inadequately presented” to us are waived (citation omitted)).

Fifth and finally, Hohn presses that, because Shillinger’s holding is appropriately narrowed to “only the most egregious . . . prosecutorial intrusions,” its structural-error rule is justified. Suppl. Br. at 23. But Morrison also dealt with facts alleging the “most egregious” behavior—an intentional, unjustified governmental intrusion—and yet the Supreme Court still tied the defendant’s remedy to the injury she suffered. 449 U.S. at 364. Finding she had suffered none, because she failed to demonstrate or even allege any prejudice, the Court denied her relief. See id. at 366–67. It didn’t simply hold that “[t]his type of misconduct should be remedied in every case.” Suppl. Br. at 23. Thus, the scope of Shillinger’s structural-error rule, however narrow, is not enough to save it.

C. Federal Circuit Caselaw

1. A majority of circuits either support or are consistent with our view that constitutional claims like Hohn’s require the defendant to show prejudice.

A majority of the circuit courts support our revised view that Sixth Amendment claims concerning purposeful, unjustified intrusions into the attorney-client relationship require the defendant to show prejudice and that such prejudice accrues “only if the intercepted communications are somehow used against the defendant . . . in connection with the underlying proceeding.” ACLU Found. of S. Cal. v. Barr, 952 F.2d 457, 472 (D.C. Cir. 1991) (emphasis

58

added); see, e.g., United States v. Collins, 799 F.3d 554, 591 (6th Cir. 2015) (“[T]o establish a violation of the Sixth Amendment right to counsel ensuing from government surveillance, a claimant must . . . show . . . that the information gained was used to prejudice the claimant’s defense in his criminal trial.”); Nordstrom v. Ryan, 762 F.3d 903, 911 (9th Cir. 2014) (contemplating the type of “harm” sufficient to “cause[] prejudice” from “an improper intrusion into the attorney-client relationship,” and concluding that “tainted evidence . . . used against [the defendant]” would qualify); see also, e.g., United States v. Esformes, 60 F.4th 621, 632–33 (11th Cir. 2023), cert. denied, 144 S. Ct. 485; United States v. Hari, 67 F.4th 903, 912–13 (8th Cir. 2023), cert. denied, 144 S. Ct. 436; United States v. Allen, 491 F.3d 178, 192 (4th Cir. 2007); Williams v. Woodford, 384 F.3d 567, 584–85 (9th Cir. 2004); United States v. Danielson, 325 F.3d 1054, 1068–70 (9th Cir. 2003); United States v. Castor, 937 F.2d 293, 297–98 (7th Cir. 1991); United States v. Kelly, 790 F.2d 130, 136–38 (D.C. Cir. 1986); United States v. Ginsberg, 758 F.2d 823, 833 (2d Cir. 1985); United States v. Mastroianni, 749 F.2d 900, 906–08 (1st Cir. 1984); United States v. Steele, 727 F.2d 580, 585–87 (6th Cir. 1984); United States v. Melvin, 650 F.2d 641, 643–44 (5th Cir. 1981). The government culls this caselaw in its supplemental brief, see Suppl. Resp. Br. at 9 n.1, to suggest that any decision from this court rejecting Shillinger’s structural-error rule would find good company among our fellow circuits.

59

Hohn attempts to distinguish some of this caselaw by alleging that these cases do not invoke the same “discrete, trial-specific harm” as the one in Shillinger. Suppl. Reply Br. at 6. But this approach assumes that the prejudice component of a Sixth Amendment claim depends on the nature of the inciting intrusion, a theory we already refuted. See Discussion II.B.2, supra. Regardless, we view these cases as (at best) supportive and (at worst) consistent with our decision to reverse Shillinger’s structural-error rule.

To start, Hohn asserts that the Fourth and Seventh Circuit cases do not undermine Shillinger because those cases were missing “the Shillinger requirement that the prosecutor become privy to strategic communications.” Suppl. Reply Br. at 6. Hohn’s assessment is incomplete. In the Fourth Circuit case, United States v. Allen, the court rejected the defendant’s Sixth Amendment claim based on the district court’s allowing the government to view a document related to the defense’s cross-examination, in part because the court took care to screen off the appropriate AUSA, but also because the defendant “d[id] not allege any prejudice” “nor [was] prejudice clear from the record.” 491 F.2d at 192. Had “the defense’s cross-examination [been] impaired in any respect” from the document’s use, the court considered, the defendant may have been prejudiced. See id. But it “was not [so] impaired,” and so without any other showing that “prejudice resulted from th[e] [district court’s] arrangement,” the Fourth Circuit rejected the defendant’s Sixth Amendment argument. See id. (“It is well settled that some showing of prejudice is a

60

necessary element of a Sixth Amendment claim based on an invasion of the attorney-client relationship.” (quoting United States v. Chavez, 902 F.2d 259, 266 (4th Cir. 1990) (citing Weatherford, 429 U.S. at 558))). Thus, the Fourth Circuit’s decision affirms that, even where the government intrudes intentionally into the defense camp, the defendant must make an allegation of prejudice to have a chance of prevailing on a Sixth Amendment intrusion claim.

So too with the Seventh Circuit case, United States v. Castor, in which the court denied the Sixth Amendment claim because the defendant “admit[ted] he cannot show prejudice,” and “[w]ithout any proof of . . . actual prejudice, the defendant cannot assert that . . . the case violates his constitutional right to counsel.” 937 F.2d at 297–98. The Seventh Circuit stressed that, “[w]here the sixth amendment right to attorney-client confidentiality exists, prosecutorial violation of that privilege might lead to reversal of a resulting conviction,” but only if “the defendant c[an] show prejudice.” Id. at 297 (emphasis added). The court did not condition this prejudice requirement on the type of prosecutorial violation or on the prosecution’s becoming privy to the contents of the communication. See id.

Similarly, in attacking the Sixth, Eighth, and Ninth Circuit cases, Hohn emphasizes that all of them lack two “Shillinger requirement[s]”: the prosecutor’s becoming privy to privileged communications and an intentional intrusion from the government. Suppl. Reply Br. at 6. But once again, Hohn ignores that these factual differences regarding the nature of the government’s

61

intrusion (the first component of a Sixth Amendment claim) existed separately from the circuit courts’ analysis of prejudice (the second component of a Sixth Amendment claim) as it applies generally.

True, in United States v. Steele, the Sixth Circuit dealt with an unintentional-government-intrusion claim, but in establishing the standard for Sixth Amendment right-to-counsel violations the court remarked broadly that “[e]ven where there is an intentional intrusion by the government . . . prejudice to the defendant must be shown.” 727 F.2d at 586 (emphasis added). In United States v. Hari, the Eighth Circuit rejected the defendant’s Sixth Amendment claim partly on the ground to which Hohn refers—that the government’s “accidently receiv[ing],” 67 F.4th at 911, confidential materials lacked the element of “deliberate intrusion” needed to establish a Sixth Amendment claim, id. at 912 (quoting United States v. Tyerman, 701 F.3d 552, 559 (8th Cir. 2012))—but also because there was “no showing the materials in question were even referred to at trial, much less used as substantive evidence in the government’s case,” id. at 913. This two-part analysis demonstrates that, had the defendant alleged an intentional government intrusion, the Eighth Circuit still would have required an additional showing that “the overheard conversations produced, directly or indirectly, any of the evidence offered at trial,” and so the court would have dismissed the claim on that alternate ground anyway. Id. (quoting Weatherford, 429 U.S. at 552). A similar outcome resulted in Williams v. Woodford, where the Ninth Circuit ascertained that “[e]ven if

62

[the court] assumed that the jailhouse monitoring” of the defendant’s calls was “deliberate state interference with the confidential relationship between [defendant] and his counsel, [defendant] fails to establish substantial prejudice.” 384 F.3d at 585. “Substantial prejudice results,” the Ninth Circuit specified, “from the introduction of evidence gained through the interference against the defendant at trial, from the prosecution’s use of confidential information pertaining to defense plans and strategy, and from other actions designed to give the prosecution an unfair advantage at trial.” Id. Because the defendant had not shown any such “use[]” of “any confidential information obtained from the monitoring . . . to achieve an unfair advantage at trial” the Ninth Circuit denied his COA application. Id.

In Ginsberg, the Second Circuit reviewed the district court’s decision to deny the defendant’s motion for an evidentiary hearing based on his claim that the government intruded into the defense camp by allowing one of its cooperating witnesses to “‘mingle’ with the other defendants prior to trial,” “to sit at the defense table during pre-trial court conferences,” and to “eat lunch with the defendants,” all before the defendant knew that this person would be called as the prosecution’s witness. 758 F.2d at 832. The Second Circuit affirmed the denial. Id. at 833. Expounding on the Sixth Amendment right to effective assistance of counsel and Weatherford, the Second Circuit held that “to require a hearing on a claimed sixth amendment violation resulting from unintentional or justifiable presence of a government informant or agent at an

63

attorney-client conference, a defendant must allege specific facts that indicate communication of privileged information to the prosecutor and prejudice resulting therefrom.” Id. (emphasis added). The court gave examples of the sort of evidence that would demonstrate “resulting” prejudice, which included “[a]llegations that a prosecution witness testified concerning privileged communications, that prosecution evidence originated in such communications, or that such communications have been used in any other way to the substantial detriment of the defendant.” Id.

All of these examples focus on the use of the overheard communications against the defendant at trial. For this reason, Hohn’s distinguishing Ginsberg because the case involved an “unintentional” and “justified” intrusion is unavailing. Suppl. Reply Br. at 6. Yes, the intrusion in Ginsberg was justified to protect the witness’s safety, but that difference had no effect on the court’s ultimate reasoning that the Sixth Amendment prejudice inquiry requires the defendant to show privileged communications were used “to [his] substantial detriment.” 758 F.2d at 833; see id. (confirming that prejudice was something “Ginsberg would need to establish” at an evidentiary hearing (cleaned up)). Hohn concedes that the intercepted CoreCivic call was not used against him at trial in any way, let alone to his substantial detriment. Besides, since Ginsberg, the Second Circuit has reaffirmed that establishing a Sixth Amendment violation requires that the “privileged information was passed to the Government” and “that prejudice to [the] defense resulted.” United States v.

64

Simels, 654 F.3d 161, 168 (2d Cir. 2011); accord Chandler, 56 F.4th at 40 (emphasizing that, even if privileged information had been passed to the prosecution, “a valid Sixth Amendment claim under Weatherford” would accrue only “if prejudice were shown”).

Likewise, Hohn writes off the Eleventh Circuit case, United States v.

Esformes, as one that dealt solely with determining the remedy for a governmental intrusion, not the structural-error question. But Esformes did touch on structural error when the Eleventh Circuit rejected the defendant’s argument that the court “should presume prejudice” under the Ninth Circuit’s burden-shifting approach. 60 F.4th at 633 (citing Danielson, 325 F.3d at 1072). The Eleventh Circuit rebuked what it dubbed a “novel approach” from the Ninth Circuit as being “foreclosed by [Eleventh Circuit] precedent.” Id. The foreclosing case the court referred to was United States v. Ofshe, where the defendant’s attorney worked as a government informant and recorded several of his meetings with the defendant that “included some unplanned discussions about his Florida case including the timing and likelihood of success on the motion to suppress.” 817 F.2d 1508, 1511 (11th Cir. 1987). In Ofshe, the Eleventh Circuit determined that the defendant had “suffered no prejudice” partly because the “taped conversation produced no tainted evidence.” Id. at 1515. So Esformes looked to Ofshe, along with a Fifth Circuit case, to affirm the Eleventh Circuit’s stance against presuming prejudice in Sixth Amendment

65

intrusion cases where the defendant shows no use of confidential conversations to generate evidence against him at trial. See Esformes, 60 F.4th at 633. 28 For his part, Hohn contends that D.C. Circuit caselaw favors his position because the D.C. Circuit recognized a “‘facially adequate’ Sixth Amendment claim when the prosecution intentionally intruded and became privy to strategic defense information.” Suppl. Reply Br. at 6 (quoting Kelly, 790 F.2d at 137– 39). But this argument misstates Kelly. In Kelly, the D.C. Circuit did not issue a merits-based Sixth Amendment ruling; rather, it held that the district court had abused its discretion by denying the defendant (a United States congressman) an evidentiary hearing on his motion for a new trial based on newly discovered evidence that the government had violated his Sixth Amendment rights during his criminal prosecution. 790 F.2d at 134. The evidence that the defendant presented showed the government had intentionally intruded into the defendant’s attorney-client meetings, stolen documents from defense counsel

28 The government also cites favorably Melvin, 650 F.2d at 643–44, but Hohn doesn’t address Melvin in his reply brief on the structural-error point. We agree with the government that Melvin too endorses Weatherford’s prejudice requirement. There, the Fifth Circuit affirmed its prior ruling in United States v. Sander, 615 F.2d 215 (5th Cir. 1980), where the court held that if the district court determined on remand that the defendant had been prejudiced by a police officer’s examining his attorney’s confidential files, then the court “must also determine whether some remedy short of dismissal . . . can be tailored to vindicate [the defendant’s] Sixth Amendment rights to counsel.” 650 F.2d at 644. So the Fifth Circuit agreed that to establish a Sixth Amendment violation and receive relief—any relief—a defendant must show “that the intrusion into [his] attorney-client relationship prejudiced the ability of [his] attorneys to provide adequate representation or otherwise prejudiced [his] defense.” Id.

66

(“including a witness list”) and passed those documents on to an FBI informant working with federal prosecutors. Id. at 132–33. The D.C. Circuit gathered that this evidence was “enough of a factual showing to merit further evidentiary development,” id. at 137, yet concluded that the defendant would still need to demonstrate at an evidentiary hearing “sufficient prejudice to establish a sixth amendment violation,” id. at 138. Thus in Kelly, even though the government intentionally intruded into attorney-client communications and relayed those communications to the prosecution, the court maintained the defendant’s need to show prejudice and indeed remanded for proceedings to allow him to do so. See id. So even under circumstances comparably egregious to those in Shillinger, the D.C. Circuit did not presume prejudice.

At best, Kelly exemplifies when a minimal showing of prejudice might entitle the defendant to an evidentiary hearing on the prejudice inquiry. See, e.g., id. at 137 (“While we cannot specify with certainty the quantum of prejudice Kelly must establish under Weatherford . . . , we are confident that he has made enough of a factual showing to merit further evidentiary development.”); Ginsberg, 758 F.2d at 833 (“To require a hearing on his sixth amendment claim, Ginsberg’s proffer would need to allege facts which, if proven,” would “establish that . . . the government intentionally invaded the attorney client relationship, and resulting prejudice” (cleaned up)); see also Morrison, 449 U.S. at 366 (acknowledging that in some cases there may be “justification for interfering with the criminal proceedings” if the defendant

67

makes some allegation or showing of prejudice). But in Hohn’s case, he stipulated that no prejudice resulted at his trial or at his sentencing, based on AUSA Morehead’s having heard the six-minute call. If an evidentiary hearing were held, Hohn has already conceded that he would have nothing to show. So we see no reason to grant him that relief. But that conclusion is limited to Hohn; given different facts, other § 2255 litigants might be entitled to an evidentiary hearing, which is a determination we leave to the district court.

The only circuit authority in concert with Hohn’s argument and Shillinger’s structural-error rule is United States v. Levy, 577 F.2d 200 (3d Cir. 1978), a Third Circuit case decided after Weatherford but before Morrison. In Levy, the Third Circuit reversed the district court’s decision to deny a defendant’s § 2255 petition where he alleged a Sixth Amendment violation because confidential information shared with his attorney had been relayed to the government through an undercover informant (the defendant’s nephew and former coconspirator). Id. at 202, 204–05, 207. Reviewing Third Circuit authority on the Sixth Amendment issue, the court reiterated its standard that “prejudice will be presumed if the informer transmits information on defense strategy to the government.” Id. at 207–08. Levy applied this rule despite the Supreme Court’s then-recent disposition in Weatherford, which the Third Circuit addressed. See id. at 209. Like Shillinger, Levy concluded that Weatherford was distinguishable because the informer in Weatherford had not relayed any intercepted trial strategy to the government. See id. at 209–10. The

68

Third Circuit understood Weatherford to “suggest[] by negative inference” that “where, as here, defense strategy was actually disclosed or where, as here, the government enforcement officials sought such confidential information,” a per se prejudice rule is appropriate. Id. at 210. This reflects Shillinger’s same reasoning. See 70 F.3d at 1140–41 (declaring a circuit split on the structural- error question and citing Levy, 577 F.2d at 210, as being on the supportive side). Hohn too leans on Levy throughout his supplemental briefing as a ballast to Shillinger.

But the Third Circuit has since rolled back Levy’s interpretation of Weatherford in light of the Supreme Court’s later decision in Morrison—issued three years after Levy. See United States v. Mitan, 499 F. App’x 187, 192 n.6 (3d Cir. 2012) (unpublished). With the benefit of Morrison and, more specifically, Morrison’s affirmative statement that a Sixth Amendment violation requires some showing of prejudice or adverse impact on the trial, see 449 U.S. at 365, the Third Circuit’s most recent discussion of Levy acknowledges that Levy may no longer be “viable,” Mitan, 499 F. App’x at 192. Though that decision did not overturn Levy, the Third Circuit noted that the facts in the case denied the court occasion to do so. See id. at 192 n.6 (opting not to “address the question of whether Morrison precludes the presumption of prejudice approach adopted in Levy” because the defendant could not show the government intentionally intruded into his attorney-client relationship). So

69

Levy may remain, but it stands on shaky ground and is certainly not enough to balance the scales of circuit authority weighing against it.

Contrary to Hohn’s assertions, most of the federal circuit caselaw that discusses Sixth Amendment intrusion claims bolsters our conclusion that, regardless of the circumstances underlying the government’s intrusion— intentional or unintentional, justified or unjustified, communicated or uncommunicated—the defendant cannot escape the second component of a Sixth Amendment intrusion violation: prejudice.

2. We disagree with the minority of circuits that construe prejudice as a rebuttable presumption in the defendant’s favor.

The First and Ninth Circuits agree that Weatherford holds “mere government intrusion into the attorney-client relationship . . . is not itself violative of the Sixth Amendment right to counsel,” unless “the intrusion substantially prejudices the defendant.” United States v. Irwin, 612 F.2d 1182, 1186–87 (9th Cir. 1980); accord Mastroianni, 749 F.2d at 907 (“A Sixth Amendment violation cannot be established without a showing that there is a ‘realistic possibility of injury’ to defendants or ‘benefit to the State’ as a result of the government’s intrusion into the attorney-client relationship” (quoting Weatherford, 429 U.S. at 558)). But they take a different tack when it comes to the prejudice inquiry.

The First and Ninth Circuits hold that prejudice should be assessed under a rebuttable presumption in the defendant’s favor, thus putting the onus on the

70

government to disprove any prejudicial effect from its actions. See Mastroianni, 749 F.2d at 907–08 (determining that a rebuttable presumption against the government balances the “competing concerns” that, on the one hand, it is “virtually impossible” for defendants to prove prejudice, and that, on the other hand, “there are certain circumstances in which the revelation of confidential communications by [an] informant is harmless” (quoting Briggs v. Goodwin, 698 F.2d 486, 494–95 (D.C. Cir. 1983)); Danielson, 325 F.3d at 1070–71 (recognizing that a defendant’s task to show prejudice presents “practical problems” because “[t]he prosecution team knows what it did and why” whereas “[t]he defendant can only guess”). This approach is driven by the circuit courts’ acknowledgement that governmental intrusions into the attorney- client relationship “pose a serious risk to [a] defendant[’s] constitutional rights,” and yet proving prejudice is “unreasonably difficult for most defendants.” United States v. DeCologero, 530 F.3d 36, 64 (1st Cir. 2008). So to alleviate the defendant’s burden, these circuits apply a rebuttable- presumption framework that “require[s] defendants to make a prima facie showing of prejudice by ‘proving that confidential communications were conveyed as a result’ of the government intrusion into the attorney-client relationship” and then shifts the burden to the government “to show that the defendant was not prejudiced.” Id. (quoting Mastroianni, 749 F.2d at 907–08); cf. Danielson, 325 F.3d at 1071 (adopting the “Mastroianni approach” from the First Circuit with the slight modification that a prima facie showing of

71

prejudice requires the defendant to show the government agent intentionally intruded into the attorney-client relationship to obtain confidential communications).

We decline to join this school because we find the rebuttable-

presumption framework incompatible with binding Supreme Court precedent. In Weatherford, the Supreme Court openly envisioned Bursey as the one who would bear the burden of showing prejudice when it stated that, even if Weatherford had communicated what he learned from the pretrial meetings to the prosecution, “Bursey would have a much stronger case,” 429 U.S. at 554 (emphasis added), not that the government would have a much weaker one. Similarly, in Morrison, the Court denied relief for the alleged Sixth Amendment violation because “respondent ha[d] demonstrated no prejudice.” 449 U.S. at 366 (emphasis added). There again, the Court put the defendant in the driver’s seat. Without any authority from the Supreme Court to suggest otherwise, we take its statements from Weatherford and Morrison to mean that defendants carry the burden under the Sixth Amendment prejudice inquiry. Cf. Cronic, 466 U.S. at 658 (noting that for effective-assistance claims “the burden rests on the accused to demonstrate a constitutional violation”). Not to mention, the circuits that have followed the rebuttable-presumption approach have done so, admittedly, under a dearth of authority from the Supreme Court. See Cinelli v. City of Revere, 820 F.2d 474, 478 (1st Cir. 1987) (contextualizing the burden-shifting arrangement with the observation that “the Supreme Court ha[s]

72

not had occasion to determine what showing of prejudice . . . is required to establish a sixth amendment violation and who bears the burden of proving it”); Danielson, 325 F.3d at 1069–70 (“[I]t is not clear from our precedents what constitutes ‘substantial prejudice’ and who bears the burden of proof . . . .”); cf. Kauer v. Maryland, 141 S. Ct. 5, 6 (2020) (mem.) (“Since Weatherford, many federal and state courts have struggled to define what burden, if any, a defendant must meet to demonstrate prejudice from a prosecutor’s wrongful or negligent acquisition of privileged information.”). And even Hohn does not ask us to follow the First and Ninth Circuits’ burden-shifting formulation. So we see no reason to adopt this “novel approach.” Esformes, 60 F.4th at 633.

CONCLUSION

Hohn’s appeal puts Shillinger squarely under the microscope and, upon closer examination, we cannot help but see its flaws. A more exacting review throws Shillinger’s misreading of Supreme Court precedents into stark relief. And given that Hohn’s claim rests entirely on the presumption of prejudice permitted by Shillinger’s structural-error rule, we cannot faithfully resolve his appeal without considering whether Shillinger still stands on solid footing. We believe it does not, and so we hold that a Sixth Amendment violation of the right to confidential communication with an attorney requires the defendant to show prejudice.

We affirm the district court’s denial of Hohn’s § 2255 petition on that alternate ground.

No. 22-3009, United States of America v. Steven M. Hohn

BACHARACH, joined by McHUGH and ROSSMAN, Circuit Judges, dissenting only as to Part II(C)(2).

This case grew out of a prosecutor’s intentional and unjustified intrusion into attorney-client communications about legal strategy. We earlier held that this kind of intrusion creates a conclusive presumption of prejudice. Shillinger v. Haworth, 70 F.3d 1132, 1142 (10th Cir. 1995). The majority abrogates this holding. Maj. Op. at 1–69. So we must decide how to gauge prejudice in the future. Do we treat the intrusion into attorney- client communications about legal strategy like most other elements of post-conviction relief, putting the burdens of production and persuasion on the defendant? Or should we recognize the unique factors bearing on the defendant’s inability to show how the prosecutor may have used the intercepted information?

The First and Ninth Circuits have zeroed in on these unique factors, creating a rebuttable presumption of prejudice when the defendant proves an intentional, unjustified intrusion by the prosecution into attorney-client communications about legal strategy. United States v. Mastroianni, 749 F.2d 900, 907–08 (1st Cir. 1984); United States v. Danielson, 325 F.3d 1054, 1073–74 (9th Cir. 2003), as amended (May 19, 2003). This approach enhances fairness because the prosecution’s misconduct typically yields superior access to information about potential prejudice.

1. The defendant should bear the threshold burden to show a prima facie case.

The defendant should bear the burden to show an intentional, unjustified intrusion into attorney-client communications about legal strategy.

We have applied the Sixth Amendment to protect the defendant from “a prosecutor’s intentional intrusion into the attorney-client relationship . . . absent a countervailing state interest.” Shillinger v. Haworth, 70 F.3d 1132, 1142 (10th Cir. 1995). For this kind of intrusion, most circuits recognize that the defendant bears the initial burden. For example, the First and Ninth Circuits create a rebuttable presumption of prejudice only upon the defendant’s initial showing of an improper intrusion. United States v. Danielson, 325 F.3d 1054, 1071–72 (9th Cir. 2003), as amended (May 19, 2003); United States v. Mastroianni, 749 F.2d 900, 907–08 (1st Cir. 1984). The Third, Fifth, and Eighth Circuits require the defendant to show an actual disclosure of attorney-client communications. United States v. Levy, 577 F.2d 200, 209 (3d Cir. 1978); United States v. Melvin, 650 F.2d 641, 645–46 (5th Cir. Unit B July 1981); United States v. Hari, 67 F.4th 903, 912–13 (8th Cir. 2023). And the Sixth and Seventh Circuits suggest that the defendant must show that the prosecution listened to attorney-client communications. United States v.

Steele, 727 F.2d 580, 586–87 (6th Cir. 1984); United States v. Castor, 937 F.2d 293, 297–98 (7th Cir. 1991).

Like these circuits, we should recognize the defendant’s threshold burden to show a prima facie case. See Maj. Op. at 14–15. That showing requires the defendant to prove two elements:

1. The prosecution’s intrusion was intentional and unjustified.

2. This intrusion resulted in the prosecution’s interception of attorney-client communications about the defendant’s legal strategy.

Mr. Hohn made that showing. The district court concluded that the prosecution had intentionally intruded into the attorney-client relationship by listening to Mr. Hohn’s phone call with his attorney. No issue of justification existed, for the government didn’t argue that the prosecution had a legitimate reason to listen to the call. And the attorney-client communication itself had related to legal strategy, including • Mr. Hohn’s desire to proceed to trial, • his criminal history, • the evidence he expected to face, • the flaws in the evidence, and

• how he and his attorney would meet and discuss the case moving forward. 1

1 The attorneys in the appeal haven’t heard the recorded phone call, and it isn’t in our record. But the district court made a factual finding about the contents of the call. Given the unavailability of the recording, I would rely on the district court’s findings about the call.

Maj. Op. at 7–8. So Mr. Hohn satisfied his prima facie burden to show the prosecution’s intentional, unjustified interception of attorney-client communications about legal strategy. 2. The burden should shift to the prosecution.

Given Mr. Hohn’s showing, the prosecution should bear the burden of negating the potential prejudice.

“The burden-shifting principle is not new or novel,” Keyes v. Sch.

Dist. No. 1, Denver, 413 U.S. 189, 209 (1973), as courts often shift the burden of proof based on factors such as a party’s superior access to evidence, Int’l Bhd. of Teamsters v. United States, 431 U.S. 324, 359 n.45 (1977). Adopting the burden-shifting principle makes sense here for two reasons:

1. The prosecution typically knows whether and how the communications affected the trial, while the defendant can only speculate.

2. It’s fair to place the burden on the prosecution when it acted wrongfully by intruding into attorney-client communications.

a. The burden may shift based on access to information and principles of fairness.

The Sixth Amendment is violated only when the intrusion is prejudicial. Maj. Op. at 20. A violation is prejudicial only when it creates “a ‘realistic possibility of injury’ to [the] defendants or ‘benefit to the

State.’” United States v. Mastroianni, 749 F.2d 900, 907 (1st Cir. 1984) (quoting Weatherford v. Bursey, 429 U.S. 545, 558 (1977)).

The question is who should bear the burden of proving that possibility. We have flexibility in answering because “[t]here are no hard- and-fast standards governing the allocation of the burden of proof in every situation.” Keyes v. Sch. Dist. No. 1, Denver, 413 U.S. 189, 209 (1973). In the absence of hard-and-fast standards, we consider various factors. Alaska Dep’t of Env’t Conservation v. E.P.A., 540 U.S. 461, 494 n.17 (2004). These factors include

• relative access to “peculiar means of knowledge,” Alaska Dep’t of Env’t Conservation, 540 U.S. at 494 n.17 (internal quotation marks omitted), and

• “question[s] of policy and fairness,” Keyes, 413 U.S. at 209 (internal quotation marks omitted).

We consider these factors against the backdrop of our own “experience.” Denning Warehouse Co. v. Widener, 172 F.2d 910, 913 (10th Cir. 1949); see Keyes, 413 U.S. at 209 (stating that allocation of the burden of proof is “a question of policy and fairness based on experience in the different situations” (quoting 9 J. Wigmore, Evidence § 2486, at 275 (3d ed. 1940))); see also Fleming James, Jr., Burdens of Proof, 47 V A . L. R EV . 51, 58 (1961) (stating that the burden of proof is allocated “on the basis of one or more of several variable factors”).

b. The prosecution is typically the only party that knows whether and how the communications affected the trial.

The information is generally asymmetrical because the prosecution typically knows what it decided, when it made the decision, and why it made that decision. Unlike the prosecution, “[t]he defendant can only guess.” United States v. Danielson, 325 F.3d 1054, 1070 (9th Cir. 2003), as amended (May 19, 2003). Given the asymmetry, the prosecution should bear the burden of negating prejudice. See Alaska Dep’t of Env’t Conservation v. E.P.A., 540 U.S. 461, 494 n.17 (2004) (stating that the burdens of production and persuasion may be placed on the party with superior access to information) ; see also United States v. N.Y., New Haven & Hartford R.R. Co., 355 U.S. 253, 256 n.5 (1957) (“The ordinary rule, based on considerations of fairness, does not place the burden upon a litigant of establishing facts peculiarly within the knowledge of his adversary.”); Elizabeth Bartholet, Proof of Discriminatory Intent Under Title VII: United States Postal Service Board of Governors v. Aikens, 70 C ALIF . L. R EV . 1201, 1211 n.43 (1982) (“Access to evidence is one of the key considerations determining who bears the burden of proof in general.”).

Similar circumstances exist in cases of securities fraud, where shareholders are not privy to the same information as corporate insiders. So when a publicly traded corporation makes material misstatements, the

Supreme Court presumes prejudice to shareholders. Goldman Sachs Grp., Inc. v. Ark. Tchr. Ret. Sys., 594 U.S. 113, 124–27 (2021); Basic Inc. v. Levinson, 485 U.S. 224, 245–27 (1988). But this presumption is rebuttable, for the defendant gets a chance to show that the misrepresentation didn’t distort the share price. Basic, 485 U.S. at 248.

The Supreme Court explained that this allocation of the burden makes sense because we can’t ordinarily expect a shareholder to have proof of prejudice. Id. at 245. The same problem exists for someone like Mr. Hohn, whose attorney-client communications have been intercepted without any way to know how the prosecution may have used the information. See, e.g., United States v. Danielson, 325 F.3d 1054, 1070 (9th Cir. 2003), as amended (May 19, 2003) (“[I]t will often be unclear whether, and how, the prosecution’s improperly obtained information about the defendant’s trial strategy may have been used, and whether there was prejudice.”). The rebuttable presumption provides a solution, just as it does in cases of securities fraud.

The informational advantage is magnified when the prosecution learns about the legal strategy of a criminal defendant. For example, consider how the defendant could show prejudice when the prosecution improperly intercepts attorney-client communications about whether to call the defendant as a witness. The intrusion could prejudice the defendant in plea bargaining, jury selection, or the prosecution’s case-in-chief. But how

8

could the defendant know if the prosecution had used the information for these purposes? The defendant has no way of knowing .

The majority says that Mr. Hohn “stipulated” that the prosecution hadn’t used the intercepted information. Maj. Op. at 11, 67. The majority is mistaken: Mr. Hohn never stipulated or admitted that the prosecution hadn’t used the intercepted information. 2 The only pertinent stipulation was this: “Mr. Hohn does not assert that he can prove that he suffered any actual–as opposed to presumptive–prejudice due to the prosecution’s becoming privy to the one attorney-client call listed in his privilege log.” Supp. R. vol. 2, at 143. There Mr. Hohn admitted only that he couldn’t prove prejudice.

Mr. Hohn presumably couldn’t prove prejudice because the pipeline for intercepted information about legal strategy had flowed only one way: The prosecution knew Mr. Hohn’s legal strategy, including what he believed would be the incriminating evidence and how to attack that

2 If Mr. Hohn had stipulated that there wasn’t any prejudice, he presumably would have waived the Sixth Amendment claim with or without a conclusive presumption. See United States v. Kieffer, 681 F.3d 1143, 1158 (10th Cir. 2012) (stating that under an analysis of structural error, there must be an error or defect that hasn’t been affirmatively waived). So if the majority were right about the alleged stipulation, the Court would have had no reason to convene en banc or to address the continued viability of Shillinger.

evidence. Maj. Op. at 7–8. But Mr. Hohn had no way of knowing whether the prosecution had previously • planned to use that evidence or • known how Mr. Hohn was going to attack that evidence.

The one-way pipeline for information made it virtually impossible for Mr. Hohn to know whether the prosecution had used the improperly intercepted information.

We use a burden-shifting test in many similar situations. See, e.g., Int’l Bhd. of Teamsters v. United States, 431 U.S. 324, 359 n.45 (1977) (justifying a burden-shifting test because the defendant “knew best what th[e relevant] factors were and the extent to which they influenced the decision-making process”). In these situations, a shift in the burden could relieve the innocent party of a need to guess about the impact.

For example, consider cases involving employment discrimination through disparate impact or disparate treatment. In these cases, the plaintiff must make a prima facie showing that creates an inference of employment discrimination. The burden then shifts to the defendant to show a business necessity or a legitimate, nondiscriminatory reason for the employment decision. See Thomas v. Metroflight, Inc., 814 F.2d 1506, 1509 (10th Cir. 1987) (business necessity); Young v. United Parcel Serv., Inc., 575 U.S. 206, 213 (2015) (legitimate, nondiscriminatory reason). The burden shifts to the defendant in order “to frame the factual issues with

10

sufficient clarity” for the plaintiff to get “a full and fair opportunity to demonstrate pretext.” Tex. Dep’t of Cmty. Affairs v. Burdine, 450 U.S. 248, 255–56 (1981).

We also shift the burden in criminal cases. For example, the burden shifts to the prosecution when a defendant alleges a racial motivation for peremptory challenges. Batson v. Kentucky, 476 U.S. 79, 97–98 (1986); Johnson v. Martin, 3 F.4th 1210, 1219 (10th Cir. 2021). The shift in the burden makes sense because only the prosecution knows why it struck particular jurors. See Hill v. Texas, 316 U.S. 400, 405 (1942) (explaining why the burden shifts to the prosecution in challenges involving racial bias in jury selection).

A shift in the burden is equally sensible here. Our inquiry turns on the existence of prejudice, and the prosecution is typically the only party that could possibly know whether it made decisions based on the intercepted information.

c. It’s fair to place the burden on the prosecution when the asymmetry of information resulted from prosecutorial misconduct.

When the asymmetry of information results from prosecutorial misconduct, a shift in the burden is particularly appropriate. Here we are addressing allocation of the burden only when the prosecution’s intrusion is intentional and unjustified. So the issue arises only when the prosecution • created the problem,

11

• could have prevented the problem, and • could have redressed the problem earlier.

See United States v. Danielson, 325 F.3d 1054, 1072 (9th Cir. 2003), as amended (May 19, 2003) (“[T]he prosecution team can avoid this burden either by not improperly intruding into the attorney-client relationship in the first place, or by insulating itself from privileged trial strategy information that might thereby be obtained.”). The prosecution should bear the burden when it created the problem. See Keyes v. Sch. Dist. No. 1, Denver, 413 U.S. 189, 209 (1973) (stating that allocation of the burden of proof is “a question of policy and fairness” (quoting 9 J. Wigmore, Evidence § 2486, at 275 (3d ed. 1940))). “[T]o require anything less would be to condone intrusions into a defendant’s protected attorney-client communications.” United States v. Mastroianni, 749 F.2d 900, 908 (1st Cir. 1984).

Our facts illustrate the fairness of putting the burden on the prosecution. The district court learned that the prosecution had “harbored multiple copies of [the] recorded calls” and ordered their disclosure. R. vol. 2, at 1757. But the prosecution refused to comply. Id. at 1756. Given this refusal, the district court explained not only how the prosecution had possessed and listened to Mr. Hohn’s attorney-client call, but also how the prosecution had taken “steps to conceal that tactical advantage,” “minimiz[ing], deflect[ing] and obfuscat[ing the

12

prosecution’s] role.” Id. at 1776–77; see id. at 1777–79. The district court thus disbelieved the prosecution’s contrary testimony. Id. at 1777–79.

The prosecution’s misconduct “raise[s] a substantial and serious question about the fundamental fairness of the process.” Bank of Nova Scotia v. United States, 487 U.S. 250, 259 (1988). Given that misconduct, it’s hardly fair to require the defendant to show why the prosecution made its strategic decisions. That burden belongs with the prosecution when it was the wrongdoer. See Int’l Bhd. of Teamsters v. United States, 431 U.S. 324, 359 n.45 (1977) (stating that the burden of persuasion shifts because the existence of a prima face showing “changed the position of the [defendant] to that of a proved wrongdoer”).

3. The competing interests are properly balanced through a shift in the burden.

Allocation of the burden should “best account[] for the competing interests at stake.” Shillinger v. Haworth, 70 F.3d 1132, 1141 (10th Cir. 1995); see United States v. Wilson, 17 F.4th 994, 1004 (10th Cir. 2021) (adopting a burden-shifting test based on “competing considerations on both sides”). Here, for example, we must reconcile the competing interests involving the necessity of prejudice and the potential subversion of justice.

On one hand, a Sixth Amendment violation is not complete until there is prejudice. Maj. Op. at 20; see Weatherford v. Bursey, 429 U.S. 545, 558 (1977). On the other hand, the prosecution’s intrusion into the

13

defendant’s attorney-client relationship and communications about legal strategy “threaten[] to subvert the adversary system of criminal justice.” Id. at 556; see United States v. Levy, 577 F.2d 200, 209 (3d Cir. 1978) (“In order for the adversary system to function properly, any advice received as a result of a defendant’s disclosure to counsel must be insulated from the government.”). This threat exists partly because the fear of eavesdropping can chill a defendant’s willingness to freely communicate with counsel. Weatherford, 429 U.S. at 554 n.4.

We can properly balance these interests through a rebuttable presumption of prejudice. To see this balance, consider what happens when the prosecution intercepts a defendant’s phone call with attorneys about their plans to impeach a government witness. Interception of the call might or might not prejudice the defendant. For example, if the prosecution had already decided not to call the witness, the interception might not be prejudicial. But other times, the interception might be prejudicial. For example, knowledge of the defense strategy might lead the prosecution to elicit testimony about impeachment material to soften the sting of later cross-examination. Or a brief call might disclose information about the attorneys’ tone or approach. In each circumstance, however, the prosecution is the only party that knows whether it used the improperly intercepted information against an unknowing defendant.

Other courts have taken various approaches. On one end of the spectrum, the Third Circuit has held that the defendant’s prima facie case triggers a conclusive presumption of prejudice. United States v. Levy, 577 F.2d 200, 210 (3d Cir. 1978). On the other end, the Fifth Circuit has suggested that the defendant must show prejudice stemming from the intrusion. United States v. Melvin, 650 F.2d 641, 644 (5th Cir. Unit B July 1981). And in the middle, the First and Ninth Circuits require the government to show the absence of prejudice. United States v. Mastroianni, 749 F.2d 900, 907–08 (1st Cir. 1984); United States v. Danielson, 325 F.3d 1054, 1073–74 (9th Cir. 2003), as amended (May 19, 2003). Like the First, Third, and Ninth Circuits, we should relieve the defendant of the threshold duty to show prejudice. And now that the majority has rejected a conclusive presumption, we should join the First and Ninth Circuits in recognizing a rebuttable presumption and allowing the government a chance to rebut that presumption. 3

3 Scholars similarly conclude that the burden should fall on the government, not the defendant. See, e.g., Loretta A. Neary-West, Right to Counsel: Balancing the Burden of Persuasion on the Adversarial Scales of Criminal Justice, 26 S AN D IEGO L. R EV . 1145, 1160–67 (1989) (urging a theory of allocating the burden on the prosecution (citing C. M C C ORMICK , M C C ORMICK ON E VIDENCE 952 (E. Cleary 3d ed. 1984))); Blake R. Hills, Unsettled Weather: The Need for Clear Rules Governing Intrusion Into Attorney-Client Communications, 50 N.M. L. R EV . 135, 160–61 (2020) (contending that an intentional, unjustified intrusion should trigger a rebuttable presumption, requiring the government “to prove that it has not used the confidential information to prejudice the defendant or benefit

15

4. The burden need not shift in the absence of an intentional, unjustified interception of the defendant’s discussion with counsel about legal strategy.

The Sixth Amendment may be implicated in other circumstances, including when the intrusions don’t uncover legal strategy or involve intentional eavesdropping of communications between defendants and their attorneys. For example, the majority points to cases in other circuits involving

• no showing of intentional eavesdropping, United States v.

Collins, 799 F.3d 554, 591–92 (6th Cir. 2015); United States v.

Ginsberg, 758 F.2d 823, 832–33 (2d Cir. 1985), no information being transmitted to prosecutors, Williams v. Woodford, 384 F.3d 567, 585 (9th Cir. 2004); United States v. Castor, 937 F.2d 293, 297–98 (7th Cir. 1991), or no governmental

itself in any manner”). One scholar identifies five reasons for putting the burden on the government:

1. The government is the party seeking a departure from the status quo because a deliberate, unjustified intrusion is a constitutional violation. Neary-West, supra, at 1161–62.

2. “Violation of procedural safeguards specifically designed to protect against trial prejudice renders the claim of no prejudice more unusual than a claim of prejudice.” Id. at 1162.

3. Only the government has “knowledge of the relevant facts.” Id.

at 1163.

4. Without access to the relevant facts, the defendant can’t typically prove prejudice, facilitating—rather than deterring—

prosecutorial intrusions into the Sixth Amendment. Id. at 1164.

5. Placing the burden on the government best balances the defendant’s constitutional right with society’s interest in the effective administration of criminal justice. Id. at 1164–67.

intrusion, United States v. Steele, 727 F.2d 580, 586 (6th Cir.

1984);

• intercepted conversations that had been suppressed before trial, United States v. Esformes, 60 F.4th 621, 629, 633 (11th Cir.

2023);

• accidental receipt of attorney-client information, which hadn’t been seen by the prosecution, United States v. Hari, 67 F.4th 903, 911–13 (8th Cir. 2023); and

• court-ordered disclosure of the defense attorney’s cross-

examination plans, with a stipulation that the plans not be shared with the cross-examining prosecutor, United States v.

Allen, 491 F.3d 178, 192 (4th Cir. 2007).

These intrusions don’t involve intentional, unjustified eavesdropping into legal strategy. 4 As a result, these intrusions don’t involve the government’s ability to benefit from its wrongdoing or an asymmetry of information. So these intrusions might not require a court to put the burden of persuasion on the government. But here, allocation of the burden is justified by an asymmetry of information resulting from prosecutorial misconduct.

4 The case law contains two exceptions.

The first appears in United States v. Kelly, 790 F.2d 130, 137–38 (D.C. Cir. 1986). Kelly did involve the improper interception of legal strategy, but the court declined to address the need to show prejudice. Instead, the court noted uncertainty about the standard for prejudice and remanded for an evidentiary inquiry. Id. at 137–38.

The second exception appears in United States v. Melvin, 650 F.2d 641, 643–44 (5th Cir. Unit B July 1981). There the court simply remanded “for further findings of fact on the question of prejudice.” Id. at 644.

For example, the majority points out that the Seventh Circuit has stated that the defendant must “show prejudice.” Maj. Op. at 60 (quoting United States v. Castor, 937 F.2d 293, 297 (7th Cir. 1991)). There the defendants alleged a personal relationship between their own investigator and an investigator for the government. Castor, 937 F.3d 297–98. So this case didn’t involve prosecutorial misconduct or interception of communications about legal strategy.

We need not explore allocation of the burden in that case or the others discussed in the majority opinion. None involve allocation of the burden for an intentional, unjustified intrusion into communications between a defendant and counsel about legal strategy. And it’s this unique context that triggers the need to shift the burden because of the prosecution’s superior access to information acquired through improper conduct. See Part 1, above.

In this context, the only circuits to address the allocation of the burden are the First and Ninth Circuits. Both adopt a rebuttable presumption of prejudice. United States v. Mastroianni, 749 F.2d 900, 907–08 (1st Cir. 1984); United States v. Danielson, 325 F.3d 1054, 1073– 74 (9th Cir. 2003), as amended (May 19, 2003). The majority creates a circuit split, 5 scuttling a rebuttable presumption adopted in the only two

5 Perhaps in part for this reason, some of the cited cases decline to state who bears the burden to show prejudice, relying on the passive voice

18

circuits to address the issue in cases of intentional intrusions into attorney- client communications.

5. Recognition of a rebuttable presumption wouldn’t violate United States v. Morrison or Weatherford v. Bursey.

The Supreme Court hasn’t said anything inconsistent with a rebuttable presumption.

a. United States v. Morrison didn’t involve interception of legal strategy or eavesdropping on attorney-client communications.

The majority says that the Supreme Court’s opinion in United States v. Morrison, 449 U.S. 361 (1981), is “incompatible” with a rebuttable presumption. Maj. Op. at 71. But the Morrison Court didn’t address intrusion into legal strategy or circumstances creating an asymmetry of information bearing on prejudice.

In Morrison, a criminal defendant hired counsel to defend against an indictment for heroin distribution. 449 U.S. at 362. Two federal agents tried to obtain the defendant’s cooperation in a related investigation. Id. The agents knew that the defendant had been indicted and had hired counsel. Despite that knowledge, the agents met with the defendant without informing her counsel. Id. In the meeting, the agents

or other indeterminate language. See, e.g., Maj. Op. at 59–60 (“[S]ome showing of prejudice is a necessary element.” (quoting United States v. Allen, 491 F.3d 178, 192 (4th Cir. 2007))), 61 (“[P]rejudice to the defendant must be shown.” (quoting United States v. Steele, 727 F.2d 580, 586 (6th Cir. 1984))).

• disparaged defense counsel,

• suggested that the defendant seek representation by the public defender, and

• discussed the benefits and drawbacks of cooperation.

Id. The defendant declined and notified her attorney. Id. “[A]t no time did [she] agree to cooperate with them, incriminate herself, or supply any information pertinent to her case.” Id. at 362–63.

Morrison didn’t involve an intrusion into attorney-client communications or an asymmetry of information from prosecutorial misconduct. To the contrary, the defendant knew what the federal agents had said and how the conversations would affect her decision-making. So she was on equal footing with the government in the ability to prove prejudice.

The majority points out that Morrison put the burden on the defendant. Maj. Op. at 71. But Morrison didn’t address allocation of the burden when the Sixth Amendment violation comes from prosecutorial misconduct or asymmetry of information bearing on prejudice.

b. Weatherford v. Bursey didn’t discuss the burden of proof for Sixth Amendment violations.

The majority also says that the Supreme Court’s opinion in Weatherford v. Bursey, 429 U.S. 545 (1977), prevents recognition of a rebuttable presumption. Maj. Op. at 71. In Weatherford, the Supreme Court had no occasion to allocate the burden of production or persuasion for a

Sixth Amendment violation. See Weatherford, 429 U.S. at 558 (“There being no tainted evidence in this case, no communication of defense strategy to the prosecution, and no purposeful intrusion by [the prosecution], there was no violation of the Sixth Amendment.”). The majority nonetheless seizes on the Weatherford Court’s reference to the defendant’s case, suggesting that this word choice must have shown an intent to require the defendant to prove prejudice. Maj. Op. at 71 (quoting Weatherford, 429 U.S. at 554).

In Weatherford, the defendant communicated with his attorney in the presence of a codefendant. 429 U.S. at 547–48. The defendant didn’t know that the codefendant was actually an undercover law-enforcement officer. Id. Despite his undercover status, the codefendant never told the prosecution what he had learned in the defendant’s meeting with counsel. Id. at 548. The Supreme Court pointed out that the defendant “would have a much stronger case” if the undercover officer had told the prosecution about what the defendant and his attorney had said. Id. at 554. The majority apparently assumes that the Supreme Court must have been using the word case to imply that the defendant had the burden of persuasion. Maj. Op. at 71.

This assumption is questionable, for the term case typically means “[a] civil or criminal proceeding, action, suit, or controversy at law or in equity”—not a burden to prove prejudice or any other element. Case,

B LACK ’ S L AW D ICTIONARY (12th ed. 2024). Suppose, for example, that a court said that a civil plaintiff would have a stronger case if it weren’t barred by the statute of limitations. Would you think that the court regarded the statute of limitations as part of the plaintiff’s burden rather than an affirmative defense? See Fed. R. Civ. P. 8(c)(1) (treating the statute of limitations as an affirmative defense). Even if you would, Weatherford contained no suggestion—in either the briefing or the opinion itself—that allocation of the burden was at issue. And the Supreme Court doesn’t typically hide important legal propositions in mouseholes—like word choices in opinions involving different issues. See In re Permian Basin Area Rate Cases, 390 U.S. 747, 775 (1968) (“[T]his court does not decide important questions of law by cursory dicta inserted in unrelated cases.”).

Nor has the Supreme Court ever addressed allocation of the burden on prejudice when the prosecution intentionally and unjustifiably intrudes into attorney-client communications about legal strategy. Only two circuits have addressed the issue, and both have adopted a presumption of prejudice that gives the government an opportunity for rebuttal. United States v. Mastroianni, 749 F.2d 900, 907–08 (1st Cir. 1984); United States v. Danielson, 325 F.3d 1054, 1073–74 (9th Cir. 2003), as amended (May 19, 2003); see Part 4, above.

6. The government should be required to show that the intercepted legal strategy didn’t prejudice the defendant.

What should the government’s rebuttal entail? To answer, we can draw guidance from the Supreme Court’s treatment of the burden when a defendant claims that the prosecution compelled testimony in violation of the Fifth Amendment. In these cases, the U.S. Supreme Court shifts the burden to the government. Kastigar v. United States, 406 U.S. 441, 460–62 (1972). In shifting the burden, the Supreme Court reasoned in part that the Fifth Amendment protects witnesses against compelled self-incrimination. Id. at 444–45. But the Court acknowledged that the prosecution can compel a witness to testify by providing immunity. Id. at 449–50, 453.

But what if the government then indicts the witness on charges related to the compelled testimony? How do we assess whether the government had improperly based the indictment on the compelled testimony? After all, the witness would lack any way of showing compulsion of the testimony.

The Supreme Court has resolved this dilemma by shifting the burden to the government. Id. at 460–61. Through this allocation of the burden, witnesses must demonstrate that they testified under a grant of immunity on matters related to the prosecution. Id. That demonstration shifts the burden to the government to show an independent, legitimate source for the evidence. Id. That showing must do more than negate the taint; the

23

government must prove that its evidence “derived from a legitimate source wholly independent of the compelled testimony.” Id. at 460; see United States v. Lacey, 86 F.3d 956, 972 (10th Cir. 1996).

A similar approach is appropriate here, for the prosecution • created the problem through misconduct and

• thereby gained superior access to the relevant information bearing on prejudice.

If the prosecution could discharge its burden just by presenting some evidence, the defendant would generally have no way to show an effect on the trial. The presumption is meaningful only if the prosecution bears the ultimate burden to disprove prejudice from the intrusion into intercepted communications about legal strategy.

Application of this burden “will vary from case to case.” United States v. Danielson, 325 F.3d 1054, 1072 (9th Cir. 2003), as amended (May 19, 2003). Given the multitude of possible scenarios, we should avoid rigid formulas to specify what the prosecution needs to show in order to rebut a presumption of prejudice. The inquiry may vary depending on the timing of the intrusion, the scope of information revealed, the prosecution’s conduct, and other circumstances. Given the variety of possible circumstances, district courts occupy an ideal position to balance the appropriate factors on a case-by-case basis. In undertaking this balancing of factors, district courts should decide in the first instance

24

whether the prosecution satisfied its burden of persuading the factfinder that the intrusion hadn’t prejudiced the defendant. See id. at 1073–74. 6

7. We should give the prosecution a chance to rebut the presumption of prejudice here.

In district court, the parties were bound by our precedent recognizing a conclusive presumption of prejudice. See Shillinger v. Haworth, 70 F.3d 1132, 1142 (10th Cir. 1995). But the majority abrogates that precedent. In the absence of a conclusive presumption, we must decide whether to foist

6 Although the Ninth Circuit acknowledged the necessity of a case-bycase analysis, the court also provided a detailed description of the government’s burden:

[T]he government must introduce evidence and show by a preponderance of that evidence that it did not use this privileged information. Specifically, it must show that all of the evidence it introduced at trial was derived from independent sources, and that all of its pre-trial and trial strategy was based on independent sources. Strategy in this context is a broad term that includes, but is not limited to, such things as decisions about the scope and nature of the investigation, about what witnesses to call (and in what order), about what questions to ask (and in what order), about what lines of defense to anticipate in presenting the case in chief, and about what to save for possible rebuttal.

Danielson, 325 F.3d at 1074. This formulation may or may not be suitable in a given case. District courts should have discretion to choose whether to require a similar showing based on the particular facts.

an impossible burden on the defendant when the prosecution wrongfully gains a monopoly of the pertinent information on prejudice.

I wouldn’t do that. Instead, I think we should shift the burden to the government because Mr. Hohn has demonstrated an intentional, unjustified intrusion into attorney-client communications about legal strategy. 7 The burden should then shift to the government for rebuttal of that

7 The majority notes that Mr. Hohn urged us to continue applying a conclusive presumption rather than to make the presumption rebuttable. Maj. Op. at 72. The majority addresses allocation of the burden anyway, presumably because the issue arises from the parties’ disagreement on who must prove prejudice. Id. at 69–71. And “when an issue or claim is properly before the court, the court is not limited to the particular legal theories advanced by the parties, but rather retains the independent power to identify and apply the proper construction of governing law.” U.S. Nat’l Bank of Or. v. Indep. Ins. Agents of Am., Inc., 508 U.S. 439, 446 (1993) (quoting Kamen v. Kemper Fin. Servs., Inc., 500 U.S. 90, 99 (1991)). So we need not adopt either party’s position when we conclude that an appellant is entitled to some, but not all, of the requested relief:

The party-presentation principle, however, restricts courts from raising new issues. The principle does not say that once an issue has been raised and responded to, a court must render its decision in accordance with the position of one of the parties.

Courts have always had authority to resolve raised issues as fairness requires.

United States v. Cortez-Nieto, 43 F.4th 1034, 1052 (10th Cir. 2022) (emphasis in original; citation omitted); accord Novella v. Westchester Cnty., 661 F.3d 128, 147 (2d Cir. 2011) (adopting a third approach after rejecting the parties’ positions); Clark v. A&L Homecare & Training Ctr., LLC, 68 F.4th 1003, 1009–1011 (6th Cir. 2023) (adopting a middle ground after rejecting the approaches urged by both sides); United States v. Arnold, 238 F.3d 1153, 1155 (9th Cir. 2001) (adopting a third approach after rejecting the parties’ positions).

presumption. Until now, however, the government hasn’t had a chance to make that showing. A remand to district court is thus appropriate.

United States v. Hohn, No. 22-3009 ROSSMAN, joined by BACHARACH, Circuit Judges, dissenting.

For nearly three decades, it has been the law of this circuit that when the prosecution unjustifiably and intentionally becomes privy to confidential attorney-client communications, the Sixth Amendment is violated, and this rarely occurring constitutional error is so fundamental and pervasive that we will deem it prejudicial in every case. Shillinger v. Haworth, 70 F.3d 1132, 1142 (10th Cir. 1995). Today, the majority undoes Shillinger’s conclusive presumption and replaces it with a new rule requiring the defense to show “a realistic possibility of injury to the defendant or benefit to the government” to establish a Sixth Amendment prosecutorial-intrusion claim. Op. at 15; see Op. at 72. Not only is the majority opinion wrong about the law, it reflects a mistaken judgment about how the law should be enforced and justice administered. I disagree with the majority’s disposition and the analysis on which it depends. To explain my reasoning, I proceed in four parts.

First, I address some of the unusual aspects of this appeal. Second, I discuss why Shillinger was correctly decided and why we should have reaffirmed its conclusive presumption of prejudice. Third, I explain why the majority’s new rule is unworkable. Fourth, I reach the confidential- communications issue presented by the parties, conclude Shillinger did not

2

include a privilege element, and hold Sixth Amendment protections attached to Mr. Hohn’s confidential attorney-client call.

Mr. Hohn’s § 2255 motion should have been granted because a Sixth Amendment violation occurred when the prosecution purposefully and without justification became privy to his confidential legal communications with defense counsel. I would reverse the district court’s contrary conclusion and remand for a determination of the appropriate remedy. Because the majority decides otherwise, I respectfully dissent.

I

We must acknowledge at the outset this appeal is unusual. First, it stems from unprecedented transgressions by federal prosecutors into the defense function. “There is no template for this case,” the district court observed, “where the fairness of the adversary system is called into question by systemic prosecutorial misconduct of the type alleged here.” United States v. Carter, 429 F. Supp. 3d 788, 903 (D. Kan. 2019), order vacated in part, No. 16-20032-02- JAR, 2020 WL 430739 (D. Kan. Jan. 28, 2020). Second, the underlying habeas petition, premised on just one example of this pervasive misconduct, asserts a narrow and rare Sixth Amendment claim. Finally, the disposition upends longstanding circuit precedent using an uncommon procedure—sua sponte initial en banc review. See United States v. Hohn, 91 F.4th 1060, 1060 (10th Cir. 2024). I briefly discuss these features before addressing the merits.

A1

1

For an unknown number of years, the United States Attorney’s Office for the District of Kansas (USAO) undertook an undisclosed “systematic practice of purposeful collection, retention, and exploitation” of confidential attorney- client communications, Carter, 429 F. Supp. 3d at 900; see, e.g., id. at 847–866, in violation of an unknowable number of defendants’ Sixth Amendment rights to attorney-client confidentiality. As we recently summarized, “the [prosecutors] intruded into a large number of defendants’ communications with their attorneys, with no legitimate law-enforcement purpose, and later tried to conceal these actions.” United States v. Orduno-Ramirez, 61 F.4th 1263, 1267 (10th Cir. 2023). When this misconduct came to light, more than one hundred federal prisoners, including Mr. Hohn, petitioned for habeas relief.

Public confidence in the fairness of the criminal process demands scrutiny of the prosecutor, whose “role transcends that of an adversary,” United States v. Bagley, 473 U.S. 667, 675 n.6 (1985), and “whose

These facts are derived from the district court’s memorandum and

1

order denying Mr. Hohn’s habeas petition, see RII.1729, and from the district court’s factual findings and conclusions of law in Carter, United States v. Carter, 429 F. Supp. 3d 788 (D. Kan. 2019), order vacated in part, No. 16-20032- 02-JAR, 2020 WL 430739 (D. Kan. Jan. 28, 2020); see RI.2728.

interest . . . in a criminal prosecution is not that it shall win a case, but that justice shall be done,” Berger v. United States, 295 U.S. 78, 88 (1935); see also Kaur v. Maryland, 141 S. Ct. 5, 7 (2020) (Sotomayor, J., statement respecting the denial of certiorari) (“Prosecutors wield an immense amount of power, and they do so in the name of the State itself.”). A prosecutor’s ethical and constitutional obligations go “to the very integrity of the legal system.” Gray v. Mississippi, 481 U.S. 648, 668 (1987). And prosecutors have a well-established affirmative obligation “not to act in a manner that circumvents and thereby dilutes the protection afforded by the right to counsel.” Maine v. Moulton, 474 U.S. 159, 170–71 (1985); see also id. at 176.

The majority opinion says the “scandal” underlying this appeal is “the Kansas USAO’s mishandling of attorney-client communications.” Op. at 51, 3 (emphasis added). That puts it mildly. My colleagues appropriately “condemn” this misconduct. Op. at 50. But condemnation demands more elaboration.2 Offutt v. United States, 348 U.S. 11, 14 (1954) (“[J]ustice must satisfy the appearance of justice.”).

In this appeal, we must decide whether “basic, constitutional guarantees that should define the framework of any criminal trial”—guarantees protected

2 This factual background was comprehensively recited by the district

court in Carter, 429 F. Supp. 3d 788, but not described so extensively in the majority’s opinion.

by the structural-error doctrine—are fundamentally disrupted when the prosecution intentionally and unjustifiably learns what is said between lawyer and client. Weaver v. Massachusetts, 582 U.S. 286, 294–95 (2017). The district court endeavored to “shine daylight” on these surreptitious incursions into attorney-client communications and took “a wide-lens view of the Government’s conduct implicating the Sixth Amendment inquiry.” Carter, 429 F. Supp. 3d at 800. I respectfully submit we must do the same.

2

At the heart of this case is something ordinary—a defense lawyer talking on the phone to his incarcerated client—and something extraordinary—the prosecutor listening. 3 The Kansas USAO maintained a routine practice of requesting and receiving recordings of phone calls that defendants placed from the Corrections Corporation of America (CCA). The district court found this practice was motivated by the USAO’s belief that the recorded calls would be useful to it for investigative purposes and to prepare for trial and other

3 Defense lawyers routinely make calls to incarcerated clients, and the

defense function depends on adversarial confidentiality. As amici Federal Public Defenders summarize, defendants’ “trial lawyers must often communicate with incarcerated clients by phone, and there is no way to be sure the calls are private”—making integrity by prosecutors and other government officials critical. Fed. Pub. Defs. Amicus Br. at 2 (heading capitalization omitted).

hearings. 4 With those objectives, the court concluded, “[f]or years, prosecutors in the Kansas City division had received, or knew others had received, attorney-client calls when they made a general request for all of a detainee’s calls from CCA.” Carter, 429 F. Supp. 3d at 854. This practice was “neither infrequent nor uncommon.” Id. “Every time the USAO made a general request for all recorded calls,” in fact, “there was a 27.96% chance that the calls would include attorney-client calls.” Id. at 856. And while the precise scope of the USAO’s practice is unclear, in part because “the USAO failed to preserve and produce electronic and paper records,” one analysis suggests prosecutors “accessed [an estimated] 1,429.21 attorney-client calls.” Id. The prosecutors did not merely possess those recordings; “[t]he record is clear,” the district court found, “that upon receiving recordings, prosecutors and their agents reviewed the calls.” Id. at 848.

4 The court explained, “[i]t was typical for the USAO to obtain audio

recordings placed by CCA detainees in a wide variety of criminal cases.” RI.2815. Prosecutors testified they obtained these recordings for a variety of reasons:

(1) for voice comparisons to aid in identifying voices on wiretaps or consensual recordings; (2) to see if the defendant had made any inculpatory statements, particularly if the case was going to trial;

(3) to investigate whether a detainee is continuing to engage in conspiratorial or otherwise criminal conduct; or (4) to investigate whether a detainee was violating a court-imposed no-contact order with other detainees or with witnesses.

Carter, 429 F. Supp. 3d at 847.

After this intrusive practice came to light, the USAO “denied that its practices implicated the Sixth Amendment or the attorney-client privilege.” Id. at 799. The district court appointed a Special Master to examine the USAO’s conduct and the cases potentially affected. The Special Master led an almost three-year investigation. Throughout this process, the Kansas USAO “did not cooperate with [the Special Master’s] investigation.” Id. “The Government’s wholesale strategy to delay, diffuse, and deflect,” the district court explained, “succeeded in denying the individual litigants their day in court for almost three years.” Id. at 800. The district court considered this conduct relevant “[a]s part of the Sixth Amendment analysis” because “the Government’s lack of meaningful cooperation in the Special Master’s investigation” had implications for its “credibility.” Id. at 799.

Mr. Hohn was detained at CCA from 2012 to 2014. During litigation of another case arising from the Kansas USAO’s misconduct, United States v. Black, Mr. Hohn discovered the prosecution team had obtained a recording of a call he placed from CCA to his then-newly appointed defense lawyer on April 23, 2012. There is no question that call was recorded by CCA and contained “discussion relating to legal advice or strategy.” RII.1754.

The evidence before the district court established how the prosecution team purposefully obtained and listened to the call between Mr. Hohn and his defense lawyer. The court determined the prosecution gained “access to the

8

audio recordings [at CCA] under circumstances where they knew or should have known the material would include attorney-client communications, with no precautions to exclude or avoid learning the content of these recordings [by] use of a filter or taint team.” RII.1772. The court continued, “[t]he government has never asserted, nor is there evidence to suggest, that any prosecution team member started listening to the April 23, 2012 call, heard [Mr. Hohn’s attorney] Campbell’s voice and the nature of the conversation, and immediately stopped listening to the call.” RII.1779. In fact, there was evidence to the contrary. Assistant United States Attorney (AUSA) Morehead “retained her own copy” of N-8, a CD containing nothing but the recording of Mr. Hohn’s call. RII.1774. And another member of the prosecution team emailed “referenc[ing] those same materials in connection with Hohn,” suggesting the prosecution team knew the content of the recording. RII.1774.

Then, after obtaining and listening to the recording of the call between Mr. Hohn and his attorney, AUSA Morehead “took steps to conceal th[e] tactical advantage” she had gained in doing so. RII.1776. The district court explained “Morehead did not disclose N-8 to Campbell in discovery, . . . admitting this fact to government counsel in a February 13, 2019 email.” RII.1776. The court concluded “[b]y declining to do for Campbell what she represented she normally does, [AUSA] Morehead made it less likely that anyone would discover that she was in possession of N-8.” RII.1776.

AUSA Morehead’s misconduct continued during litigation of Mr. Hohn’s habeas petition. “When the USAO began the process of disgorging calls to the Court, she resisted.” RII.1777. Although she “had every opportunity to explain how, when, and why she obtained access and became privy to Hohn’s attorney- client call, . . . she continued to minimize, deflect, and obfuscate her role in Hohn’s Sixth Amendment claim.” RII.1777. The district court observed, for example, although “she stated in her May 29, 2020 affidavit that she did provide Hohn’s April 23, 2012 call to Campbell,” “she reversed her position once again” in August 2021, “testifying that she was never aware that the prosecution team had obtained the April 2012 calls and therefore did not produce those calls to Campbell in discovery.” RII.1777. During that testimony, she also

equivocated about whether she subpoenaed Hohn’s and [codefendant ] Redifer’s calls; attempted to minimize her role in requesting and obtaining CCA calls; attempted to minimize her knowledge of the USAO’s call-collection procedures between 2012 and 2015; equivocated about a specific defendant’s case;

equivocated about discovery procedures; equivocated about what calls she did and did not produce in discovery; equivocated about threats to government witnesses; and denied any involvement with ‘the second batch’ of calls, despite keeping a copy of N-8 in Hohn’s case file.

RII.1777–78.

“[E]ven after turning over scores of attorney-client calls that ha[d] been in its possession for years, including the call at issue in this case,” the

10

government nevertheless “steadfastly refused to acknowledge the problem before the [district] [c]ourt” and instead “disclaim[ed] any responsibility for fixing that problem.” RII.1781. 5 And “despite evidence of her conduct in both this and other criminal cases, the government has confirmed that it has not imposed internal sanctions or discipline against AUSA Morehead on the basis of untruthfulness.” RII.1782.

B

Premised on this prosecutorial interference with his right to counsel, Mr.

Hohn filed a timely habeas motion raising a Sixth Amendment claim under Shillinger. 6 I generally agree with the majority opinion’s recitation of Shillinger’s facts, but it bears emphasizing just how narrow is Shillinger’s rule.

5 In the district court, AUSA Morehead “denied [she] had any idea that

the prosecution team was in possession of such calls or that they listened to the recordings.” RII.1772; see also Carter, 429 F. Supp. 3d at 898 (“[T]he AUSAs and their agents deny watching or listening to the recordings.”). The government now appears to concede it did obtain and listen to the recording. See Aplee. Supp. Br. at 4 (discussing several of the district court’s findings in the memorandum and order on Mr. Hohn’s habeas petition without disputing the finding that “the prosecution had copies of the recording, knew the call contained attorney-client communications, and nevertheless intentionally listened to it before Mr. Hohn’s trial” (RII.1775–79)).

6Throughout the habeas litigation, the district court observed, the government “continue[d] to trivialize the circumstances precipitating Hohn’s Sixth Amendment claim, . . . referring to his claim for relief as a ‘windfall.’” RII.1781 (quoting ECF No. 1028 at 1).

In Shillinger, we held “when the state becomes privy to confidential communications because of its purposeful intrusion into the attorney-client relationship and lacks a legitimate justification for doing so, a prejudicial effect on the reliability of the trial process must be presumed.” 70 F.3d at 1142. As the district court described, Shillinger “sets a high bar to establish a per se Sixth Amendment violation.” CCA Recordings 2255 Litigation v. United States, No. 19-cv-2491-JAR-JPO, at 19 (Jan. 18, 2021), ECF No. 730. That’s right. Unless the prosecution intentionally (meaning not inadvertently) and unjustifiably (meaning not for a legitimate reason) became privy to (meaning not just simply possessed) confidential communications between lawyer and client, there will be no per se Sixth Amendment violation at all.

It does not take much to remove a case from Shillinger’s slim ambit. Did the prosecution intrude, but by accident? No Shillinger claim. Did the prosecution intrude intentionally, but with a legitimate justification? No Shillinger claim. Did the prosecution intrude, even intentionally and without justification, but without learning the substance of the attorney-client communications? No Shillinger claim. Are all other preconditions satisfied, but the communications cannot reasonably be described as attorney-client confidences? Again, no Shillinger claim. And we have recently limited Shillinger’s application to pretrial intrusions. See Orduno-Ramirez, 61 F.4th at 1273 (concluding “[a] post-plea intrusion is less likely to cause prejudice

12

than a pretrial intrusion because the latter can taint any part of a criminal prosecution—trial, sentencing, or both—and greatly expand the task of ascertaining prejudice as compared to a post-plea intrusion”); United States v. Spaeth, 69 F.4th 1190, 1211 (10th Cir. 2023) (holding a prosecutorial intrusion that would otherwise violate the Sixth Amendment under Shillinger does not invalidate a guilty plea, because Shillinger “does not concern [the] guilty-plea situation” and “has nothing to do with whether a guilty plea is voluntary or knowing”).

Fortunately, the circumstances giving rise to a Sixth Amendment violation under Shillinger will not come up often. “[T]radition and experience justify our belief that the great majority of prosecutors will be faithful to their duty.” 7 United States v. Mezzanatto, 513 U.S. 196, 210 (1995) (quoting Newton v. Rumery, 480 U.S. 386, 397 (1987)). Counting generously—but excluding the appeals stemming from the aberrant misconduct in Kansas—a substantive discussion of Shillinger has come up approximately four times in thirty years in our circuit’s jurisprudence, and never as a basis for granting relief to a

7 The district court reasonably observed “[t]here is not much precedent

for the [c]ourt to draw from [in ruling on a Shillinger claim] for obvious reasons; such governmental intrusions into defendants’ attorney-client relationships are easily prevented by the use of a taint team or other precautions.” RII.1781.

prisoner. 8 There can be no serious question Shillinger’s conclusive presumption of prejudice is reserved for truly “limited” and “exceptional” circumstances. Greer v. United States, 593 U.S. 503, 513 (2021) (internal quotation omitted). Contrary to the government’s framing, that over 100

8 In United States v. Kennedy, 225 F.3d 1187 (10th Cir. 2000), we observed in a footnote the case was “distinguishable from Shillinger in that” the claim concerned “prosecutorial misconduct flowing from invasion of [the defendant’s] relationship with his attorney during the investigative stage of the prosecution,” whereas Shillinger involved the pre-trial phase. Id. at 1195 n.5. We also “agree[d] with the district court’s ultimate conclusion” that the alleged invasion into the attorney-client relationship at issue did not violate the Sixth Amendment. Id.

In Reali v. Abbot, 90 F. App’x 319 (10th Cir. 2004) (unpublished), we discussed Shillinger in outlining when a defendant claiming a Sixth Amendment violation need not show prejudice. Id. at 323 & n.3. We held, because the defendant “fail[ed] to show the prosecution purposefully intruded upon her attorney-client relationship, she is not entitled to a presumption of prejudice.” Id. at 323 (citing Shillinger v. Haworth, 70 F.3d 1132, 1142 (10th Cir. 1995)). This case therefore also distinguished Shillinger.

In United States v. Singleton, 52 F. App’x 456 (10th Cir. 2002)

(unpublished), we denied Mr. Singleton’s request for a certificate of appealability because he did not meet his burden to show a constitutional violation, including under the Sixth Amendment. We again discussed Shillinger in outlining the law surrounding prosecutorial intrusions into attorney-client communications with criminal defendants. Id. at 459. But we ultimately held “[w]e need not decide whether a per se Sixth Amendment violation occurred here” because no remedy would be available even if one had. Id.

In United States v. Harssfell, 735 F. App’x 553 (10th Cir. 2018)

(unpublished), we dismissed an appeal invoking Shillinger frivolously, because the defendant did “not support[] his claim with sufficient evidence of . . . misconduct” that would implicate Shillinger. Id. at 554.

defendants invoked Shillinger after this misconduct is not evidence of Shillinger’s supposedly wide reach as much as it is evidence of this misconduct’s wide reach. Contra Aplee. Supp. Br. at 17–18.

C

Finally, while my main disagreement is with the outcome, I am also concerned about the path taken to achieve it. Mr. Hohn requested and received a certificate of appealability on two issues: (1) did the district court err in ruling that Mr. Hohn failed to prove the elements of his Sixth Amendment claim? And (2) did the district court err in ruling that the government proved Mr. Hohn waived his Sixth Amendment right? The parties briefed those issues and argued the appeal in September 2023 before a three-judge panel. No panel opinion issued.

In January 2024, over dissent, this court ordered initial en banc review sua sponte and posed two new questions. Hohn, 91 F.4th at 1060. The majority opinion now answers only one: whether Shillinger “correctly h[e]ld that it is structural error for the government to purposefully intrude without legitimate justification into the attorney-client relationship and that prejudice must be presumed[.]” 9 Id. En banc review is already “an extraordinary procedure” and

9 The other is “[w]hen, if ever, . . . the government unjustifiably intrude[s] into the attorney-client relationship by intentionally obtaining

15

is “not favored.” Id. at 1061 (Rossman, J., dissenting) (quoting 10th. Cir. R. 35.1(A); Fed. R. App. P. 35(a)). Even more extraordinary is when a court initiates en banc review without being asked.10 Id. Appellate courts are not “self-directed boards of legal inquiry and research.” State v. U.S. Env’t Prot. Agency, 989 F.3d 874, 885 (10th Cir. 2021) (quoting Nat’l Aeronautics & Space Admin. v. Nelson, 562 U.S. 134, 147 n.10 (2011)). But here, the majority, sua sponte, has chosen the issue it wants to decide and decided it—overruling circuit precedent without a request from the parties, without a change in Supreme Court law, and without the participation of two active but recused members of the en banc court. I regret we have so readily bypassed the norms of the appellate process.

II

I now explain why Shillinger should not be disturbed. The majority “conclude[s] that the case—and its structural-error rule—is untenable under Supreme Court law” and therefore “overrule[s] Shillinger.” Op. at 4. To be

attorney-client communications that are not privileged[.]” United States v. Hohn, 91 F.4th 1060, 1060 (10th Cir. 2024).

10 The majority observes “[i]n the panel briefing, the government did

argue that the district court erred by relying on” the structural-error rule in Shillinger because it supposedly “runs contrary to the rule and rationale of” multiple “Supreme Court cases.” Op. at 12 n.13 (quoting Ans. Br. at 24). But nowhere did the government ask this court to overrule Shillinger formally, and as the majority admits, that “the Shillinger decision bound the panel” is “[o]bvious[].” Op. at 12 n.13.

clear, our circuit has not done away entirely with Shillinger. The majority acknowledges, as it must, “a Sixth Amendment violation of the right to confidential communication with an attorney” could exist in some cases, namely where the defendant shows prejudice. Op. at 72; see also Op. at 21 (recognizing a Sixth Amendment claim still exists for some “intentional, unjustified intrusions into the attorney-client relationship”); Op. at 67 (ruling district court shall determine when “other § 2255 litigants might be entitled to an evidentiary hearing” to examine possibility of prejudice).

Though purporting to overrule Shillinger, the majority opinion interrogates only one aspect of its holding: “whether we should retain Shillinger’s structural-error rule or reverse it.” Op. at 4. The majority picks the latter and holds “a Sixth Amendment violation of the right to confidential communication with an attorney requires the defendant to show prejudice.”11 Op. at 4; see also Op. at 14 (“[T]o establish a Sixth Amendment violation, the defendant must show (1) that the government intentionally intruded into the defense camp and (2) that the intrusion caused prejudice.”). While acknowledging the elements of a Sixth Amendment prosecutorial-intrusion claim, the majority departs from Shillinger by holding “the violation is not

11 It is correct to say the conclusive presumption of prejudice in Shillinger

has been “abrogated.” See Judge Bacharach’s Partial Dissent at 1 (“We earlier held that this kind of intrusion creates a conclusive presumption of prejudice. The majority abrogates this holding.” (emphasis added) (citation omitted)).

complete until the defendant establishes prejudice.” Op. at 15. Accordingly, I focus only on the portion of Shillinger the majority has “review[ed], reverse[d], and replace[d]”—its conclusive presumption of prejudice. Op. at 16.

The decision to abrogate Shillinger’s conclusive presumption does not withstand scrutiny.

First, the majority’s comprehensive reliance on Strickland’s prejudice prong—which applies to Sixth Amendment claims based on defense counsel’s performance—is misplaced. Unlike the ineffective-assistance-of-counsel claims controlled by Strickland, the Sixth Amendment violation at issue here is based on “direct governmental interference with the right to counsel,” which the Supreme Court has “expressly noted . . . is a different matter.” Perry v. Leeke, 488 U.S. 272, 279 (1989). Supreme Court precedent confirms the error recognized in Shillinger is structural, meaning prejudice must be presumed. Requiring the defendant to show prejudice here, because Strickland did, shows how the majority misunderstands the nature of the Sixth Amendment right at issue.

Second, Shillinger was correct at inception, and its per se prejudice rule does not, as the majority claims, conflict with the Supreme Court’s jurisprudence on government intrusions into attorney-client communications.

Third, traditional stare decisis factors, unaddressed by the majority, uniformly support retaining Shillinger in full.

A

The foundational assumption underlying the majority opinion is that Strickland prejudice applies to the prosecutorial intrusion claim recognized in Shillinger. This is wrong. The majority recognizes that many Supreme Court cases, like Shillinger, adopt a conclusive presumption of prejudice when the government interferes with the right to counsel. Op. at 45. Attempting to distinguish these precedents, the majority reasons “the prejudicial impact was tangible” in each of those cases “[b]ecause the judicial interference . . . jeopardized the integrity and fairness of the trial itself.” Op. at 46. But the majority’s reasoning fails to account for the different ways the Supreme Court analyzes ineffective-assistance violations based on who causes them.

A defense counsel’s Sixth Amendment violations, at issue in Strickland, look very different from the government’s Sixth Amendment violations, at issue in Shillinger. And this fundamental difference has led the Supreme Court to treat them differently. Relying so centrally on Strickland—without regard to the critical differences in how different actors can violate the Sixth Amendment—is misguided. This section explains why. I first show the majority makes an incorrect doctrinal assumption—all ineffective-assistance- of-counsel claims are alike and all are subject to Strickland’s prejudice requirement. I then show why a better reading of Supreme Court caselaw supports Shillinger’s structural-error rule.

a

The majority begins by stating “[t]he Sixth Amendment guarantees a right to the effective assistance of counsel.” Op. at 13 (citing Strickland v. Washington, 466 U.S. 668, 686 (1984); McMann v. Richardson, 397 U.S. 759, 771 n.14 (1970)). It then calls the Sixth Amendment “right to communicate confidentially with an attorney” “[p]art and parcel of” that same effective- assistance right. Op. at 13–14 (citing Weatherford v. Bursey, 429 U.S. 545, 554 n.4 (1977)). The majority asserts “the legal principles that govern effective- assistance claims apply equally to”—that is, govern—claims based on intrusions into “attorney-client confidentiality.” Op. at 55.

“[B]ecause we derive ‘the right to effective representation from the purpose of ensuring a fair trial,’” the majority reasons, “we should ‘also derive[] the limits of that right from that same purpose.’” Op. at 55 (second alteration in original) (quoting United States v. Gonzalez-Lopez, 548 U.S. 140, 147 (2006)). This limit, the majority suggests, is the requirement that a defendant must show prejudice. Because “[t]he right to communicate confidentially with an attorney originates from the Sixth Amendment’s promise of effective assistance of counsel,” the majority concludes incursions on confidentiality do not violate the constitution unless they too involve prejudice—that is, “unless [they] call[] into question the basic justice of a defendant’s conviction or

20

sentence.” Op. at 54–55 (quoting Lafler v. Cooper, 1082 566 U.S. 156, 178 (2012) (Scalia, J., dissenting)).

The majority thus collapses two distinct guarantees under the Sixth Amendment—to be free of prosecutorial intrusion into attorney-client confidences and to have effective performance by defense counsel. Baked into the majority’s reasoning is a tacit premise that “a fair trial,” Op. at 56, is at risk, and thus prejudice is present, only in the way Strickland recognized when the claim is defense counsel performed ineffectively. See Strickland, 466 U.S. at 686. There is, the majority suggests, only one kind of ineffective-assistance violation. 12 This reasoning does not withstand scrutiny, as I will explain.

b

At root, the Sixth Amendment guarantees, in relevant part, “[i]n all criminal prosecutions, the accused shall enjoy the right . . . to have the Assistance of Counsel for his defence.” U.S. Const. amend. VI. “Th[is] right to counsel is the right to the effective assistance of counsel.” Strickland, 466 U.S. at 686 (emphasis added) (quoting McMann, 397 U.S. at 771 n.14). Put

12 Notably, in a case arising from the same prosecutorial intrusions that

affected Mr. Hohn, this court recognized as an unjustified analytical “shortcut[]” a defendant’s attempt to “equate[] lack of effective assistance of counsel” in the sense required by Shillinger “with ‘ineffective assistance of counsel’ as required by” Strickland and similar cases. United States v. Spaeth, 69 F.4th 1190, 1211 (10th Cir. 2023). Why this court has now abandoned its distinct understanding of these two kinds of Sixth Amendment violations is unclear.

differently, the word “effective” is implied before the word “Assistance” in the Sixth Amendment. As the majority acknowledges, see Op. at 13, “[t]he purpose” of this effective-assistance guarantee “is simply to ensure that criminal defendants receive a fair trial.” Strickland, 466 U.S. at 689.

It is true the Court has framed government interference as implicating the “right to the effective assistance of counsel,” Weatherford, 429 U.S. at 547— the same general right recognized in Strickland. But nothing indicates the Supreme Court intended Strickland’s prejudice prong, applicable to situations where counsel allegedly performs below the required standard, to apply where the prosecution intentionally and without justification intrudes on attorney- client communications. See Strickland, 466 U.S. at 692. The Court has drawn a sharp distinction in how it evaluates the effective-assistance right based on who interferes with it.

Multiple parties can render assistance ineffective. Typically, as in Strickland, counsel’s performance implicates the Sixth Amendment guarantee. See id. But the government can also interfere with the right to counsel. And the Court has “expressly noted that direct governmental interference with the right to counsel is a different matter.” Perry, 488 U.S. at 279. Such direct interference “is a different matter” in a particular way. The Court has frequently found a Sixth Amendment violation “without any showing of prejudice when counsel was . . . prevented”—including by the government—

“from assisting the accused during a critical stage of the proceeding.” United States v. Cronic, 466 U.S. 648, 659 n.25 (1984) (emphasis added) (collecting cases).

This difference goes to the central fair-trial right animating the Sixth Amendment’s effective-assistance guarantee. See Strickland, 466 U.S. at 689. The Court has, reasonably enough, applied different standards throughout its Sixth Amendment jurisprudence based on who interferes with this fair-trial right. When defense counsel does so, the Court begins its analysis with the premise that “a fair trial is one in which evidence subject to adversarial testing is presented to an impartial tribunal for resolution of issues defined in advance of the proceeding.” Id. at 685. Thus, in that context, “[t]he benchmark for judging any claim of ineffectiveness must be whether counsel’s conduct so undermined the proper functioning of the adversarial process that the trial cannot be relied on as having produced a just result.” Id. at 686. In other words, an ineffective defense counsel creates a sufficiently unfair trial only when her performance renders the adversarial process altogether unreliable.

This sort of defense-counsel-caused “breakdown in the adversary process,” id. at 687, is something a defendant must show based on the particular facts of the case. Only in limited cases has the Court found defense counsel to violate the Sixth Amendment’s guarantees without an individualized showing of prejudice. See Cuyler v. Sullivan, 446 U.S. 335, 349

(1980) (when “a defendant . . . shows that [his attorney’s] conflict of interest actually affected the adequacy of his representation”); McCoy v. Louisiana, 584 U.S. 414, 427–28 (2018) (when the attorney admits his client’s guilt over the client’s objection). 13 But the Court has recognized that the government can undermine the fairness of trial, and thus violate the Sixth Amendment, in ways that do not require the defendant to prove prejudice. For instance, the Court has held the government always violates the Sixth Amendment when it:

• disallows direct examination of the defendant, Ferguson v. Georgia, 365 U.S. 570, 596 (1961);

• holds certain proceedings without the opportunity to access counsel, Williams v. Kaiser, 323 U.S. 471, 475–76 (1945); Hamilton v.

Alabama, 368 U.S. 52, 55 (1961); White v. Maryland, 373 U.S. 59, 60 (1963); Gideon v. Wainwright, 372 U.S. 335, 344–45 (1963);

• disallows closing arguments, Herring v. New York, 422 U.S. 853, 863 (1975);

13And even in Cuyler, the “presumption of prejudice” the Court recognized was “more limited” than in government-interference cases. Strickland v. Washington, 466 U.S. 668, 692 (1984).

• prevents attorney-client consultations in the evening during the defendant’s testimony, Geders v. United States, 425 U.S. 80, 91 (1976);

• denies a defendant’s request to proceed pro se, McKaskle v. Wiggins, 465 U.S. 168, 177 & n.8 (1984);

• denies a public trial or hearing, Waller v. Georgia, 467 U.S. 39, 49 & n.9 (1984);

• discriminates unconstitutionally in grand jury selection, Vasquez v.

Hillery, 474 U.S. 254, 261–64 (1986);

• fails to provide a reasonable-doubt jury instruction, Sullivan v.

Louisiana, 508 U.S. 275, 279–80 (1993);

• rejects a defendant’s choice of counsel, Gonzalez-Lopez, 548 U.S. at 147–50; or • fails to recuse when the Constitution so requires, Williams v.

Pennsylvania, 579 U.S. 1, 14–15 (2016).

While certainly not all, or even most, trial errors render criminal proceedings fundamentally unfair, the Supreme Court has not hesitated to recognize that especially unfair conduct can do so—including when the government interferes with a defendant’s Sixth Amendment right to effective representation. See, e.g., Kaiser, 323 U.S. at 475–76; Hamilton, 368 U.S. at 55; White, 373 U.S. at 60; Gideon, 372 U.S. at 344–45; Geders, 425 U.S. at 91;

McKaskle, 465 U.S. at 177 n.8; Gonzalez-Lopez, 548 U.S. at 149. And while much of the relevant jurisprudence involves interference by the judicial and legislative branches, there can be no serious question obligations owed by “the government” extend to prosecutors. See Moulton, 474 U.S. at 171 (“We have on several occasions been called upon to clarify the scope of the State’s obligation in this regard, and have made clear that . . . the prosecutor and police have an affirmative obligation not to act in a manner that circumvents and thereby dilutes the protection afforded by the right to counsel.” (emphasis added)).

Strickland itself recognized “various kinds of state interference with counsel’s assistance” involve a presumption of prejudice. 466 U.S. at 692 (citing Cronic, 466 U.S. at 659 & n. 25); see also Weaver, 582 U.S. at 308 (Alito, J., concurring in the judgment on behalf of himself and Justice Gorsuch) (“The Court has relieved defendants of the obligation to make this affirmative [prejudice] showing in only a very narrow set of cases . . . includ[ing] the actual or constructive denial of counsel, state interference with counsel’s assistance, or counsel that labors under actual conflicts of interest.” (emphasis added)). What Strickland identified as “subject to a general requirement that the defendant affirmatively prove prejudice” are claims “alleging a deficiency in [defense] attorney performance.” 466 U.S. at 693. Recall, the Court has “expressly noted that direct governmental interference with the right to counsel is a different matter.” Perry, 488 U.S. at 279. The majority’s analysis is fatally premised on

26

such interference being analytically the same. 14 By equating all ineffective- assistance violations with the counsel-caused violations discussed in Strickland, the majority’s reasoning is incompatible with the Court’s Sixth Amendment jurisprudence. 15

14 The majority is correct that “the government’s intrusion into the attorney-client relationship ‘inhibit[s] [the] free exchanges between defendant and counsel’ and therefore constrains an attorney’s ability to effectively represent a defendant.” Op. at 14 (alterations in original) (quoting Weatherford v. Bursey, 429 U.S. 545, 554 n.4 (1977)). But this is just one of the problems with government intrusion. That government intrusion often hinders a particular defense attorney plainly does not mean government intrusion violates the Sixth Amendment only when it hinders a particular defense attorney.

15 The majority argues I “err[] by” using “the overbroad term ‘governmental interference.’” Op. at 21 n.15. Using that term, the majority says, “merges prosecutorial interference with judicial interference.” Op. at 22 n.15. I am not persuaded. First, the terms the Supreme Court uses to describe this class of cases are, tellingly, “governmental interference,” or “state interference,” not a more limited term. E.g., Perry v. Leeke, 488 U.S. 272, 279 (1989); Strickland, 466 U.S. at 692. “[A] good rule of thumb for reading [Supreme Court] decisions is that what they say and what they mean are one and the same . . . .” Mathis v. United States, 579 U.S. 500, 514 (2016). Second, some of the cases the majority itself places in this canon do not exclusively involve judicial interference. See, e.g., Op. at 46 (citing Ferguson v. Georgia, 365 U.S. 570, 594–95 (1961), and Herring v. New York, 422 U.S. 853, 863 (1975)—both of which involved legislative, not judicial, interference with the effective assistance of counsel via a statute—as exemplar structural-error cases).

My colleagues continue: “if that line of cases requires structural error for prosecutorial intrusions, Weatherford and Morrison failed to notice so and blundered by repeatedly discussing the need for prejudice.” Op. at 22 n.15. But nowhere do I argue, as the majority suggests, all prosecutorial intrusions amount to structural error. My point is, more simply, government-interference

c

This Supreme Court-recognized distinction makes sense. Practically speaking, a Sixth Amendment violation looks very different when a defendant’s counsel causes it through ineffective performance versus when the government causes it through intentional and unjustified intrusion. These differences are at least threefold.

For counsel-caused errors, first, because “[r]epresentation is an art, . . .

an act or omission that is unprofessional in one case may be sound or even brilliant in another.” Strickland, 466 U.S. at 693. A set of precise rules for lawyering, without looking at the impact of counsel’s errors in an individual case, would be unmanageable, as “[a]ttorney errors come in an infinite variety.” Id. Second, “[t]he government is not responsible for, and hence not able to prevent, attorney errors that will result in reversal of a conviction or sentence.” Id. Third, in a world where ineffectiveness claims were easy to prove, “[c]riminal trials resolved unfavorably to the defendant would increasingly come to be followed by a second trial” on the question of the first counsel’s ineffective performance. Id. at 690. As a result, the Supreme Court worried that a low bar for establishing ineffective assistance would deter lawyers from

cases cannot be subject to the same analytical framework as Strickland. And as I will explain, Weatherford and Morrison present situations wholly unlike that presented in Shillinger or this case. See infra section II.B.2.a.–b.

representing criminal defendants. Id. Largely because of the weight of these concerns, the Strickland Court adopted a test that, while often asserted, has been extremely difficult to satisfy.

None of these rationales applies to the kind of government-caused violation in Shillinger—and in this case. First, prosecutorial intrusions into a defendant’s attorney-client relationship do not admit of nearly as much nuance as attorney errors in representing a defendant. Shillinger recognized a few different variations—for instance, based on whether the intrusions are intentional or justified, 70 F.3d at 1140—but that is hardly enough open-ended variety to deem government intrusions into attorney-client relationships “an art,” Strickland, 466 U.S. at 693. Second, “[t]he government” is clearly “responsible for,” and thus easily “able to prevent,” its own errors. Id. Third, a too-frequent “second trial,” id. at 690, is no concern given Shillinger’s extreme rarity outside irregular circumstances like the misconduct in Kansas. 16 It is

16 According to the majority opinion, per se rules like Shillinger’s “‘cut[]

much too broadly’ to safeguard the Amendment’s guarantees” because they “indiscriminately recognize constitutional violations” without a showing of prejudice. Op. at 40 (alteration in original) (quoting Weatherford, 429 U.S. at 557). I disagree. Shillinger established a per se rule only as to a specific type of intrusion (intentional, unjustified), in a specific context (pre-trial), regarding a specific type of communication (confidential and between a defendant and his attorney). Again, since we decided Shillinger, we have never again granted relief on that basis. And, in a reality in which “nine out of ten [criminal charges] are resolved by plea and the remaining trials favor conviction . . . fairness, honesty, and morality are not an undue burden on accomplished justice.”

therefore much easier to envision the Framers endorsing a remedy whenever the government violates the rule “do not intentionally and unjustifiably become privy to defendants’ confidential attorney-client communications” than whenever a defense attorney violates the rule “do not represent your client ineffectively.”

d

Shillinger properly understood the violation at issue as analytically distinct from counsel-caused violations recognized in Strickland, and as more analogous to the government-caused violations recognized in the cases described above. The majority now unjustifiably elides this distinction.

Shillinger, unlike the majority opinion, recognized how intentional, unjustified intrusions implicate the overriding structural concern about “render[ing] a trial fundamentally unfair.” 70 F.3d at 1142 (quoting Rose v. Clark, 578 U.S. 570, 577 (1986)). Shillinger therefore properly found a Sixth Amendment violation because the prosecutor’s actions constituted “a direct interference with the Sixth Amendment rights of a defendant”—not just because a conviction was more likely. Id. (emphasis added). Indeed, the idea that the violation alleged in Shillinger is amenable to analysis under Strickland would have come as quite a surprise to the Shillinger panel, who

Shillinger, 70 F.3d at 1142 (quoting Haworth v. State, 840 P.2d 912, 919 (Wyo. 1992) (Urbigkit, J., dissenting)).

recognized the Supreme Court’s explicit rejection of collapsing government- and counsel-caused violations. See Shillinger, 70 F.3d at 1141 (“‘[I]n certain Sixth Amendment contexts, prejudice is presumed.’ This is particularly true with regard to ‘various kinds of state interference with counsel’s assistance.’” (quoting Strickland, 466 U.S. at 692)); see also id. (citing Perry, 488 U.S. at 279–80, and Cronic, 466 U.S. at 658 & n.24 for this general proposition, and then discussing Ferguson, 365 U.S. 570, Brooks v. Tennessee, 406 U.S. 605 (1972), Herring, 422 U.S. 853, and Geders, 425 U.S. 80, as cases in this tradition).

And this distinct understanding could not be otherwise. The governmental interference at issue in Shillinger and in this case fundamentally “affec[ts] the framework within which the trial proceeds.”17 Gonzalez-Lopez, 548 U.S. at 148 (quoting Fulminante, 499 U.S. at 310). This is so for at least two reasons.

First, if prosecutors may effectively listen in without consequence (as here), that is all but certain to affect defense attorneys’ “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.” Gonzalez-Lopez, 548 U.S. at 150. Under these circumstances,

17As I will discuss more infra section II.A.2.c., this rationale is one reason Shillinger was correct to deem the error at issue “structural.”

“[i]t is impossible to know what different choices . . . counsel would have made, and then to quantify the impact of those different choices on the outcome of the proceedings.” Id.

As Shillinger recognized, the Court has not hesitated to find a Sixth Amendment violation “without any showing of prejudice when counsel was . . . prevented from assisting the accused during a critical stage of the proceeding.” Cronic, 466 U.S. at 659 n.25 (collecting cases); Shillinger, 70 F.3d at 1141–42. The same must be true when counsel is “prevented from assisting the accused” openly and freely, without the specter of prosecutorial intrusion clouding the communications. Holding otherwise, as the majority does, violates the Supreme Court’s exhortation that “there can be no restrictions upon the function of counsel . . . in accord with the traditions of the adversary factfinding process.” Herring, 422 U.S. at 857; see also id. (noting the “right to the assistance of counsel . . . ensures to the defense in a criminal trial the opportunity to participate fully and fairly in the adversary factfinding process” (emphasis added)); Weatherford, 429 U.S. at 556 (recognizing the possibility that confidential information communicated to the prosecution “unfairly advantaged the prosecution[] and threatened to subvert the adversary system of criminal justice”).

Second, and relatedly, defendants themselves may behave differently—

less candidly, say—when the adversary may be intruding on their attorney-

client conversations. “And then we would have to speculate upon what effect those different choices or different intangibles might have had.” 18 Gonzalez- Lopez, 548 U.S. at 151. That is especially true when, as in Shillinger and this case, the communications “involved legal advice or strategy.” RII.1760. Ensuring a fully informed defense attorney, and a fully candid defendant, goes to the very “framework within which the trial proceeds.” Gonzalez-Lopez, 548 U.S. at 148 (quoting Fulminante, 499 U.S. at 310); see also Perry, 488 U.S. at 284 (describing Geders as involving a per se Sixth Amendment violation because of government interference with “the normal consultation between attorney and client . . . [on] matters that the defendant does have a constitutional right to discuss with his lawyer, such as the availability of other witnesses, trial tactics, or even the possibility of negotiating a plea bargain”). As Shillinger recognized, government interference in the counsel relationship “disabl[es] [defendants’] counsel from fully assisting and representing [them].” 70 F.3d at 1141 (quoting United States v. Decoster, 624 F.2d 196, 201 (D.C. Cir. 1976)).

18 There is no need to speculate, claims the majority, because “Hohn

concedes that he suffered no prejudice.” Op. at 4. I would not read nearly as much into that supposed “conce[ssion]” as the majority does. As I will explain, with Shillinger on the books, Mr. Hohn’s defense counsel had no obligation to show prejudice on an individualized basis.

I understand the right at stake here as safeguarding the capacity of the adversarial system to produce just results—which requires both effective performance by counsel and fair adversarial conditions systemically. The majority focuses only on the former, but Shillinger correctly accounted for the latter.

2

The particular error here, like those in many of the government-

interference cases described above, is structural. Shillinger gave three reasons for finding purposeful and unjustified prosecutorial intrusions into confidential attorney-client communications “structural”: (1) “no other standard can adequately deter” the type of Sixth Amendment violations at issue; (2) “[p]rejudice in these circumstances is so likely that case-by-case inquiry into prejudice is not worth the cost”; and (3) “such intentional and groundless prosecutorial intrusions are never harmless because they ‘necessarily render a trial fundamentally unfair.’” 70 F.3d at 1142 (alteration in original) (first quoting Strickland, 466 U.S. at 692; and then quoting Rose, 478 U.S. at 577).

These rationales map cleanly onto the Supreme Court’s most recent comprehensive statement of what makes an error “structural” in Weaver v. Massachusetts, 582 U.S. 286. As the majority opinion correctly describes,

[t]he Supreme Court generally classifies an error as structural (1) “if the right at issue is not designed to protect the defendant from erroneous conviction but instead protects some other

34

interest”; (2) “if the effects of the error are simply too hard to measure”; and (3) “if the error always results in fundamental unfairness.”

Op. at 22 (quoting Weaver, 582 U.S. at 295–96). The Weaver Court stressed “one point is critical: An error can count as structural even if the error does not lead to fundamental unfairness in every case.” 582 U.S. at 296 (citing Gonzalez–Lopez, 548 U.S. at 149 n.4). As Justice Alito observed, these rationales place “state interference with counsel’s assistance” among the “very narrow set of cases” in which “[t]he Court has relieved defendants of the obligation” to show prejudice on an individualized basis. Id. at 308 (Alito, J., concurring in the judgment). The conclusive presumption of prejudice in Shillinger comports fully with each of these three rationales. The majority deems none apply, but it is mistaken.

a

The first Weaver rationale—that “the right at issue is not designed to protect the defendant from erroneous conviction but instead protects some other interest,” 582 U.S. at 295—aligns with our stated objective in Shillinger and supports a structural error rule in this context. In Shillinger, we said one objective underlying the rule was to “adequately deter this sort of misconduct.” 70 F.3d at 1142; see also id. at 1142 (stating “a categorical approach is appropriate” in part because “these impediments . . . are susceptible to easy correction by prophylactic rules” (quoting Decoster, 624 F.2d at 201)). Our

“discussion of deterrence” in Shillinger, Mr. Hohn persuasively explains, “fits comfortably within structural error’s ‘other interest’ rationale.” Aplt. Supp. Br. at 20. I agree the structural error rule here protects some “other interest,” implicating the first Weaver rationale.

Weaver describes the “other interest” as one in which “harm is irrelevant to the basis underlying the right.” 582 U.S. at 295 (citing Gonzales-Lopez, 548 U.S. at 149 n.4). For example, one right justified under this rationale is the right to self-representation, “a right that when exercised usually increases the likelihood of a trial outcome unfavorable to the defendant.” McKaskle, 465 U.S. at 177 n.8 (emphasis added). Denying that right constitutes structural error, the Court held, because that harms “the accused’s individual dignity and autonomy.” Id. at 178.

The majority extrapolates a great deal from this example of one “other interest”—autonomy—that can underlie structural error. It summarizes Mr. Hohn’s invocation of the “other interest” rationale, then pivots to calling the rights the Court recognized under this prong “autonomy rights.” Op. at 53. Thus, the majority suggests, because Mr. Hohn did not claim his autonomy was impaired but instead asserted “effective-assistance rights,” this other- interest justification fails. Op. at 54.

Yet the idea that one articulated “other interest” is the only cognizable “other interest” does not follow. True, the Court has recognized several errors

36

as structural in service of preserving defendants’ autonomy. See McCoy, 584 U.S. at 427–28 (when a defense attorney admits his client’s guilt over the client’s objection); McKaskle, 465 U.S. at 177 n.8, 178 (when the state disallows self-representation); Gonzalez-Lopez, 548 U.S. at 147–49 (when the state rejects a defendant’s choice of counsel). But autonomy is plainly not the only “other interest” that could underlie structural error.

As one of our amici helpfully explains, the Court has also suggested (albeit before it coined the phrase “structural error”) that deterrence of prosecutorial misconduct is another interest supporting treatment of the error as structural. See Nat’l Ass’n of Crim. Def. Laws. Amicus Br. at 14–15 (quoting Vasquez, 474 U.S. 254). 19 Failing to deter misconduct that undermines “the structural integrity of the criminal tribunal” itself undermines the structural integrity of the criminal tribunal. Vasquez, 474 U.S. at 263–64. And, like the “other interests” Weaver recognizes, individualized “harm is irrelevant,” 582

19 In Vasquez v. Hillery, the Court reversed the defendant’s conviction

based on the state’s intentional racial discrimination in the selection of the grand jury, refusing the state’s invitation to hold the violation “amounted to harmless error” because it had affected only the indictment process rather than the trial. 474 U.S. 254, 260–62 (1986). The Court stated it has “rejected all arguments that a conviction may stand despite racial discrimination in the selection of the grand jury,” highlighting this error is “possible only under color of state authority” and is thus “wholly within the power of the State to prevent.” Id. at 261, 262. “If grand jury discrimination becomes a thing of the past,” the Court concluded, “no conviction will ever again be lost on account of it.” Id. at 262.

U.S. at 287, when deterrence is necessary “to eliminate [a] systemic flaw,” Vasquez, 474 U.S. at 264. Our stated objective in Shillinger to deter the sort of prosecutorial misconduct that undermines the adjudicatory framework— “government intrusion of the grossest kind upon the confidential relationship between the defendant and his counsel,” Hoffa v. United States, 385 U.S. 293, 306 (1966)—thus expressed the same purpose captured in the first Weaver rationale. The majority’s narrow focus on autonomy does not persuade otherwise. 20 Even if autonomy were the main concern, I cannot see how the specter of consequence-free prosecutorial intrusions on confidential attorney-client communications could avoid impinging on “the defendant’s power to steer the ship of his own defense.” Op. at 53. Full and candid conversations with one’s

20 Likewise, the Supreme Court has recognized “dignity” is an interest

separate from autonomy. McKaskle v. Wiggins, 465 U.S. 168, 179 (1984); see also id. (“Appearing before the jury in the status of one who is defending himself may be equally important to the pro se defendant. . . . From the defendant’s own point of view, the right to appear pro se can lose much of its importance if only the lawyers in the courtroom know that the right is being exercised.” (emphasis added)). The quoted language suggests dignity has additional value even when, in fact, the law respects autonomy. Consider the indignity to the defendant the prosecution introduces by intentionally and unjustifiably becoming privy to confidential communications with counsel. For the defendant, the rules have been inverted and expectations of adversarial confidentiality wholly undermined by the defendant’s adversary— compromising not only the defense function but the public’s perception of our justice system.

attorney—possible only when the adversary does not unjustifiably listen in— are necessary to ensure a fully informed and freely chosen defense strategy.

b

The second Weaver rationale—that the effects of the error “are simply too hard to measure,” 582 U.S. at 295—is also consistent with our reasoning in Shillinger and supports its structural-error rule. “In adopting th[e] [structural error] rule,” Shillinger recognized, “[p]rejudice in these circumstances is so likely that case-by-case inquiry into prejudice is not worth the cost.” 70 F.3d at 1142 (quoting Strickland, 466 U.S. at 692). Given the impossibility of determining “what might have occurred in an alternate universe,” Gonzalez- Lopez, 548 U.S. at 150, the costs of proving prejudice in each case are indeed very high.

As Mr. Hohn explains, “[h]armless-error analysis in [this] context would be a speculative inquiry into what might have occurred in an alternate universe.” Aplt. Supp. Br. at 19 (alterations in original) (quoting Gonzalez- Lopez, 548 U.S. at 150). In support, Mr. Hohn refers to the Third Circuit’s analysis in United States v. Levy, 577 F.2d 200 (3d Cir. 1978), which similarly adopted a per se rule for Sixth Amendment claims premised on “knowing invasion[s] of the attorney-client relationship . . . where confidential

39

information is disclosed to the government.” 21 Aplt. Supp. Br. at 19–20 (citing Levy, 577 F.2d at 208). There, the court justified its decision not to require the defendant to show prejudice by observing

it is highly unlikely that a court can . . . arrive at a certain conclusion as to how the government’s knowledge of any part of the defense strategy might benefit the government in its further investigation of the case, in the subtle process of pretrial discussion with potential witnesses, in the selection of jurors, or in the dynamics of trial itself.

Levy, 577 F.2d at 208. Without a structural-error rule, in a case like this one, a trial court “would face the virtually impossible task of reexamining the entire proceeding to determine whether the disclosed information influenced the government’s investigation or presentation of its case or harmed the defense in any other way.” Id.

21 The majority attempts to downplay Levy by discussing a subsequent

unpublished case, United States v. Mitan. Op. at 68 (citing United States v. Mitan, 499 F. App’x 187 (3d Cir. 2012) (unpublished)). Through Mitan, the majority reasons, “the Third Circuit has . . . rolled back Levy’s interpretation of Weatherford in light of . . . Morrison.” Op. at 68. But Mitan changes nothing.

First, as the majority acknowledges, Mitan “did not overturn Levy.” Op.

at 68; see Mitan, 499 F. App’x at 192 n.6 (“We need not address the question of whether Morrison precludes the presumption of prejudice approach adopted in Levy.”). Nor could it, as an unpublished panel decision. The Third Circuit’s conclusive presumption from Levy remains. Second, Mitan cites Shillinger for the (correct) proposition “that Morrison had left open the question of whether ‘intentional and unjustified intrusions upon the attorney-client relationship may violate the Sixth Amendment even absent proof of prejudice.’” 499 F. App’x at 192 n.6 (quoting Shillinger, 70 F.3d at 1140). Any reinterpretations of Weatherford in light of Morrison are not germane to this case, which, as I will explain, hinges on exactly what Morrison left open.

The Supreme Court has likewise observed the many insidious ways that potential Sixth Amendment violations can affect the course of a trial. In Herring, for example, the Court justified a presumption of prejudice for denying defense counsel the opportunity to give a closing argument in part by noting the difficulty of showing prejudice in that context. See 422 U.S. at 863. Conceding the per se rule would apply even in some cases in which the closing argument would have “le[ft a] judge just where it found him,” the Court nevertheless concluded a specific inquiry into prejudice was not required because in that context, the counterfactual was unknowable. Id. at 863 (quoting Robert H. Jackson, The Struggle for Judicial Supremacy: A Study of a Crisis in American Power Politics 301 (1941)).

Similarly, Justice Sotomayor described in her statement respecting the denial of certiorari in Kaur several examples “of the many ways in which the prosecutors’ possession of Kaur’s privileged information could have subtly but indelibly affected the course of her trial.” 141 S. Ct. at 7. For example,

The prosecutors, either intentionally or subconsciously, may have selected a different mix of jurors. They may have changed their pretrial preparation, perhaps by emphasizing different parts of the State’s case or focusing on different weaknesses in the defense. Or they may have considered different lines of questioning, brainstormed different objections, or anticipated different arguments.

Id. Because of the infinite permutations of strategic decisions resulting from an adversarial disclosure, “[i]t would be an impossible task for any court, no

41

matter how diligent, to identify and assess all potential sources of prejudice simply by comparing the records of two trials.” Id. This reasoning echoes the Court’s understanding in seminal state-interference cases.

As Weaver affirms, this rationale remains viable. See 582 U.S. at 295 (“[A]n error has been deemed structural if the effects of the error are simply too hard to measure.”). The Court recently held in Gonzalez-Lopez that denying a defendant’s choice of counsel was appropriately considered structural error in part because “[i]t [wa]s impossible to know what different choices the rejected counsel would have made, and then to quantify the impact of those different choices on the outcome of the proceedings.” 548 U.S. at 150.

When the prosecution becomes privy to confidential defense communications revealing trial strategy and preparations, that information almost certainly affects a range of the prosecution’s decisions. These decisions could, as the Third Circuit described, include jury selection, its own case preparation, its lines of witness questioning, or its anticipation of the defense’s counterarguments. See Levy, 577 F.2d at 208. They could also affect the government’s plea offer and ultimately the defendant’s decision to proceed to trial. None of these effects would necessarily be measurable; indeed, the prosecution itself might not even be aware its violation had affected the trial. Cf. Herring, 422 U.S. at 863–64. Thus, the difficulty of ascertaining the subtle

42

and indelible effects of intrusion provide another ground supporting Shillinger’s structural-error rule.

c

Finally, Shillinger’s conclusive presumption of prejudice finds support in the third Weaver rationale: the error “cause[s] fundamental unfairness, either to the defendant in the specific case or by pervasive undermining of the systemic requirements of a fair and open judicial process.” Weaver, 582 U.S. at 301. We relied on this same rationale in Shillinger. See 70 F.3d at 1142 (“[S]uch intentional and groundless prosecutorial intrusions are never harmless because they ‘necessarily render a trial fundamentally unfair.’” (quoting Rose, 478 U.S. at 577)). “To provide prosecutors with defense counsel’s legal advice or strategy,” Mr. Hohn argues, “is to skew the adversarial system in a way that makes it impossible ‘to achieve a fair system of justice.’” Aplt. Supp. Br. at 5 (quoting Gideon, 372 U.S. at 344). 22 I agree. The intentional nature of the

22 As noted, the prosecutors clearly believed they would, and did, gain a

“tactical advantage” from these intrusions. RII.1775. Recall, also, the district court’s description of the ways in which the USAO used the recordings for their benefit, including

(1) for voice comparisons to aid in identifying voices on wiretaps or consensual recordings; (2) to see if the defendant had made any inculpatory statements, particularly if the case was going to trial;

(3) to investigate whether a detainee is continuing to engage in conspiratorial or otherwise criminal conduct; or (4) to investigate

43

violation, and its commission by a party acting with the imprimatur of state authority, makes that conduct an even greater affront to the system’s integrity.

According to the majority opinion, that Mr. Hohn “never argue[d] that the prosecutor” actually “had an ‘upper hand’ at his trial . . . takes the wind out of his sails.” Op. at 53. But whether Mr. Hohn made such an argument is beside the point. The Court’s state-interference cases make clear the focus is on systemic fairness, not individualized unfairness. 23 And the third Weaver rationale calls for a consideration of the systemic effects of the intrusion on the fairness and integrity of the judicial process for all defendants. See Weaver, 582 U.S. at 301 (explaining the unfairness may be “either to the defendant in the specific case or by pervasive undermining of the systemic requirements of a fair and open judicial process” (emphasis added)); see also Justin Murray, A Contextual Approach to Harmless Error Review, 130 Harv. L. Rev. 1791, 1813,

whether a detainee was violating a court-imposed no-contact order with other detainees or with witnesses.

Carter, 429 F. Supp. 3d at 847.

23 For example, in Ferguson, the Court struck down a state statute prohibiting defense counsel from eliciting the defendant’s testimony through direct examination. 365 U.S. at 596. Under such conditions, the Court concluded, “it will not be surprising if [a defendant’s] explanation is incoherent, or if it overlooks important circumstances” if he must give it without guiding questions from his counsel. Id. In doing so, the Court relied on the extent to which such a rule would create unfair conditions for the testimony of that defendant and others, without reference to any individualized allegation or showing by Mr. Ferguson that it had harmed him particularly.

1822 (2017) (explaining the “eclectic normative objectives of criminal procedure” include ensuring “that the administration of justice should reasonably appear to be disinterested” (quoting Pub. Utilities Comm’n of D.C. v. Pollak, 343 U.S. 451, 467 (1952) (Frankfurter, J., concurring))).

We therefore correctly considered the intentional prosecutorial misconduct in Shillinger an error that would inherently undermine the “fairness, honesty and morality” of the justice system. 70 F.3d at 1142. The third Weaver rationale provides another basis for the Shillinger rule.

B

I now turn to the majority’s conclusion that binding Supreme Court precedent foreclosed Shillinger’s per se rule at its inception. Contrary to the majority opinion’s understanding, the Supreme Court has never required an additional showing of discrete, trial-specific harm to establish the narrow Sixth Amendment claim at issue in Shillinger and here. I will first briefly describe the Supreme Court decisions addressing violations of the Sixth Amendment right to counsel based on governmental intrusion into attorney-client communications in Black v. United States, 385 U.S. 26 (1966), O’Brien v. United States, 386 U.S. 345 (1967), Hoffa, 385 U.S. 293, Weatherford, 429 U.S. 545, and United States v. Morrison, 449 U.S. 361, 365 (1981). Then, I will explain how the majority opinion misreads these precedents. Finally, I will show Shillinger correctly understood and faithfully applied Supreme Court

45

precedent in defining the Sixth Amendment violation and in deeming it structural error.

1

a. Black In Black, the Court considered a petition for rehearing, which was filed after the Solicitor General voluntarily advised the Court that the prosecution had become privy to pre-trial attorney-client conversations. See 385 U.S. at 27– 28. The Solicitor General explained agents of the Federal Bureau of Investigation—in an investigation for an unrelated matter—had installed monitoring devices in Mr. Black’s hotel room, which had “overheard, among other conversations, exchanges between petitioner and the attorney who was then representing him . . . in this case.” Id. (internal quotation omitted). “Reports and memoranda of the intercepted conversations were examined by the Tax Division attorneys” responsible for prosecution of the case against Mr. Black. Id. at 28. The prosecutors retained copies of the reports, although they maintained they “found nothing in the F.B.I. reports or memoranda which they considered relevant to the tax evasion case.” Id. Mr. Black was ultimately convicted on those tax evasion charges.

Following the Solicitor General’s revelation, he “suggest[ed] that the judgment be vacated and remanded to the District Court in which the relevant materials would be produced and the court would determine, upon an

46

adversary hearing, whether petitioner’s conviction should stand.” Id. (internal quotations omitted). The Court instead ordered a new trial to “afford the petitioner an opportunity to protect himself from the use of evidence that might be otherwise inadmissible.” Id. at 29. In dissent, Justice Harlan emphasized, “the Court today orders a totally new trial in spite of the fact that the disclosure commendably made by the Solicitor General reveal no use of ‘bugged’ material in Black’s prosecution.” Id. at 30–31. That is, Mr. Black showed no prejudice.

b. O’Brien The following year, the Court in O’Brien considered a petition for a writ of certiorari seeking to challenge Mr. O’Brien’s convictions on several counts of removing merchandise from a bonded area. 386 U.S. at 345 (Harlan, J., dissenting). The Court granted the petition, summarily vacated the convictions, and remanded for a new trial, citing Black. Id. at 345 (majority opinion). The dissent provided additional factual context. As in Black, the Solicitor General “commendably notified the Court that pursuant to a general review of the use of ‘electronic eavesdropping or wiretapping,’ he discovered that a microphone had been installed in a commercial establishment owned by an acquaintance of petitioner O’Brien.” Id. at 346 (Harlan, J., dissenting). A pre-trial conversation between Mr. O’Brien and his attorney was recorded. Id. Although the conversation was “overheard by the monitoring agents and summarized in their logs,” it was neither “mentioned in any F.B.I. report” nor

47

conveyed to the attorneys who prosecuted the case. Id. (internal quotations omitted).

The Solicitor General “indicated that he would ‘not oppose’ a remand of the case for an adversary hearing as to the effect of this activity on the validity of petitioners’ convictions.” Id. at 346. The Court instead remanded for a new trial, relying on its decision in Black. Id. at 345 (majority opinion). Justice Harlan reprised his dissent from Black, repeating that “a new trial is not an appropriate vehicle for sorting out the eavesdropping issue because until it is determined that such occurrence vitiated the original conviction”—that is, until prejudice is shown—“no basis for a retrial exists.” Id. at 347 (Harlan, J., dissenting). He concluded that, in his view, “this Court’s action put[] the cart before the horse.” Id.

c. Hoffa In Hoffa, the Court considered a challenge to Mr. Hoffa’s conviction for attempting to bribe members of the jury in an earlier trial in which he was charged with violating the Taft-Hartley Act. 385 U.S. at 294–95. An informant reported to the government and ultimately testified at the later trial about several incriminating statements Mr. Hoffa made during the earlier trial. Id. at 296. The Court granted certiorari to decide whether this prosecutorial conduct invalidated the conviction in the later trial. Id. at 295.

The Court assumed without deciding “the proposition that a surreptitious invasion by a government agent into the legal camp of the defense may violate the protection of the Sixth Amendment.” Id. at 307. “Consequently,” the Court continued, “if the [earlier] trial had resulted in a conviction instead of a hung jury, the conviction would presumptively have been set aside as constitutionally defective.” Id. at 307 (citing Black, 385 U.S. 26). But the Court concluded the conviction in the later trial, which was at issue, needed not be reversed because the fruits of the intrusion were unrelated in “time and subject matter” to that conviction. Id. at 309. The Court reasoned even if a situation existed in which “previous activities in undermining a defendant’s Sixth Amendment rights at one trial would make evidence obtained thereby inadmissible in a different trial on other charges,” this case “d[id] not remotely approach such a situation.” Id. at 308.

d. Weatherford The Court in Weatherford reviewed a Fourth Circuit rule establishing that “whenever the prosecution knowingly arranges and permits intrusion into the attorney-client relationship the right to counsel is sufficiently endangered to require reversal and a new trial.” 429 U.S. at 549–50 (emphasis added) (internal quotation omitted). There, an undercover government informant was present during pre-trial attorney-client meetings. Id. at 547. The district court made an “express finding,” left undisturbed in the Court of Appeals, “that [the

49

informant] communicated nothing at all to his superiors or to the prosecution about [the defendant] Bursey’s trial plans or about the upcoming trial.” Id. at 556; see also id. at 548.

The Court emphasized the agent “went, not to spy, but because he was asked and because the State was interested in retaining his undercover services on other matters.” Id. at 557. Thus, the agent’s presence in the meeting was “necessary to avoid raising the suspicion that he was in fact the informant whose existence Bursey and [his attorney] Wise already suspected.” Id. The Court considered this interest a legitimate justification for the intrusion. See id. (reasoning the Court’s “cases have recognized the unfortunate necessity of undercover work and the value it often is to effective law enforcement”).

The Court reversed the Court of Appeals’ rule that a new trial was per se required under those circumstances. Id. at 558. It reasoned, “[t]here being no tainted evidence in this case, no communication of defense strategy to the prosecution, and no purposeful intrusion by Weatherford, there was no violation of the Sixth Amendment.” Id. But, it also explained,

[h]ad Weatherford testified at Bursey’s trial as to the conversation between Bursey and [his attorney] Wise; had any of the State’s evidence originated in these conversations; had those overheard conversations been used in any other way to the substantial detriment of Bursey; or even had the prosecution learned from Weatherford, an undercover agent, the details of the Bursey-Wise

50

conversations about trial preparations, Bursey would have a much stronger case.

Id. at 554 (emphasis added).

e. Morrison In Morrison, the Court reviewed a Third Circuit order dismissing with prejudice the indictment charging Ms. Morrison with drug distribution. 449 U.S. at 362–64. When Ms. Morrison was indicted, she had retained private counsel. Id. at 362. “Two agents of the Drug Enforcement Agency, aware that [Ms. Morrison] had been indicted and had retained counsel, sought to obtain her cooperation in a related investigation.” Id. The agents “met and conversed with her without the knowledge or permission of her counsel.” Id. In that conversation, the agents “disparaged [her] counsel” and “indicated that [she] would gain various benefits if she cooperated but would face a stiff jail term if she did not.” Id.

Ms. Morrison “moved to dismiss the indictment with prejudice on the ground that the conduct of the agents had violated her Sixth Amendment right to counsel.” Id. at 363. Her motion to dismiss

contained no allegation that the claimed violation had prejudiced the quality or effectiveness of [her] legal representation; nor did it assert that the behavior of the agents had induced her to plead guilty, had resulted in the prosecution having a stronger case against her, or had any other adverse impact on her legal position.

Id. Instead, it alleged the “egregious behavior of the agents . . . ha[d] ‘interfered’” with her right to counsel. Id.

The district court denied her motion to dismiss. Id. The Third Circuit reversed, holding Ms. Morrison’s “Sixth Amendment right to counsel had been violated and that whether or not any tangible effect upon [her] representation had been demonstrated or alleged, the appropriate remedy was dismissal of the indictment with prejudice.” Id. The Supreme Court granted certiorari “to consider whether this extraordinary relief was appropriate in the absence of some adverse consequence to the representation respondent received or to the fairness of the proceedings leading to her conviction.” Id. at 363–64.

Ultimately, the Court assumed without deciding a Sixth Amendment violation had occurred, id. at 364, but held, “absent demonstrable prejudice, or substantial threat thereof, dismissal of the indictment is plainly inappropriate” to remedy the violation, id. at 365. It explained cases involving “Sixth Amendment deprivations are subject to the general rule that remedies should be tailored to the injury suffered from the constitutional violation and should not unnecessarily infringe on competing interests.” Id. at 364. Therefore, “the solution provided by the Court of Appeals [was] inappropriate” as Ms. Morrison had not demonstrated the Sixth Amendment violation that she alleged had an “adverse impact upon the criminal proceedings.” Id. at 367.

I now describe why the majority’s understanding of Weatherford, Morrison, and the three predecessor cases is infirm.

a. The Majority Opinion’s Understanding of Weatherford One of the majority’s central misunderstandings, underlying the entire thesis of its disposition, is Weatherford prefigured the outcome in Shillinger. The majority believes “Weatherford . . . established a prejudice requirement for intrusion-based Sixth Amendment claims.” Op. at 28. It did no such thing.

While Weatherford did require an individualized prejudice showing, see 429 U.S. at 558, it clearly limited the reach of its holding, and that requirement, to the particular intrusion in that case. The Court reasoned,

[h]ad Weatherford testified at Bursey’s trial as to the conversation between Bursey and [his attorney] Wise; had any of the State’s evidence originated in these conversations; had those overheard conversations been used in any other way to the substantial detriment of Bursey; or even had the prosecution learned from Weatherford, an undercover agent, the details of the Bursey-Wise conversations about trial preparations, Bursey would have a much stronger case.

Id. at 554 (emphasis added). The conduct in Weatherford did not present “a situation where the State’s purpose was to learn what it could about the defendant’s defense plans.” Id. at 557. And the intrusion in Weatherford was justified by the “unfortunate necessity of undercover work,” which provides “value . . . to effective law enforcement.” Id. The Court also drew a distinction

53

between government intrusion “by an undercover agent” and accomplished through “electronic eavesdropping,” finding the former causes less “inhibition [to] free exchanges between defendant and counsel.” Id. at 554 n.4.

Another way to put these myriad limitations is the Weatherford Court required the defendant to show prejudice when Shillinger’s elements are absent. Recall, Shillinger applies only when the prosecution becomes privy to confidential attorney-client communications through intentional, unjustified intrusions. 70 F.3d at 1142. In Weatherford, the prosecution never became privy to the communications, and the intrusion was unintentional and justified. 429 U.S. at 556–58. And the intrusion into Mr. Hohn’s communications did not involve “an undercover agent,” but involved surreptitious overhearing, which the Weatherford Court explicitly called out as more concerning. Id. at 554 n.4. Nothing indicates Weatherford intended the defendant to show prejudice under the circumstances present in Shillinger.

In arguing otherwise, the majority seizes on Weatherford’s particular language that, if certain other facts were present, defendant “Bursey would have had only ‘a much stronger case’ in proving a Sixth Amendment violation.” Op. at 27 (quoting Weatherford, 429 U.S. at 554). As an initial matter, the critical word “only” does not appear in the Weatherford passage, meaning the majority seems to have assumed its own conclusion that Weatherford placed an upper limit on how much stronger the case would be. And, more

54

fundamentally, the majority opinion later clarifies that “stronger case” means “even then his case might not have been strong enough.” Op. at 36–37. But that does not follow. The phrase appears in a section in Weatherford disputing the Fourth Circuit’s logic and illustrating the unclear “contours of” “the Sixth Amendment right to counsel.” 429 U.S. at 553–54. The better reading is Mr. Bursey would have had a “stronger case” for establishing a Sixth Amendment violation, not for showing prejudice, if he could show the prosecution became privy to the communications, and the intrusion was intentional and unjustified. Under this reading and assuming those facts, Shillinger’s per se rule would indeed give him “a much stronger case.”

The majority then insists,

even in the worst cases, where the informant purposefully intrudes into confidential attorney-client conversations or where the informant relates those conversations to the prosecution, Weatherford still advises against assuming that the confidential information “has the potential for detriment to the defendant or benefit to the prosecutor’s case.”

Op. at 27–28 (quoting Weatherford, 429 U.S. at 557). Not exactly. Weatherford does not “advise against” anything. The Court just said it will not assume any information overheard by informants would be communicated to prosecutors, or, if communicated, that it would be “detriment[al].” Weatherford, 429 U.S. at

55

557. 24 The central point remains: in Weatherford, unlike in Shillinger or here, prosecutors never became privy to any attorney-client communications. The Court simply refused to assume otherwise. Weatherford does not say what doctrinal significance would attach to deeming communications to the prosecutor “detriment[al],” or even what constitutes “detriment”—that is, this passage is consistent with the notion that prosecutors becoming privy to legal confidences is necessarily “detriment[al].” Id.

Finally, the majority suggests Weatherford “general[ly]

repudiat[ed] . . . per se rules to protect attorney-client confidentiality.” Op. at 40; see also Op. at 36 (referring to “Weatherford’s holding that denounced per se Sixth Amendment rules against government intrusions”). This suggestion is similarly unpersuasive. At issue in Weatherford was a per se rule totally unlike the one in Shillinger. The Fourth Circuit’s rule mandated reversal and a new trial whenever the prosecution “arranges and permits” an informant to be present during the defense’s attorney-client communications. See

24 More specifically, the Court said,

Nor do we believe that federal or state prosecutors will be so prone to lie or the difficulties of proof will be so great that we must always assume not only that an informant communicates what he learns from an encounter with the defendant and his counsel but also that what he communicates has the potential for detriment to the defendant or benefit to the prosecutor’s case.

Weatherford, 429 U.S. at 556–57.

Weatherford, 429 U.S. at 549. Shillinger’s rule is nothing like that. A conclusive presumption of prejudice applies only in the narrow circumstances I have outlined above—circumstances notably absent in Weatherford. The Court’s holding that “the [Fourth Circuit’s] per se rule cuts much too broadly” says nothing about whether Shillinger’s rule, or any other similarly cabined per se rule, cuts too broadly. Id. at 557. Weatherford therefore does not stand for a blanket repudiation of structural error for prosecutorial intrusions.

b. The Majority Opinion’s Understanding of Morrison Next, the majority insists Morrison “reinforce[d] Weatherford’s prejudice requirement” in assessing Sixth Amendment violations. Op. at 28. Again, I am unconvinced Morrison even applies to Shillinger or to this case. In Morrison, as in Weatherford, Shillinger’s elements simply were not met. The Sixth Amendment violation assumed by the Court in Morrison involved no attorney- client confidences and no prosecutors becoming privy to surreptitiously obtained information between a lawyer and a client. See Morrison, 449 U.S. at 362–63. It is, once again, a different set of facts. That there was “no effect of a constitutional dimension which need[ed] to be purged” in that case, id. at 366, says nothing about whether that sort of “effect” existed in Shillinger or exists in Mr. Hohn’s case. Morrison necessarily left that question open. The majority’s understanding that Morrison “reinforce[d] Weatherford’s” supposed “prejudice requirement” in cases like Shillinger, Op. at 28, is thus incorrect. The Sixth

Amendment violation assumed in Morrison does not conflict with Shillinger’s heavily cabined rule and does not control this case.

In arguing otherwise, the majority summarizes, “[t]he [Morrison] Court considered that, once a ‘constitutional infringement [has been] identified,’ there must be some ‘threat[]’ of an ‘adverse effect upon the effectiveness of counsel’s representation’ or ‘some other prejudice to the defense’ to have a remediable Sixth Amendment claim.” Op. at 29 (third and fourth alterations in original) (quoting Morrison, 449 U.S. at 365). Even if this language applies to Shillinger’s entirely distinct facts, it takes us nowhere. For the reasons I give throughout, see supra sections II.A.1.d., II.A.2.c.; infra section III.B., intentional and unjustified intrusions into confidential attorney-client communications carry a true “‘threat[]’ of an ‘adverse effect upon the effectiveness of counsel’s representation,’” both in the intruded-on cases and beyond. By using the term “threat[],” the Morrison Court did not suggest the “adverse effects” had to be demonstrable by the defendant in the case at bar.

Further, the majority acknowledges the Morrison Court expressly “assume[d], without deciding, that the Sixth Amendment was violated.” Op. at 42 (alteration in original) (emphasis added) (quoting Morrison, 449 U.S. at 364). But the majority believes “the only sensible way to read” that statement is the Court meant it assumed an “intrusion,” not that it assumed a “violation.”

Op. at 42. I would instead believe the Court in Morrison meant what it said. 25 Thus, an essential difference remains: At issue in Shillinger and this case is whether a Sixth Amendment violation occurred and whether that violation merits a conclusive presumption of prejudice. At issue in Morrison was what remedy, if any, would be appropriate assuming a Sixth Amendment violation occurred. See 449 U.S. at 364. The latter inquiry is simply separate from the former.

The majority asks, “How could Morrison have presumed prejudice and then gone on to deny the defendant relief because she ‘demonstrated no prejudice’?” Op. at 42 (quoting Morrison, 449 U.S. at 366). That rhetorical question, though, is answerable. I see no reason the Court could not have presumed the prosecution’s misconduct was “inherently detrimental to [the

25 After all, the Court later discussed the remedy available given “[t]he

Sixth Amendment violation, if any,” United States v. Morrison, 449 U.S. 361, 366 (1981) (emphasis added)—not the intrusion, if any. The context in which the Court assumed a Sixth Amendment violation makes clear it was referring to a complete violation:

The United States initially urges that absent some showing of prejudice, there could be no Sixth Amendment violation to be remedied. Because we agree with the United States, however, that the dismissal of the indictment was error in any event, we shall assume, without deciding, that the Sixth Amendment was violated in the circumstances of this case.

Id. at 364. That is, Morrison assumed a violation in specific contrast to the United States’ argument that a Sixth Amendment violation necessarily requires a showing of prejudice.

defendant] . . . and threatened to subvert the adversary system of criminal justice,” Weatherford, 429 U.S. at 556, for purposes of establishing a Sixth Amendment violation, then subsequently inquired into the impact of the violation for purposes of appropriately tailoring the remedy. As Mr. Hohn correctly points out, “[w]hile structural error justifies a remedy regardless of harm to the defendant, what remedy is a separate question.” See Aplt. Supp. Br. at 14 n.1 (citing Waller, 467 U.S. at 49–50, in which the Court “agree[d]” that “the defendant should not be required to prove specific prejudice in order to obtain relief” for a structural error but noted “the remedy should be appropriate to the violation”). Case-specific prejudice could be both irrelevant to whether a Sixth Amendment violation occurred and relevant to whether “the drastic relief granted by the Court of Appeals”—dismissing the indictment with prejudice—is available. Morrison, 449 U.S. at 367.

The majority further posits, “[b]y directing us to calibrate the appropriate remedy from a defendant’s injury, Morrison presupposes that by the remedies stage some demonstration of prejudice has already occurred.” Op. at 43 (citing Morrison, 449 U.S. at 365). I cannot agree. The Morrison Court observed, unless “the constitutional infringement identified has had or threatens some adverse effect upon the effectiveness of counsel’s representation or has produced some other prejudice to the defense, . . . there is no basis for imposing a remedy in that proceeding.” 449 U.S. at 365 (emphasis

60

added). Again, Morrison invoked prejudice to decide what remedy is available, not whether a violation occurred. 26 Finally, the majority contends, “Morrison bolsters Weatherford’s prejudice requirement by reiterating that a ‘constitutional infringement’ under the Sixth Amendment requires ‘some adverse effect’ to the defendant . . . .” Op. at 44 (quoting Morrison, 449 U.S at 365). But, as I explained, Morrison “simply conclude[d] that the solution provided by the Court of Appeals”—dismissal of the indictment with prejudice—“is inappropriate where the violation, which [the Court] assume[d] ha[d] occurred, has had no adverse impact upon the criminal proceedings.” 449 U.S. at 367. By the Court’s own explicit limitation, Morrison articulated no requirement that a defendant prove prejudice to establish a Sixth Amendment violation.

26 Notably, this statement in Morrison suggests the showing at the remedies stage is not “a showing of actual prejudice,” see Op. at 38, but rather a showing that the intrusion “has had or threatens some adverse effect.” 449 U.S. at 365 (emphasis added). Thus, even if the majority were referring only to the remedies stage, its assertion that the Court requires “demonstration of prejudice” is incorrect. Op. at 43. As noted above, prejudice could be “threaten[ed].”

c. The Majority Opinion’s Understanding of Black, O’Brien, and Hoffa

Shillinger’s conclusive presumption also does not defy Black, O’Brien, or Hoffa.27 The majority acknowledges those cases but concludes they suggest a defendant must “tether governmental intrusion to a realistic possibility of injury from the use of confidential communications at trial.” Op. at 56. I understand them differently.

In Black and O’Brien, the Court remanded for a new trial without finding (or even discussing) prejudice. 385 U.S. at 29; 386 U.S. at 345. To that extent, as Mr. Hohn observes, those cases support the notion that a Sixth Amendment

27 The majority asserts Mr. Hohn has waived his argument that Black,

O’Brien, and Hoffa implied “a prosecutor who purposefully learns about the contents of confidential attorney-client communications commits a structural constitutional violation” by inadequately briefing it. Aplt. Supp. Br. at 22; see Op. at 56–57. Because Mr. Hohn “doesn’t address Weatherford’s unfavorable discussions of” those cases, the majority believes the argument was “inadequately presented” and thus “waived.” Op. at 56–57 (quoting United States v. Walker, 918 F.3d 1134, 1151 (10th Cir. 2019)).

I would not consider Mr. Hohn’s affirmative arguments under Black, O’Brien, and Hoffa inadequately briefed simply because he did not address one particular counterargument. See Walker, 918 F.3d at 1151 (clarifying it concerned arguments that were advanced in an opening brief “only in a perfunctory manner” (internal quotations omitted)). In any event, applying our prudential waiver doctrine, see In re Syngenta AG MIR 162 Corn Litig., 61 F.4th 1126, 1182 (10th Cir. 2023), seems unwise under the circumstances. There are, at most, five Supreme Court cases concerning government intrusions into attorney-client communications. We should fully engage with all reasonable arguments regarding each of them, especially because what is at stake is whether to overturn decades-old circuit precedent.

violation is established—and thus prejudice is presumed—in the context of prosecutorial intrusions into attorney-client trial preparations. See Aplt. Supp. Br. at 11 (citing Black, 385 U.S. at 27–29; O’Brien, 386 U.S at 345). The majority asserts, however, “Black, O’Brien, and Hoffa,” as expounded in Weatherford, suggest a defendant must show “a realistic possibility of injury from the use of confidential communications at trial.” Op. at 56 (citing its own discussion of Weatherford). To be sure, Black’s reasoning—relied upon by O’Brien—was that a new trial was warranted to “afford the petitioner an opportunity to protect himself from the use of evidence that might be otherwise inadmissible.” 385 U.S. at 28–29 (emphasis added). But nothing in this aside specifically concerns the use of the fruits of the intrusion at trial or that anything beyond a remedy determination hinged on that use.

Justice Harlan, dissenting in Black, explained the only justification he could imagine for the Court’s decision was “that any governmental activity of the kind here in question automatically vitiates, so as at least to require a new trial, any conviction occurring during the span of such activity.” 385 U.S. at 31 (Harlan, J., dissenting) (emphasis added). He did not interpret the Black majority as holding a petitioner must prove the intrusion prejudiced him to establish a constitutional violation. Neither do I.

And, contrary to the majority’s reasoning, Weatherford’s reading of Black and O’Brien is consistent with a suggestion that prejudice need not be shown

63

to establish a Sixth Amendment violation in this context. Contra Op. at 56. The Weatherford Court stated it “c[ould] not agree that these cases, individually or together, either require or suggest the rule announced by the Court of Appeals.” 429 U.S. at 551. “If anything is to be inferred from these two cases with respect to the right to counsel,” the Court continued, “it is that when conversations with counsel have been overheard, the constitutionality of the conviction depends on whether the overheard conversations have produced, directly or indirectly, any of the evidence offered at trial.” Id. at 552 (emphasis added). Because this reading was “a far cry from the per se rule announced by the Court of Appeals below,” the Court concluded those two cases did not support that rule. Id. But the fact that those cases did not support the Fourth Circuit’s broad per se rule says nothing about whether they conflict with Shillinger’s much narrower rule. 28 Finally, the majority, citing Weatherford’s discussion of Hoffa, asserts “the Supreme Court has never held that the Sixth Amendment right to attorney-client confidentiality ‘subsumes a right to be free from intrusion’ by government agents into the attorney-client relationship.” Op. at 14 (emphasis added) (quoting Weatherford, 429 U.S. at 553). First, Weatherford did not read

28 What is more, perhaps nothing at all is “to be inferred from these two

cases with respect to the right to counsel”; perhaps they were decided on “Fourth Amendment grounds,” not Sixth Amendment grounds. Weatherford, 429 U.S. at 551–52.

Hoffa to deny the Sixth Amendment subsumes such a right. It concluded Hoffa did not “furnish[] grounds” for the Fourth Circuit’s per se rule. Weatherford, 429 U.S. at 553–54. Hoffa, as Weatherford describes, had nothing at all to say on that topic. See id. at 553. Second, whatever Hoffa and Weatherford have to say about intrusion by “government agents” in general is beside the point. In Hoffa, as in Weatherford and Morrison, no prosecutor eavesdropped on attorney-client confidences. Because Hoffa did not address or concern whether a defendant must show prejudice to state a Sixth Amendment prosecutorial- intrusion claim, and because Weatherford’s interpretation appropriately indicated Hoffa’s limited import in this context, I do not agree Weatherford’s discussion “defeats” Shillinger’s structural-error rule. Contra Op. at 56.

3

The majority says “Shillinger’s holding contradicts those pronounced in Weatherford and its progeny”—i.e., Morrison—“because those cases affirm that, even when the prosecution becomes privy to attorney-client communications without a legitimate law-enforcement purpose, the defendant still must demonstrate a prejudicial use of the overheard information at trial.” Op. at 24 (citing Weatherford, 429 U.S. at 553–54). I disagree.

a. Shillinger’s Understanding of Weatherford Shillinger properly understood Weatherford. As Mr. Hohn correctly observes, Shillinger “discussed” the holding in Weatherford “at length.” Reply

Br. at 31. There, we observed the district court in Weatherford had made an “express finding that Weatherford communicated nothing at all to his superiors or to the prosecution about Bursey’s trial plans or about the upcoming trial” and that the Court had emphasized “the absence of purposefulness in the prosecutor’s intrusion and the legitimate law enforcement interests at stake.” Shillinger, 70 F.3d at 1139 (quoting Weatherford, 429 U.S. at 556). Then we concluded “the instant case presents a vastly different situation.” Id. at 1141. Unlike in Weatherford, we reasoned, “the intrusion here was not only intentional, but also lacked a legitimate law enforcement purpose” and “attorney-client communications were actually disclosed” to the prosecution. Id. at 1139, 1141. We therefore determined Weatherford did not preclude a conclusive presumption of prejudice, which was animated by concerns applicable to intentional, unjustified prosecutorial intrusions. See id. at 1141–42. These are the exact differences I described above.

But the majority insists Shillinger “misconstrued [Weatherford’s]

language . . . to circumvent [its] holding.” Op. at 36. Unlike in Shillinger, the majority argues, “the thrust of the [Weatherford] Court’s analysis focused on whether Bursey could show substantial detriment from the use of the confidential information at trial.” Op. at 37. But, as outlined above, Weatherford did not require use of the attorney-client communications to

66

establish a Sixth Amendment violation, much less use at trial specifically (versus, for example, to prepare for trial). Shillinger understood this distinction. See 70 F.3d at 1139 (explaining Weatherford focused not on use but on certain facts, comprising many of the Shillinger elements, that were conspicuously absent).

The majority then says Shillinger noted some distinctions I have described, then “inferred that when these conditions are flipped—when the government intrudes intentionally and without a legitimate law-enforcement purpose—prejudice must be presumed.” Op. at 38. True, but not without reason. Though it establishes mere intrusion by government agents does not constitute a Sixth Amendment violation, Weatherford does not reveal what is required in addition. Shillinger thus did not base its “infer[ence]” on Weatherford alone; it simply observed Weatherford did not control, then—for legitimate reasons described throughout Shillinger and this dissent—filled that gap by conclusively presuming prejudice. See 70 F.3d at 1139–40 (noting “commentators and courts have suggested,” when prosecutorial intrusions are intentional and unjustified, “such intrusions might not be wholly governed by the Weatherford decision” (emphasis added)); id. at 1141–42 (explaining why the conclusive presumption “best accounts for the competing interests at stake”).

b. Shillinger’s Understanding of Morrison Shillinger also correctly understood Morrison. Under Shillinger, the prejudice element of the constitutional violation is satisfied presumptively, 70 F.3d at 1142; then, under Morrison, the remedy is tailored to address the nature and extent of the intrusion’s impact, 449 U.S. at 364, 367. Shillinger understood and applied Morrison in just this way.

Shillinger acknowledged Morrison had “left open the question of whether intentional and unjustified intrusions upon the attorney-client relationship may violate the Sixth Amendment even absent proof of prejudice,” and acknowledged its holding “that even if the defendant’s Sixth Amendment rights were violated, dismissal of the indictment was an inappropriate remedy in that case.” 70 F.3d at 1140. Once the Shillinger panel determined the prosecution had violated the Sixth Amendment by intentionally intruding on Mr. Haworth’s attorney-client communications, it then turned to “ascertain[ing] the appropriate remedy,” applying Morrison. Id. at 1142 (acknowledging the remedy inquiry requires the court to “identify and then neutralize the taint [of the violation] by tailoring relief appropriate in the circumstances to assure the defendant the effective assistance of counsel and

68

a fair trial” (quoting Morrison, 449 U.S. at 365)). 29 The majority seems to agree this is a correct understanding of the analysis required by Morrison. See Op. at 41 (quoting the same language).

The majority insists “Shillinger wrongly interpreted Morrison as further proof that ‘Weatherford . . . does not necessarily govern intentional intrusions by the prosecution that lack a legitimate purpose.’” Op. at 44 (quoting Shillinger, 70 F.3d at 1140). “Rather,” the majority contends, “Morrison bolsters Weatherford’s prejudice requirement by reiterating that a ‘constitutional infringement’ under the Sixth Amendment requires ‘some adverse effect’ to the defendant . . . .” Op. at 44 (quoting Morrison, 449 U.S at 365). But, as I explained, the kind of Sixth Amendment violation assumed in Morrison simply did not control Shillinger, whose per se rule applies only when all of its elements are met. And Morrison was about remedies, in any event.

***

For these reasons, I submit Shillinger correctly understood and applied existing Supreme Court precedent when holding that intentional and unjustified prosecutorial intrusions into confidential pre-trial attorney-client

29 We ultimately “remand[ed] the case to the district court for factfinding

procedures to determine the extent of the intrusion as well as the proper remedy.” Shillinger, 70 F.3d at 1143.

communications constitute a per se Sixth Amendment violation. Shillinger, 70 F.3d at 1142.

C

Finally, before overturning longstanding precedent, especially in the face of egregious misconduct, I would apply stare decisis. The Supreme Court has called this doctrine “a foundation stone of the rule of law.” Michigan v. Bay Mills Indian Cmty., 572 U.S. 782, 798 (2014). Adherence to precedent “is the preferred course because it promotes the evenhanded, predictable, and consistent development of legal principles, fosters reliance on judicial decisions, and contributes to the actual and perceived integrity of the judicial process.” Id. (quoting Payne v. Tennessee, 501 U.S. 808, 827 (1991)). Although “not an inexorable command,” Payne, 501 U.S. at 828, stare decisis is “necessary to ensure that legal rules develop ‘in a principled and intelligible fashion,’” Bay Mills, 572 U.S. at 798 (quoting Vasquez, 474 U.S. at 265); see also United States v. Games-Perez, 695 F.3d 1104, 1116 n.16 (10th Cir. 2012) (noting “consideration of . . . the stabilizing influence of stare decisis [wa]s perfectly appropriate” when the precedent was “long-standing and firmly entrenched” and the court had raised the option of overruling the precedent sua sponte).

A decision to upend settled precedent “demands special justification.”

Arizona v. Rumsey, 467 U.S. 203, 212 (1984). 30 Such “special justification” might be the precedent was “poorly reasoned,” has “led to practical problems and abuse,” is “inconsistent with other . . . cases,” or is not sufficiently justified by “reliance interests.” Janus v. Am. Fed’n of State, Cnty., & Mun. Emps., Council 31, 585 U.S. 878, 886 (2018); 31 see Kimble v. Marvel Ent., LLC, 576 U.S. 446, 455 (2015) (“[I]t is not alone sufficient that we would decide a case differently now than we did then.”).

I cannot conclude the majority’s marked departure from our binding precedent is justified. As I have explained, Shillinger was not “poorly reasoned.” Janus, 585 U.S. at 886. The majority says “Shillinger is a twenty- nine-year-old case” and “is out of step with the Supreme Court’s cases on

30 As I have outlined, the Supreme Court follows these principles when

deciding whether to set aside its own precedents. And at least some of our sister circuits have followed these same principles in deciding whether to set aside their precedents. See, e.g., Riccio v. Sentry Credit, Inc., 954 F.3d 582, 590 (3d Cir. 2020) (describing the circuit’s adherence to the Supreme Court’s principles of stare decisis, and collecting similar cases from other circuits). Absent a contrary mandate from this circuit or an argument from any party to do otherwise, I too endorse the same principles the Supreme Court expounds.

31 In that case, the Court relied on all of those “strong reasons” combined

to overrule precedent in the First Amendment context. See Janus v. Am. Fed’n of State, Cnty., & Mun. Emps., Council 31, 585 U.S. 878, 886 (2018). It is not clear what combination and strength of the enumerated reasons might suffice to justify overruling precedent in any given case. Because none is present here, I need not attempt to answer this question.

structural error.” Op. at 4. It also quotes Greer, 593 U.S. at 513, for the proposition that only a “very limited class of cases” involve structural error. Op. at 4. 32 Yet I have shown how the Supreme Court’s structural-error jurisprudence since Shillinger has reinforced, not undermined, the viability of its reasoning and its structural-error rule. Shillinger thus fits comfortably within that “limited class.” That structural error involves a “limited class” of cases is a description, not an invitation to narrow the field further. See Weaver, 582 U.S. at 295 (acknowledging “the precise reason why the Court has deemed [an error] structural . . . varies in a significant way from error to error.”).

Nor has the conclusive presumption in Shillinger led to practical problems. A Sixth Amendment claim under Shillinger will be exceptionally rare—unless, as here, the prosecutorial misconduct affects over one hundred cases. The narrow decision in Shillinger is not to blame for the scope of misconduct; that blame lies solely with the prosecutors. And the potential

32 As I described, this type of prosecutorial intrusion is within the “limited class” of cases the Court has recognized as structural errors based on state interference with the right to counsel. See supra section II.A. Shillinger, contrary to the majority’s assertion, did “grapple with Cronic’s limited categories for recognizing structural error in Sixth Amendment right-tocounsel claims,” Op. at 44, when it discussed Ferguson, Brooks, Herring, and Geders as “cases in which state interference with the right to counsel has been held to violate the defendant’s Sixth Amendment rights per se,” and when it considered factors later described in Weaver. 70 F.3d at 1141.

availability of relief to defendants cannot be understood as an administrability problem.

Moreover, substantial reliance interests are necessarily at stake. Why would defense counsel and their clients ever have to think the adversary was listening without justification? The norms protected by Shillinger’s prophylactic rule meant the government should not, and in the normal course would not, intentionally and without justification become privy to confidential attorney-client communications. See generally Fed. Pub. Defs. Amicus Br. At the outset of the proceedings against Mr. Hohn, the conclusive presumption of prejudice in Shillinger applied to his case. And Mr. Hohn invoked it. See RI.2661. This helps explain why Mr. Hohn did not attempt to show prejudice in the district court—it was not a concession (as the majority mistakenly assumes) but reasonable reliance on long-settled precedent. 33 Apart from the reliance interests at stake, there are other important reasons to interrogate the wisdom of abrogating Shillinger in the context of the Kansas misconduct. As Mr. Hohn points out, “[o]verruling Shillinger’s structural-error rule would also condone the pattern of prosecutorial misconduct involved here.” Aplt. Supp. Br. at 18 (citing Weaver, 582 U.S. at 301). The majority’s “condemn[ation]” of “the Kansas USAO’s practice,” Op. at

33 I would at least not hold this litigation choice against Mr. Hohn and would thus not apply today’s ruling retroactively to pending cases.

73

50, while undoubtedly significant, is cold comfort to the defendants who may have suffered worse legal outcomes, through subtle and unmeasurable differences in the proceedings, because of the misconduct.

This misconduct occurred while Shillinger’s deterrence-focused rule was the law, so the need for deterrence is reinforced, not abated. And, of course, the government holds the power to ensure no defendant ever enjoys relief under Shillinger again. “If [intentional and unjustified prosecutorial intrusion into confidential attorney-client communications] becomes a thing of the past, no conviction will ever again be lost on account of it.” Vasquez, 474 U.S. at 262. As the district court found, “such governmental intrusions into defendants’ attorney-client relationships are easily prevented by the use of a taint team or other precautions,” RII.1781—measures the Kansas USAO has now implemented, see Aplee. Supp. Br. at 1 (noting it has “adopted ‘a comprehensive policy . . . that is largely curative of many of the issues that’ sparked the litigation” (quoting RI.2847)).

Finally, the Supreme Court has not overruled Shillinger or its structural-error rule. I have already shown Shillinger is consistent with the Supreme Court’s precedents. But as the district court recognized, there is a widely acknowledged circuit split over whether defendants must show prejudice to establish a Sixth Amendment violation when prosecutors wrongfully invade the attorney-client relationship. See RI.2878–79 (citing

74

Cutillo v. Cinelli, 485 U.S. 1037, 1037–38 (1988) (White, J., dissenting from the denial of certiorari), as “noting conflicting approaches within the Circuits in cases where the Sixth Amendment violation involves the transmission of confidential defense strategy information”); see also Wayne R. LaFave et al., 3 Criminal Procedure § 11.8(b) (4th ed.) (“[T]he Morrison opinion left open the possibility that the Court might adopt a per se standard for those state invasions of the lawyer-client relationship that are not supported by any legitimate state motivation. The federal lower courts have divided on this issue.”); Kaur, 141 S. Ct. at 6 (Sotomayor, J., statement respecting the denial of certiorari) (“Since Weatherford, many federal and state courts have struggled to define what burden, if any, a defendant must meet to demonstrate prejudice from a prosecutor’s wrongful or negligent acquisition of privileged information.” (emphasis added)); United States v. Mastroianni, 749 F.2d 900, 907 (1st Cir. 1984) (“[W]e have not previously had the opportunity to discuss what constitutes prejudice and who bears the burden of proving it under these circumstances. The circuit courts have thus far split on this issue.”).

Perhaps the Court should weigh in to resolve the matter, but it has not yet. 34 The majority opinion speculates how the Supreme Court would rule. But

34 “A majority of circuits either support or are consistent with our view,”

the majority opinion says, “that constitutional claims like Hohn’s require the defendant to show prejudice,” and such “prejudice accrues ‘only if the

75

intercepted communications are somehow used against the defendant . . . in connection with the underlying proceeding.’” Op. at 57 (alteration in original) (emphasis in original) (quoting ACLU Found. of S. Cal. v. Barr, 952 F.2d 457, 472 (D.C. Cir. 1991)). The cases the majority cites illustrate three of the four positions within the circuit split: those that rebuttably presume prejudice, those that require a defendant to prove prejudice to obtain certain stronger remedies, and those that always require the defendant to prove prejudice. See Op. at 58.

Without commenting on the majority opinion’s characterization of each case, I make four observations. First, as Judge Bacharach correctly explains, “the only circuits to address the allocation of the burden are the First and Ninth Circuits,” both of which “adopt a rebuttable presumption of prejudice.” Judge Bacharach’s Partial Dissent at 17. The majority’s claim that these cases “require the defendant to show prejudice,” Op. at 57 (emphasis added), is therefore not accurate.

Second, the majority omits from its list cases that conclusively presume prejudice, including those from the Third and Tenth Circuits. See Shillinger, 70 F.3d at 1142; United States v. Levy, 577 F.2d 200, 208–10 (3d Cir. 1978) (holding prejudice is presumed, as discussed above).

Third, the majority opinion also overlooks that one additional circuit has suggested a per se prejudice rule might apply when the intrusion is “manifestly and avowedly corrupt,” but it has not yet encountered such a case. See United States v. Gartner, 518 F.2d 633, 637 (2d Cir. 1975); accord United States v. Ginsberg, 758 F.2d 823, 833 (2d Cir. 1985) (“[T]o require a hearing on a claimed sixth amendment violation resulting from unintentional or justifiable presence of a government informant or agent at an attorney-client conference, a defendant must allege specific facts that indicate communication of privileged information to the prosecutor and prejudice resulting therefrom.” (emphasis added)).

Fourth, the majority opinion also does not discuss state courts’ stances, including one that takes the very position in Shillinger. See State v. Quattlebaum, 527 S.E.2d 105, 109 (S.C. 2000) (“Deliberate prosecutorial misconduct raises an irrebuttable presumption of prejudice.”).

76

we should “decline to read its tea leaves” for what the Court might say and instead “await word from the Court itself.” United States v. Wilson, 98 F.4th 1204, 1231 n.5 (10th Cir. 2024).

Under the circumstances, the majority’s decision sua sponte to abrogate Shillinger’s structural-error rule proactively is profoundly destabilizing. Under traditional notions of stare decisis, I would refuse to abandon the well-reasoned position our circuit has maintained for thirty years, unless and until the Supreme Court speaks to the contrary.

III

The majority’s new rule has serious problems. Judge Bacharach’s partial dissent ably identifies many, as I will highlight here. I will then show why his sound logic justifies a conclusive presumption of prejudice, not just a rebuttable one. 35

Thus, contrary to the majority’s view that “any decision from this court rejecting Shillinger’s structural-error rule would find good company among our fellow circuits,” Op. at 58, I would emphasize the circuit split identified above remains alive and well—as to both whether prejudice must be shown and who must show it.

35 To clarify, I read Judge Bacharach’s partial dissent as suggesting what

rule this circuit should adopt on the assumption that Shillinger’s conclusive presumption is to be abrogated. I therefore read it as taking no stance on whether Shillinger should, in fact, be abrogated—the focus of this dissent.

77

A

I begin my analysis by endorsing much of what Judge Bacharach writes in his well-stated partial dissent. My colleague persuasively shows “[t]he [kind of] intrusion [at issue in Shillinger] could prejudice the defendant in plea bargaining, jury selection, or the prosecution’s case-in-chief.” Judge Bacharach’s Partial Dissent at 7. He then asks, “But how could the defendant know if the prosecution had used the information for these purposes? The defendant has no way of knowing.” Judge Bacharach’s Partial Dissent at 7–8. Indeed, the information flows as a “one-way pipeline”: the prosecution gains information about the defense, but it gives none back. Judge Bacharach’s Partial Dissent at 9. Thus, it is “virtually impossible for Mr. Hohn to know whether the prosecution had used the improperly intercepted information.” Judge Bacharach’s Partial Dissent at 9.

Judge Bacharach also presents a helpful example of the subtle ways a trial may proceed differently because of intentional and unjustified prosecutorial intrusions on defendants’ confidential attorney-client communications:

[C]onsider what happens when the prosecution intercepts a defendant’s phone call with attorneys about their plans to impeach a government witness. . . . For example, knowledge of the defense strategy might lead the prosecution to elicit testimony about impeachment material to soften the sting of later cross-

examination. Or a brief call might disclose information about the attorneys’ tone or approach.

78

Judge Bacharach’s Partial Dissent at 13. I therefore agree “it’s hardly fair to require the defendant to show why the prosecution made its strategic decisions.” Judge Bacharach’s Partial Dissent at 12.

B

The sound logic of Judge Bacharach’s partial dissent necessarily extends further. It is “virtually impossible for Mr. Hohn to know whether the prosecution had used the improperly intercepted information.” Judge Bacharach’s Partial Dissent at 9. I would add that it is virtually impossible for anyone—even the prosecution itself—to know how that information actually shaped the trial.

When the government intentionally and unjustifiably becomes privy to a defendant’s confidential attorney-client communications, some effects on a trial are knowable in some cases. For instance, as Judge Bacharach correctly observes, listening to a conversation about impeaching a prosecutor’s witness “might lead the prosecution to elicit testimony about impeachment material to soften the sting of later cross-examination.” Judge Bacharach’s Partial Dissent at 13. The evidence that establishes that effect could be direct, such as contemporaneous notes or a recorded conversation about the government’s strategy, or circumstantial, such as an unusual amount of direct examination about the impeachment material.

79

If, however, that evidence establishes “the prosecution had already decided not to call the witness,” Judge Bacharach reasons “the interception might not be prejudicial.” Judge Bacharach’s Partial Dissent at 13. The key word is “might.” Returning to the witness example, even if the prosecutor had already decided not to call the witness, she may still learn the defense team’s overall approach to impeachment, as well as those attorneys’ overall demeanor, their apparent level of knowledge or competence, or even their rapport with the defendant. And it takes little more imagination to see how that information might shape the course of a trial, even in ways the prosecution might not consciously appreciate. Under traditional understandings of our adversarial system of criminal justice, the defendant should not assume the risk of that possibility. Because of the virtually infinite ways each criminal proceeding can progress, “[i]t is impossible to know what different choices the [prosecutor] would have made” were she not intruding, “and then to quantify the impact of those different choices on the outcome of the proceedings.” Gonzalez-Lopez, 548 U.S. at 150.

A rebuttable presumption of prejudice cannot address the fundamental issue that these effects are practically unmeasurable and unknowable. Even the government cannot know exactly how its decisions, tone, questions, writing, objections, and so on might have differed in the counterfactual world with no intrusion. And no one can know how those differences might have influenced

80

the course of the proceedings. Even if a rebuttable presumption would be a defensible alternative on the assumption Shillinger had to be overruled, the truth remains that the only standard that comports with this fundamental unknowability in a meaningful way is a conclusive presumption of prejudice, consistent with Shillinger.

IV

I would reaffirm the conclusive presumption in Shillinger and conclude the district court correctly found itself bound by it. Under these circumstances, I now reach the district court’s only reason for denying Mr. Hohn’s habeas petition. Notwithstanding its otherwise thorough analysis, the district court believed—mistakenly—that Mr. Hohn’s voluntary disclosure of his attorney- client communications to a third party (by consenting to recording or monitoring by CCA) meant the communication was not protected by the Sixth Amendment, or alternatively that Mr. Hohn waived his Sixth Amendment protections. See RII.1745–48; RII.1762–69. I agree with the majority that “Sixth Amendment attorney-client confidentiality is distinct from and broader than the attorney-client privilege.” Op. at 19. But the majority needed only “assume without deciding” that Mr. Hohn demonstrated confidentiality, because it found the absence of a showing of prejudice dispositive. Op. at 20. I would reverse because the district court added a privilege element not contemplated by the Sixth Amendment.

81

A

Recall, under Shillinger, a per se Sixth Amendment violation occurs when (1) there is a “confidential” attorney-client communication; (2) the government becomes “privy to” the communication; (3) it becomes privy because of a purposeful intrusion; and (4) the intrusion was not justified by any legitimate interest. See 70 F.3d at 1142. The parties do not appear to dispute the district court’s findings that the government became privy to the content of Mr. Hohn’s call with his attorney because of a purposeful intrusion that was not justified by any legitimate law enforcement interest. The dispositive inquiry was (and should still be) whether Mr. Hohn’s call with his lawyer was a “confidential” communication under the Sixth Amendment.

The district court understood “principles relating to the attorney-client privilege” to be “an appropriate framework for showing that the recordings between petitioner and counsel [were] protected communications under the Sixth Amendment.” RII.1763–64. Applying attorney-client-privilege principles, the court held the call was not a “confidential communication” because Mr. Hohn voluntarily disclosed it to a third party by consenting to recording or monitoring by CCA. See RII.1745–48 (making findings of fact regarding voluntary disclosure); RII.1765–66 (concluding Mr. Hohn did not have a “reasonable expectation of confidentiality” in the attorney-client call).

82

To support its conclusion that the attorney-client privilege did not protect the call, the district court highlighted these facts: (1) after Mr. Hohn “had studied and understood the Inmate Handbook and phone monitoring consent forms he signed, Hohn placed the call to Campbell from a CCA phone that he believed and understood was monitored and recorded”; (2) Mr. Hohn “testified that he believed and understood that his attorney-client calls were subject to recording by Securus and CCA”; (3) Mr. Hohn acknowledged he “consented to the monitoring and/or recording of his attorney-client calls”; and (4) Mr. Hohn acknowledged “he understood the procedure to except attorney- client calls from monitoring” but “never followed the procedure to make an unmonitored call.” RII.1765–66. The court concluded “this conduct [wa]s inconsistent with an objectively reasonable expectation of confidentiality in the attorney-client communications, and thus the attorney-client privilege and the Sixth Amendment right to confidential attorney-client communications did not attach to the April 23 call.” RII.1766.

B

I agree with the majority it is not appropriate to interpret the Sixth Amendment right to communicate with one’s counsel as limited by the attorney-client privilege. See Op. at 19. The government suggests we should extrapolate from a reference to the attorney-client privilege in a footnote in Weatherford to conclude “the Sixth Amendment’s protection of confidential

83

attorney-client communications goes hand-in-hand with the attorney-client privilege.” Ans. Br. at 32 (citing Weatherford, 429 U.S. at 554 n.4, which in turn cited Fisher v. United States, 425 U.S. 391, 403 (1976)—a case involving attorney-client privilege). In fact, the citation to attorney-client privilege principles in that footnote served only to explain the Court’s rejection of an argument by Mr. Weatherford that an intrusion by electronic surveillance should be treated the same as a physical intrusion by a government agent.

The footnote explained “one threat to the effective assistance of counsel”

posed by government intrusion in communications is the “inhibition of free exchanges” between defendant and counsel—a threat similarly addressed by the attorney-client privilege. Weatherford, 429 U.S. at 554 n.4. The Court then reasoned there is a greater chilling effect from possible (undetectable) surveillance than there is from a physical third party who can be observed and definitively excluded, including by citing Fisher. Id. Weatherford’s citation to Fisher, an attorney-client-privilege case, thus did no more than acknowledge an overlapping purpose of attorney-client privilege and the Sixth Amendment right to counsel; it did not imply the two are coextensive or privilege principles govern Sixth Amendment analyses. Overlapping purposes are not even sufficient to support the argument that the right and the privilege go “hand- in-hand,” Ans. Br. at 32, let alone to establish the former is limited by the latter.

84

Moreover, the facts of Shillinger foreclose the reading of “confidential communication” the government urges and the district court accepted. It cannot be true Shillinger created a rule limiting the Sixth Amendment’s protections to privileged communications yet granted relief in a case involving unprivileged communications. See 70 F.3d at 1134 (explaining trial preparation sessions took place in the presence of a deputy sheriff). Even if it were possible the communications in Shillinger were privileged (for example, if Mr. Haworth reasonably believed the third party was a member of the defense team or took reasonable steps to keep the communications private), we could not have affirmatively held Mr. Haworth carried his burden of proving a Sixth Amendment violation without addressing one of the elements. See Reply Br. at 9 (making this point). Therefore, it follows privilege is not an element of a Sixth Amendment claim under Shillinger. Contra Ans. Br. at 28 (arguing the court should decline to give persuasive value to Shillinger’s “silence” on whether attorney-client privilege principles govern, because that question merely “lurk[ed] in the record” and was not ruled upon (quoting Webster v. Fall, 266 U.S. 507, 511 (1925))).

The absence of a privilege element is not a “lurking” question; it is a necessary corollary of the holding. Thus, the district court mistakenly grafted an additional attorney-client-privilege requirement onto the elements of a claim under Shillinger.

85

C

So what did Shillinger mean by “confidential communications?”

The government proposes “confidential communications” are privileged communications, or at least communications as to which the defense has a “reasonable expectation of confidentiality,” where confidentiality is defined by attorney-client-privilege law. See Aplee. Supp. Br. at 19 (“The Sixth Amendment’s protection against government intrusion into attorney-client communications generally applies only to privileged communications, and certainly does not protect communications in which a defendant has no reasonable expectation of confidentiality.”); Ans. Br. at 31 (arguing privilege principles “should be the starting point for determining whether a defendant has satisfied the confidential communications element of a Sixth Amendment intentional-intrusion claim”). In the government’s view, “Weatherford provides a blueprint for this approach.” Aplee. Supp. Br. at 21. For this proposition, the government relies on (1) the already-addressed reference in footnote 4 of Weatherford to Fisher, a case about attorney-client privilege, and (2) a sentence in the same footnote describing that defendants might be able to “exclude[e] third parties from defense meetings or refrain[] from divulging defense strategy when third parties are present at those meetings.” Aplee. Supp. Br. at 21 (citing Weatherford, 429 U.S. at 554 n.4).

86

I am unpersuaded. Weatherford likely forecloses the government’s argument. The Court specifically rejected the notion that for Sixth Amendment purposes,

whenever a defendant converses with his counsel in the presence of a third party thought to be a confederate and ally, the defendant assumes the risk and cannot complain if the third party turns out to be an informer for the government who has reported on the conversations to the prosecution . . . .

429 U.S. at 554. There, the Court seems to acknowledge, the Sixth Amendment would apply even though the defendant consented to the presence of a third (non-adversarial) party. Thus, the government’s suggestion to define “confidential communications” by the “reasonable expectation of confidentiality,” in turn defined by privilege law, is unavailing.

Mr. Hohn’s suggested definition also does not seem quite right. Mr. Hohn believes “confidential communications” means any “substantive attorney- client communications”—that is, communications concerning “legal advice or strategy.” Op. Br. at 37, 47. He derives this definition from the facts of Shillinger, arguing, “after all, ‘substantive’ is one thing that the communications in Shillinger actually were.” Op. Br. at 47 (citing Shillinger, 70 F.3d at 1137). He also proposes communications can be considered “inherently” confidential in a constitutional sense even if they are not confidential in an “evidentiary” sense. Op. Br. at 47.

87

Under this definition, how far this constitutional confidentiality should stretch is an important question. Drawing the line too broadly could sweep in conversations with no legitimate reason for constitutional protection. But drawing the line too narrowly risks not protecting attorney-client conversations that could provide useful—though perhaps quite subtle—clues to the prosecution. And this lack of clarity may make defendants and attorneys hesitant to speak openly.

This court has never decided whether Shillinger’s rule sweeps in seemingly mundane conversations, as when “all the attorney says to the defendant is, ‘Hello, how are you? When are you available to meet?’” Op. at 48. Contrary to the majority’s suggestion, though, I do not see intruding on such conversations as clearly innocuous. With the specter of the adversary eavesdropping, a defendant who may wish to respond, “I am worried about what I have done and must meet immediately,” may instead respond, “I am fine and can meet anytime.” Prosecutors may pick up on subtler hints from these conversations. What if the prosecutor believes any competent defense attorney would ask a certain question during a call, but the defense attorney instead simply says “Hello, how are you?”, with no attention to that question? It would be unsurprising if the prosecution’s strategy shifts in light of its impressions of the defense lawyer. Or the prosecution can pick up clues from tones of voice. “And then we would have to speculate upon what effect those

88

different choices or different intangibles might have had.” Gonzalez-Lopez, 548 U.S. at 151. 36 But drawing that line precisely is unnecessary to resolving this appeal.

Given how closely the facts here hew to those in Shillinger, we need not define “confidential” in this case. The district court “confirmed that the communication” between Mr. Hohn and his attorney “involved legal advice or strategy.” RII.1760. Shillinger undeniably reached at least that far.

V

There was no reason to revisit Shillinger. But, having done so sua sponte, we should have reaffirmed its conclusive presumption of prejudice. The district court correctly understood and applied that conclusive presumption and erred only by adding a privilege element to the Sixth Amendment violation recognized in Shillinger. Under a proper reading of Shillinger, Mr. Hohn’s § 2255 motion should be granted. I would reverse the district court’s contrary conclusion and remand for a determination of the appropriate remedy.

36 In any event, as the majority recognizes, this case does not provide an

occasion to decide whether Shillinger would reach these purportedly “harmless subjects.” Op. at 48.