Walker v. United States

• Court: District of Columbia Court of Appeals
• Decided: June 20, 2024
• Docket: 23-CO-0417
• Status: Published

Opinion

Notice: This opinion is subject to formal revision before publication in the Atlantic and Maryland Reporters. Users are requested to notify the Clerk of the Court of any formal errors so that corrections may be made before the bound volumes go to press.

DISTRICT OF COLUMBIA COURT OF APPEALS

No. 23-CO-0417

ALPHONSO J. WALKER, APPELLANT,

V.

UNITED STATES, APPELLEE.

Appeal from the Superior Court of the District of Columbia (2018-CF1-019051)

(Hon. Rainey R. Brandt, Trial Judge)

(Argued October 17, 2023 Decided June 20, 2024)

Daniel Gonen, Public Defender Service, with whom Samia Fam, Public Defender Service, was on the brief, for appellant.

Kevin Birney, Assistant United States Attorney, with whom Matthew M. Graves, United States Attorney, and Chrisellen R. Kolb, John P. Mannarino, Natalie Hynum, and Emma McArthur, Assistant United States Attorneys, were on the brief for appellee.

Before DEAHL, HOWARD, and SHANKER, Associate Judges.

DEAHL, Associate Judge: The Fifth Amendment’s Double Jeopardy Clause

protects criminal defendants from successive prosecutions for the same offense. A

critical component of that protection is the defendant’s right to have his case decided 2

by the first jury sworn in to hear it. That right is the focus of this appeal where, two-

and-a-half weeks into Alphonso Walker’s trial for a double homicide, prosecutors

introduced what all parties now agree was inadmissible and highly prejudicial

hearsay, effectively telling jurors (over defense objection) that “everyone was

saying” Walker committed the murders. The parties debated a variety of remedial

measures to address that evidentiary error, but neither Walker nor the government

requested a mistrial without prejudice, which would prompt a retrial. Walker

stressed that a retrial was the last thing he wanted because he believed he was on the

road to acquittal despite the improper prejudice admitted against him. The trial court

nonetheless declared a mistrial over his objections, reasoning that “all that does is

put everybody back at the starting point” as if “this little gaffe didn’t happen.”

We agree with Walker that the trial court erred in declaring a mistrial over his

objections. A mistrial cannot be declared over defense objection unless “manifest

necessity” requires it. Douglas v. United States, 488 A.2d 121, 132 (D.C. 1985).

There simply was no manifest necessity here, both because the case did not present

any extraordinary circumstances that precluded Walker’s trial from proceeding, and

because there were reasonable alternatives to a mistrial that might have sufficiently

cured the prejudice. When mid-trial “prosecutorial error” like the one here

prejudices a defendant, our precedents show an unflinching commitment to the

principle that the defendant must “retain primary control over the course to be 3

followed.” See Oregon v. Kennedy, 456 U.S. 667, 676 (1982) (quoting United States

v. Dinitz, 424 U.S. 600, 609 (1976)). This is true even where the “prosecutorial error

[is] of a degree sufficient to warrant a mistrial” at the defendant’s request. Id. at

676.

That is because a prosecutor’s blunder is generally not a good justification for

depriving a defendant of his constitutional right to “conclude his confrontation with

society” before a jury that he believes to be “favorably disposed to his fate.”

Douglas, 488 A.2d at 130 (quoting United States v. Jorn, 400 U.S. 470, 486 (1971)

(plurality opinion)). If the defendant would rather proceed, despite the unfair

prejudice the prosecution introduced against him, there is no cogent basis for

denying him that right. That is particularly clear once a trial is in its evidentiary

phase. A defendant’s interest “in retaining a chosen jury may intensify as the trial

proceeds,” and where “the defendant senses the trial is going well and the factfinder

is leaning toward acquittal,” stripping him of that jury because of prosecutorial error

accords with no sense of fairness or justice. Id. Because there was no manifest

necessity for declaring a mistrial over Walker’s objections, his indictment must be

dismissed and his retrial is barred. 4

I. Factual and Procedural Background

The Murders

Dalante “Bebe” Wilson and Antone “Mini Mo” Brown were murdered one

night in 2018. Bebe ran a trap house, where he and others stored, dealt, and used

illegal drugs; Mini Mo was one of his friends and a regular at the trap house. There

was no dispute that a gunman came into the trap house that night and, after Mini Mo

apparently reached for a gun, shot Mini Mo in the head and killed him. The gunman

then demanded that Bebe give him all his “money and [his] drugs,” “give me

everything.” When Bebe apparently failed to pony up, the gunman shot him in the

head as well. The only question at trial was who did it.

The government posited that Walker did it. Its theory was that Walker had

been pistol-whipped up the block from the trap house earlier that night by someone

named Marcus, and Walker then broke into the trap house and killed Bebe and Mini

Mo as revenge. Marcus was not affiliated with the trap house or its occupants, so it

is fair to say that the government’s motive theory was pretty tenuous. But the

government had three central witnesses who would point the finger at Walker. There

was James “White Boy” Morrow, who was the only person in the living room with

Bebe and Mini Mo when the gunman came in, and he claimed that Walker was the

gunman. There was Jackie Taper, who was in one of the trap house’s rooms and 5

testified that she saw Walker in the house during the home invasion. Then there was

Juanita “Gangster” Borum, who said that Walker confessed to her mere hours after

the killings.

The defense countered with a third-party perpetrator defense, positing that the

real culprits were White Boy and a rival drug dealer named Alonzo Williams, whom

White Boy let inside the trap house to commit the offenses. White Boy was a clear

suspect himself; he was seen apologizing to Bebe as he stripped the dying man of

his belongings and he was spotted later that night, still covered in Bebe and Mini

Mo’s blood, laughing and smiling alongside Williams. As for the government’s

other key witnesses, Taper was a convicted perjurer who had previously falsely

identified somebody in court for the sake of protecting her drug dealer (and Williams

was one of her drug dealers). And Borum was a sex worker and a longtime drug

addict whose drugs of choice included “a little bit of everything.” While she claimed

that Walker confessed to her from inside of his truck mere hours after the murders,

it turns out that Walker’s truck (a work vehicle) was equipped with a GPS device.

Its GPS data showed that Walker was never in the District when he purportedly

confessed to Borum, but was instead traveling around Maryland in the hours

surrounding when Borum said he had confessed.

That is all just preamble. Because the state of the evidence is relevant to

Walker’s double jeopardy claim, we now go into greater detail. 6

The Government’s Case

Over the course of two-and-a-half weeks, a jury heard nine days of evidence

and twenty-six government witnesses (out of an expected forty) before a mistrial

was ultimately declared. Of the government’s central witnesses mentioned above,

only Taper had completed her testimony, whereas Borum was in the middle of her

direct testimony, and White Boy was still waiting in the wings when a mistrial was

declared.

The government witnesses who set the table for what happened in the trap

house that night were Taper, Mary “Lovey” Linder, and Lenny Watson. That trio

was in Lovey’s room watching Wheel of Fortune and preparing to get high on heroin

and marijuana when they heard nearby gunshots. Lovey turned off the lights in her

room and the three hunkered down. They then heard somebody from inside the

house saying that he had been “watching” the house “all day” and knew there was

“shit in this house” because Bebe had not left, i.e., the day’s proceeds had to still be

there. At first Watson thought Mini Mo was robbing Bebe, because he believed they

were the only other two people in the house, but eventually he realized the voice was

not Mini Mo’s. The gunman eventually kicked in the door to Lovey’s room and

briefly entered with Bebe. Lovey and Lenny kept their heads down, but Taper

claimed to get a look at the shooter at this point. The gunman kept demanding money

and drugs, while Bebe kept saying he did not have anything. Bebe and the gunman 7

then left the room, and after the gunman’s continued demands to give him the “shit,”

and Bebe’s repeated protests that he did not have anything, the gunman shot Bebe

in the head.

Taper was the first to exit Lovey’s room once things quieted down, and she

saw both Bebe and Mini Mo either dead or dying in the trap house’s living room. It

appeared that Bebe was still breathing so Taper went to check on him, at which point

White Boy—a regular at the trap house—“climb[ed] over [Taper’s] back” in order

to steal a necklace off of Bebe’s neck. Taper assumed White Boy had been in the

house the whole time because “as soon as [she] came out of the room” he started

stealing Bebe’s belongings. In addition to taking the necklace, White Boy also

grabbed Bebe’s cell phones and PlayStation while, according to Taper, “he kept

saying he was sorry”: “I’m sorry, Bebe; I’m sorry; I’m sorry,” though Taper “didn’t

understand why he was saying he was sorry.”

As noted, Taper was the one witness from Lovey’s room who claimed that

she could identify Walker as the killer, but as the government fairly puts it in its

brief, she was “thoroughly impeached” by the defense. Like the others in Lovey’s

room, Taper maintained for years that she could not identify who killed Bebe and

Mini Mo. But during the trial—after she learned of a $50,000 reward in the case—

she took the government by surprise when she testified that she could positively

identify Walker as the killer. When pressed about why her story had suddenly 8

changed, she said that she had simply lied to both the police and the grand jury, but

added, “Today, I’m not lying though.” She seemingly contradicted herself later in

her testimony when she stated that she was in fact telling the truth when she told

police she could not identify the shooter and she admitted that she did not recognize

Walker when she saw him just two days after the shootings.

Perhaps most damning for Taper’s credibility was that she had a prior perjury

conviction for falsely identifying a person in court (1) in order to protect a drug

dealer who actually committed the crime, and (2) after the real culprit’s girlfriend

offered her just $300 to falsely implicate a scapegoat. And Williams, who the

defense posited was the real killer, was one of Taper’s drug dealers; plus, Taper

admitted that she asked the government for drug money multiple times in the run-up

to her testimony (even putting the $50,000 reward aside). The government was

caught by surprise when Taper’s perjury conviction was admitted into evidence, as

it believed the conviction was too old to come in as impeachment evidence, though

it ultimately recognized that it was admissible. Taper had also been arrested for

lying to the police in yet another case entirely unrelated to the one that led to her

perjury conviction. And Taper’s demeanor was so bizarre that the judge at one point

described her as “catatonic” and speculated that she might have gotten high during 9

a break in her testimony. 1 She had multiple outbursts, telling defense counsel, “You

funny” (he was not being funny), answering a question with “Oh, my God. Can

[defense counsel] get the fuck out of here?” and interrupting another question by

exclaiming “Do I have to sit up here for this shit?”

White Boy was the government’s only other witness who claimed to have seen

Walker in the trap house. Unlike Taper, White Boy had not yet testified, so we do

not know the full extent of either his testimony or the impeachment against him, but

the parties’ proffers provide a rough sketch of how it would have gone. White Boy

was critical to establishing Walker’s motive, because he was the only government

witness expected to testify to seeing Walker and Marcus’s altercation up the block

from the trap house some ninety minutes before the murders. Walker did not dispute

that he got into it with Marcus, but he instead offered that dust up as a reason why

White Boy would choose him as a scapegoat—people saw him in a dispute in the

general vicinity of the trap house not too long before the murders. In White Boy’s

1 The government points out that Taper’s identification of Walker came during her direct examination, when there was no specter of her being “under the influence,” as the trial court described her after the break. But the defense’s attack on Taper was not that she identified Walker because she was high on the stand; the defense posited that she falsely implicated him because there was some money in it for her, as with her prior perjury conviction. That she might have gotten high mid-testimony, and her overall demeanor, were broader attacks on her credibility as a witness and showed how deep her drug addiction ran, suggesting that she was the kind of person who might falsely implicate another for a little drug money. 10

telling of that earlier altercation, Walker was complaining that he had paid money

for a prostitute without getting his end of the bargain. Marcus eventually quieted

Walker down by pistol-whipping him, and then Marcus apparently walked in the

general direction of the trap house. There was no indication that Marcus was

affiliated with the trap house or any of its occupants, nor was there any suggestion

that Marcus had ever been inside of the trap house.

White Boy was also expected to testify that he saw Walker come into the trap

house and commit the murders, but he had his own rather glaring credibility

problems, even before being cross-examined. Multiple witnesses saw White Boy

robbing Bebe of his jewelry, cell phones, and PlayStation in the immediate aftermath

of the murders, apologizing to him all the while. Neighbors who saw White Boy

leaving the scene of the murders with Bebe’s belongings asked him what happened,

and he told them that two men came into the house and shot it up (consistent with

Walker’s third-party perpetrator defense, but not with White Boy’s later story that

Walker was a lone gunman). White Boy’s ex-girlfriend, Amber Keener, testified

that she saw him shortly after the shooting, and he was covered in what he admitted

was Bebe and Mini Mo’s blood, hanging out with Williams as both men were

“laughing and smiling” in an “odd” way. White Boy also posted pictures of himself

on Facebook wearing Bebe’s stolen jewelry. White Boy then left the District the

day after the murders as rumors predictably swirled that he was involved in them. 11

And, like Taper, it was nearly five years after the murders when White Boy first told

authorities that he could identify Walker as the killer, and that was only after he

independently researched the ongoing investigation into the murders and admittedly

reviewed the legal filings in Walker’s case.

One other government witness worth discussing—aside from Borum, whom

we get to in a moment—is the government’s ballistics expert. That expert testified

about a gun recovered from Walker’s apartment eight months after the murders. He

testified that there was “strong support” that the gun fired the cartridges and bullets

found at the crime scene, and there was “weak support” that a different gun fired

them. But as the trial judge opined, the ballistics expert was subject to a “hard cross-

examination,” in which he confirmed he was just “eyeballing it,” conceded that his

conclusions were “not based on any statistically-derived measurement,” and

acknowledged that there was no “generally accepted way of measuring the weight”

of any of his conclusions.

To quote the government’s brief, the sum total of its evidence through its first

twenty-six witnesses—Taper and the firearms examiner included—amounted to

“scant evidence that Walker had committed the murders.” 12

The Defense Theory

The defense did not have an opportunity to present its case, but its theory was

coming into focus before the trial court declared a mistrial. The defense posited that

White Boy let a rival drug dealer, Alonzo Williams, into the trap house for a heist-

turned-murder. White Boy then implicated Walker in order to deflect attention from

himself—an obvious suspect.

The government’s own witnesses lent considerable support to that defense.

The government had no explanation why, for instance, Walker would exact revenge

against Marcus by targeting a trap house full of individuals who were not affiliated

with Marcus in any way. And the trio inside of Lovey’s room did not describe

anything that sounded like a revenge killing, but instead sounded very much like a

plotted heist. Recall that those witnesses testified consistently that the gunman came

into the house announcing that he had been “watching” the house “all day” so that

he knew there was money and drugs in the house, and he then walked Bebe around

the house looking for a stash of cash and drugs. As Taper told the police on the night

of the murders, all signs were that “somebody set up Bebe to be robbed,” consistent

with the defense’s theory, but not the government’s.

The defense further stressed that this had to be an inside job because the

government’s witnesses consistently testified that the trap house was kept 13

fastidiously secured—as one might expect given the business transacted inside—so

that one of the house’s occupants had to let the killer in. The front door was boarded

up and the back door was always kept locked; Lenny, who was among the trio in

Lovey’s room, testified to locking it “just a little while” before the shootings. Yet

there were no signs of forced entry. And there was not even a key to that back door,

so whenever somebody wanted to get inside, they had to be granted entrance from

someone inside the house. That someone inside the house, the defense theory went,

was White Boy; the person he let inside was Williams, a rival drug dealer.

Beyond that, the defense hammered the credibility of the government’s key

witnesses, and we have already detailed the defense attacks on White Boy and Taper.

As for Borum, whom we turn to now, she had her own set of issues.

Borum’s Aborted Testimony

Juanita “Gangster” Borum was not far into her direct examination when the

government elicited the hearsay that ultimately led the court to declare a mistrial.

Although her testimony was still nascent, the parties’ proffers provide some sense

of what her testimony and the impeachment against her would have been. Borum

was a homeless sex worker and drug addict who counted both Walker and Williams

among her regular clients, whom she referred to as “dates.” She was going to testify

that several hours after the murders, Walker met her in the District and she got into 14

his truck, where Walker confessed to her. The government also planned to introduce

a so-called “bra recording,” which was a recording of Borum telling one of her

friends the same thing she would testify to: that Walker confessed to her. 2

The defense’s most targeted counterpunch to Borum’s testimony, aside from

her general credibility issues, was GPS evidence showing that Walker’s truck was

in fact not even in the District when Borum claimed he had confessed. The

government has offered no indication of how it would have countered that evidence.

The defense would further stress that Borum was a heavy drug user—she used “a

little bit of everything,” “crack cocaine, marijuana, heroin, PCP . . . Percocets and

Oxy”—and that she was close with Williams. Borum admitted that she was friends

with and “grew up” with Williams, and that he was one of her regular “dates,” or as

she put it, they “solicited together.” Though Borum claimed not to know that

Williams was a drug dealer. 3

2 The friend said she hid an iPhone in her bra and secretly recorded Borum recounting the supposed confession. The trial court ruled that the government could admit the bra recording over the defense’s hearsay objections, and the parties do not wrangle over that ruling in this appeal. 3 It is not disputed that Williams was in fact a drug dealer, as multiple government witnesses testified to that fact, including Taper and Keener, while White Boy was expected to say the same. Borum, despite her drug addictions and relationship with Williams, was the only government witness who claimed ignorance about that. 15

The Prejudicial Hearsay that Led to a Mistrial

Now comes the part that prompted a mistrial. During Borum’s testimony, the

government asked why she went to meet Walker in the area of the murders in the

hours after the murders. Borum replied that she had gotten “a phone call from

somebody saying, your date just killed—” at which point the defense objected.

“State of mind,” the government responded—an exception to the rule against

hearsay, see Fed. R. Evid. 803(3) 4—and the court overruled the defense objection.

Defense counsel immediately interrupted to press his objection, but the court

permitted the government to direct Borum to answer the question without further

discussion. Borum continued that her friend Wobbie had called her and told her

“your date just killed Bebe and Mini Mo; you need to get down here now.”

Shortly thereafter, the government began quoting a portion of Borum’s grand

jury testimony, reading in front of the jury: “Where was it that night where everyone

was saying, ‘your date did this; your date did this.’” The defense objected again and

moved to dismiss the case with prejudice, arguing that the government introduced

inadmissible hearsay to goad a mistrial because it was unhappy with how its case

4 The government seems to have meant something more typically phrased as “not for the truth” or “effect on the listener,” as it would later argue that it elicited this statement to explain why Borum went to meet Walker, not for its truth. The parties now agree that this was inadmissible hearsay that should have been precluded because it was far more prejudicial than the probative force of explaining why Borum went to meet Walker that night, a seeming irrelevancy in the case. 16

was going. “The Government doesn’t like how their firearms [expert’s] cross went.

They don’t like how Jackie Taper’s cross went, and so they’re doing things that

create a mistrial.” The defense was careful to clarify that it was seeking only a

dismissal with prejudice, so that a retrial would be barred, because Walker “has a

right for this jury to rule on the evidence” and the government had just grossly (and,

the defense argued, purposely) tainted the jury.

The court then seemed to realize its error in permitting the hearsay after some

further discussion. The court queried why the government had never raised these

statements in the extensive pretrial litigation covering the admissibility of

out-of-court statements at trial. “[G]iven the countless hours that we spent litigating

everything that was anticipated to come out of Ms. Borum’s mouth,” the court

continued, “I’m not exactly sure . . . how this missed that exercise.” The court

opined that the government “knew [what it was] doing”; it had “litigated everything

else in this case ad nauseam” but kept these particular statements in its back pocket

“as an opportunity to give that left/right punch across the bow.” Still, the court was

clear from the start that the defense’s request to dismiss the case with prejudice was

“off the table.” It acknowledged that there was a “level of subtle prejudice [or]

taint . . . that needs to be erased here,” but the court noted that it could “fix it” by,

“at minimum, . . . just strik[ing] that whole line of testimony from the trial.” After

extended discussion, the court ultimately dismissed the jury for the day so that the 17

parties could continue discussing how to properly remedy the error the following

morning.

The Court Declares a Mistrial

When the case was recalled the next morning, the judge expressly agreed with

defense counsel that the objectionable statements were inadmissible and that the

government should not have put them before the jury. If the government had

included the statements in the extensive pretrial litigation about the admissibility of

out-of-court statements, the court noted that it would have “exclude[d] [them]

because [they’re] too prejudicial.” The court also said that, while it did not agree

with the defense that the prosecution was purposefully goading a mistrial, it

recognized that “this gaffe has been created by the government.”

The parties then set forth their respective positions about how to move

forward. The defense offered two acceptable options, and ranked what it viewed as

unacceptable options in order of preference: (1) it maintained that a dismissal with

prejudice was the appropriate remedy; (2) barring that, it asked the court to strike

“Borum’s testimony in its entirety, with an instruction that her testimony is being

struck due to the government’s improper conduct in presenting her testimony”;

(3) while the defense argued forcefully that merely striking the problematic portion

of Borum’s testimony with a curative instruction was too weak a remedy, it also 18

made clear that it preferred that to the last option; (4) the last thing the defense

wanted was a mistrial without prejudice, which it described as putting Walker “in a

far worse place than” simply moving forward without remedial action. The

government preferred option number three. It argued that Borum’s problematic

statements should simply be stricken from the record with an instruction telling the

jurors to disregard them.

The court was initially “in the middle” of options two and three, opining that

a mistrial with prejudice was too severe a remedy while the government’s proposal

of simply striking the problematic testimony was “too weak.” With the viable

options seemingly narrowed to striking Borum’s testimony in its entirety and

striking the problematic testimony with a stronger curative instruction, the court was

concerned that striking Borum entirely would be “a huge windfall for” Walker,

because then the jury would never hear of Walker’s alleged confession. And while

the court was similarly concerned that the government’s curative instruction was too

weak, it opined that a stronger curative instruction could work. After a break, the

government proposed a stronger instruction that included striking Borum’s

testimony about the phone call with Wobbie, telling jurors to disregard it, and

instructing jurors that “Borum’s entire testimony should be considered with caution

and scrutinized with care.” The defense was not placated and continued to insist on 19

striking Borum as a witness entirely, while making clear its view that outright

dismissal with prejudice was warranted.

By this point, the defense’s continued insistence that a curative instruction

was insufficient seemed to exasperate the court into suggesting a mistrial without

prejudice so that the trial could start anew: “[Y]ou didn’t even offer up a straight up

mistrial as an option here. You went from . . . dismissal with prejudice down to,

let’s strike Ms. Borum because we want this jury to hear this evidence.” The defense

again explained that it did not want a mistrial without prejudice, pointing out that a

retrial “would compound the problem” and “put Mr. Walker in a far worse position,

because Mr. Walker has a right for this jury to [deliver] the verdict.” The defense

elaborated that the prosecution now had a preview of the defense case—from its

third-party perpetrator defense, which it did not disclose pretrial, to its potent

impeachment of government witnesses—so that a retrial would simply give the

government a chance to shore up its weaknesses. The prosecution countered that the

defense could not coherently argue that the jury would abide an instruction striking

the entirety of Borum’s still nascent testimony, while at the same time arguing that

the jury would not follow a more targeted curative instruction simply striking the

particularly problematic testimony. The court agreed with that sentiment, saying

that “either the jury is presumed to follow the Court’s instruction[s] in their entirety,

or they’re not.” The court said that it believed “an appropriate curative instruction” 20

would solve the problem, and issued an ultimatum: When it returned from (another)

recess, “the Court is either going to declare a mistrial or the Court is going to have

in hand a curative instruction to give to this jury for Ms. Borum, one of the two,” for

the first time implicitly foreclosing the possibility of striking Borum as a witness

entirely, without explanation.

When the court returned, it declared that it had “crafted an instruction that the

Court believes will cure the problem.” The proposed instruction, largely tracking

the government’s most recent proposal, was as follows:

There is no evidentiary basis for the testimony given by Juanita Borum regarding her receiving a phone call the night of the homicides. You may infer that the lack of evidence is unfavorable to the Government. You are to disregard her testimony regarding any phone call she received, and that testimony has been stricken from the record. Ms. Borum’s testimony should be considered with caution and scrutinized with care.[5]

The defense again insisted that was “not sufficient” and “only highlights and

compounds the problem,” urging the court to strike Borum’s testimony entirely. The

court responded that striking Borum’s testimony entirely “would be overly punitive

to the government when this group of 15 [jurors] took an oath to follow the Court’s

5 Neither the government’s nor the trial court’s proposed curative instructions addressed the government’s question: “Where was it that night where everyone was saying, your date did this; your date did this?” While counsel’s questions are not evidence, that question presupposed a highly prejudicial fact that was not in evidence—that “everyone was saying” Walker did it. 21

instructions,” so that a curative instruction would suffice. The court was adamant

that it had offered “an appropriate corrective instruction to fix this mess,” but

because the defense “can’t work with that instruction,” the court believed that things

had “risen to the point where manifest necessity is almost warranting” a mistrial.

The court then declared a mistrial.

The parties returned to court two days later, ostensibly to schedule a new trial.

The defense moved to dismiss the indictment on double jeopardy grounds because

the court “declared a mistrial over the defense’s objection.” The trial court denied

Walker’s motion to dismiss the indictment, and Walker now appeals that ruling,

which is an appealable interlocutory order. See Abney v. United States, 431 U.S.

651, 662-63 (1977) (“[A] pretrial order denying a motion to dismiss an indictment

on double jeopardy grounds” is immediately appealable.).

II. Analysis

A. Manifest Necessity and Our Standard of Review

Walker’s only claim on appeal is that his retrial is barred by the Double

Jeopardy Clause. The Fifth Amendment provides that no person can “be twice put

in jeopardy . . . for the same offense.” U.S. CONST. amend. V. Jeopardy “attaches”

once a jury is sworn in for an initial trial. United States v. Martin Linen Supply Co.,

430 U.S. 564, 569 (1977). This constitutional protection is not just a guard against 22

being punished twice for the same offense, but “also embraces the defendant’s

valued right to have his trial completed by” the first tribunal to hear it. Arizona v.

Washington, 434 U.S. 497, 503 (1978) (citations omitted). The principle

undergirding this constitutional protection is that the government “with all its

resources and power should not be allowed to make repeated attempts to convict an

individual for an alleged offense.” Green v. United States, 355 U.S. 184, 187 (1957).

This right to be tried just once is not absolute; retrials happen. Where a

defendant successfully requests a mistrial, for instance, there is generally no bar to

his retrial on the theory that he effectively consented to it. Lee-Thomas v. United

States, 921 A.2d 773, 778 (D.C. 2007). The same is true if the defendant

successfully appeals a conviction—a retrial is virtually always permitted. Price v.

Georgia, 398 U.S. 323, 326 (1970) (double jeopardy does not “prevent[] a second

trial when a conviction had been set aside”); Smith v. United States, 599 U.S. 236,

242 (2023) (noting that retrials are permitted except where grounds for appellate

reversal were “violations of the Speedy Trial Clause”). Yet even these exceptions

are subject to a caveat: If the prosecution intentionally “goads” a mistrial, i.e., if it

sows error into the first trial for the sake of getting a do over, then even a mistrial

granted at the defendant’s request will bar a retrial. See Oregon v. Kennedy, 456

U.S. 667, 676 (1982); Fletcher v. United States, 569 A.2d 597, 598-99 (D.C. 1990)

(applying goading doctrine in the context of appellate reversal). 23

But when a mistrial is declared over defense objection, a retrial is barred

unless there was “manifest necessity” for the mistrial. Hinton v. United States, 979

A.2d 663, 680 (D.C. 2009) (en banc). On appellate review, we ask “whether the

[trial] judge exercised sound discretion” when determining that manifest necessity

compelled a mistrial. 6 See Vega v. United States, 709 A.2d 1168, 1172 (D.C. 1998).

The “manifest necessity standard is the balance struck by the Constitution between

‘the public interest in affording the prosecutor one full and fair opportunity to present

his evidence to an impartial jury’ and an objecting defendant’s ‘valued right’” to

have their original tribunal decide their fate. Sanchez v. United States, 919 A.2d

1148, 1155 (D.C. 2007) (quoting Washington, 434 U.S. at 505).

Our precedents instruct that “a mistrial over the objection of the defense is

justified ‘only in very extraordinary and striking circumstances,’” Bailey v. United

6 We have said conflicting things about how exacting our scrutiny of manifest necessity determinations is. On the one hand, as the government highlights, we have said that this court “should accord the highest degree of respect to the trial judge’s determination of manifest necessity.” Douglas, 488 A.2d at 133 (quoting Washington, 434 U.S. at 511). On the other hand, as Walker highlights, we have said that “a reviewing court must ‘resolve any doubt’” about the existence of manifest necessity “‘in favor of the liberty of the citizen.’” Id. (quoting Downum v. United States, 372 U.S. 734, 738 (1963)); see also Vega, 709 A.2d at 1171 (same). We think it suffices to say that, “in the final analysis,” our role is to ensure that the reasons for aborting the trial were sufficient to overcome “the importance to the defendant of being able, once and for all, to conclude his confrontation with society through the verdict of a tribunal he might believe to be favorably disposed to his fate.” Coleman v. United States, 449 A.2d 327, 329 (D.C. 1982) (quoting Jorn, 400 U.S. at 486). If not, then a retrial is barred. 24

States, 676 A.2d 461, 465 (D.C. 1996) (quoting Douglas, 488 A.2d at 132), but that

is a half-truth. There is one fairly ordinary and not so striking circumstance that

satisfies the manifest necessity standard: Where the jury is deadlocked and cannot

reach unanimous agreement on the charges before it. “The reasons for ‘allowing the

trial judge to exercise broad discretion’” to declare a mistrial over defense objection

“are ‘especially compelling’” in the event of juror deadlock. Renico v. Lett, 559 U.S.

766, 774 (2010) (quoting Washington, 434 U.S. at 509).

Aside from a deadlocked jury, though, only truly exceptional circumstances

provide valid grounds for declaring a mistrial over defense objection. For instance,

the onset of the COVID-19 pandemic scuttled trials across the country. See, e.g.,

State v. Smith, 244 A.3d 296, 300 n.5 (N.J. Super. Ct. App. Div. 2020) (counting

twenty-six federal district courts and dozens of states, plus the District, that had

“suspended jury trials” around the pandemic’s onset). Even when the defense

objected to the mistrials, courts generally permitted retrials in those cases because

manifest necessity demanded that the first trial be terminated—stay-at-home orders

prevented litigants and jurors from even convening. See, e.g., id. at 305; Hightower

v. State, 883 S.E.2d 335, 336-37 (Ga. 2023).

This court typically bifurcates the manifest necessity inquiry into two steps.

First, we ask whether some development gave “rise to a ‘“high degree” of necessity’

to terminate the trial.” Douglas, 488 A.2d at 132 (citation omitted); Washington, 25

434 U.S. at 506 (noting that “the key word ‘necessity’ cannot be interpreted

literally,” but need only be of a “high degree”). Second, we ask whether there was

any reasonable “alternative measure—less drastic than a mistrial—that [could have]

cure[d] the problem.” Bailey, 676 A.2d at 463; Douglas, 488 A.2d at 135, 139 (the

“less drastic alternative” must be “reasonable”). 7 We conclude that the first step of

the analysis is dispositive here because there was no high degree of necessity to

declare a mistrial over Walker’s objections. And even if there were a high degree

of necessity for the mistrial, we conclude at the second step of the analysis that there

were reasonable alternatives to a mistrial, so that neither of the prerequisites for a

manifest necessity finding was satisfied in this case.

B. There Was No High Degree of Necessity to Declare a Mistrial

This first step of the manifest necessity analysis requires us to examine the

7 It is fairly artificial to treat the availability of alternatives to a mistrial as a second and standalone requirement for a manifest necessity finding, when it is more natural to say that the availability of reasonable alternatives simply precludes any finding that a high degree of necessity for a mistrial existed. Still, our precedents typically bifurcate the analysis in this manner, and so we do likewise, mindful that these two factors should not be siloed because there is substantial interplay between them. For instance, as we explain in the next section, which (if either) party is to blame for creating the problem is a critical part of the “high degree of necessity” part of our inquiry. But that consideration might also come into play when determining which of the potential alternatives to a mistrial are reasonable; whether a particular remedy is too punitive to one of the parties might depend on the relative blameworthiness of the parties for creating the issue in the first place. 26

totality of circumstances and ask, at bottom, whether “the reasons for aborting a trial

[were] sufficient to override the defendant’s double jeopardy interests.” Douglas,

488 A.2d at 132 (citation omitted). “[P]rosecutorial error even of a degree sufficient

to warrant a mistrial” at the defendant’s request does not necessarily clear this first

bar. Kennedy, 456 U.S. at 676; United States v. Sanford, 429 U.S. 14, 16 n.2 (1976)

(“manifest necessity” is not required for a mistrial requested by the defense).

One factor in particular drives our analysis at this first step of our inquiry: that

it was the government who introduced the error that ultimately prompted a mistrial

over Walker’s objections. See Wayne R. LaFave, et al., 6 Crim. Pro. § 25.2(c) (4th

ed. 2023) (listing “the source of the difficulty” as the first among factors that courts

consider “on a fairly regular basis” as part of the manifest necessity inquiry). We

now explain why that is of the utmost significance before addressing several

additional factors that counsel against finding any high degree of necessity for a

mistrial in this case.

1. Prosecutorial Error Rarely Justifies a Mistrial Over Defense Objection

Manifest necessity most frequently arises when neither party is to blame for

circumstances that make it untenable to proceed with trial—such as the classic case

of a deadlocked jury or the COVID-19 outbreak discussed above. See Smith, 244

A.3d at 311 (stressing that the “predicament” created by the pandemic was “beyond 27

the control of all involved,” and was “not the result of prosecutorial or defense

misconduct”). And when defense error prejudices the government, then courts take

a somewhat relaxed approach to the manifest necessity inquiry. See, e.g.,

Washington, 434 U.S. at 510 (upholding manifest necessity ruling where

“defendant’s lawyer made improper and prejudicial remarks” disparaging the

government during opening statement); State v. Green, 992 N.W.2d 56, 59 (Wis.

2023) (retrial permitted where mistrial was prompted by the defense’s failure to

preclear third-party perpetrator defense with the court); Moussa Gouleed v. Wengler,

589 F.3d 976, 980 (8th Cir. 2009) (mistrial was manifestly necessary because “the

state was deprived of a fair trial when the defense did not provide full disclosure of

its expert’s basis for opinions going to the very heart of the” case); see also Vega,

709 A.2d at 1173 (distinguishing Washington on the basis that it concerned

“society’s interest in giving the prosecutor one complete opportunity to convict”

(quoting Washington, 434 U.S. at 509)). After all, if the defense deprived the

government of a full and fair shot at conviction, then it has relatively little cause to

complain if its own missteps require a reset.

By contrast, when prosecutorial error prejudices a defendant, the virtually

ironclad rule is that the defendant retains the option to proceed with trial if they wish,

so that a mistrial cannot be granted over their objections. In the parlance of our

precedents, when “a mistrial in response to judicial or prosecutorial error” is 28

contemplated, “[t]he important consideration, for purposes of the Double Jeopardy

Clause, is that the defendant retain primary control over the course to be followed.”

Dinitz, 424 U.S. at 609; Speaks v. United States, 617 A.2d 942, 955 (D.C. 1992)

(same); Vega, 709 A.2d at 1174 (same); see also James F. Ponsoldt, When Guilt

Should Be Irrelevant: Government Overreaching as a Bar to Reprosecution Under

the Double Jeopardy Clause After Oregon v. Kennedy, 69 Cornell L. Rev. 76, 85

(1983) (Dinitz “established a formalistic rule of defendant choice in situations

involving prosecutorial or judicial error”); Michael J. Klarman, Note, Mistrials

Arising From Prosecutorial Error: Double Jeopardy Protection, 34 Stanford L.

Rev. 1061, 1078 n.96 (1982) (“Dinitz can be plausibly interpreted to mean that when

prosecutorial error prejudices a jury, the double jeopardy clause prohibits mistrial

over the defendant’s objection.”).

To illustrate the point that prosecutorial error is rarely cause for a mistrial over

a defendant’s objection, consider the Supreme Court’s decision in Downum. In that

case, shortly after a jury was empaneled—before opening statements or the

introduction of any evidence—the government alerted the trial court that it had been

unable to find one of its key witnesses in the days leading up to trial. 372 U.S. at

735. The court granted a mistrial over defense objection to allow the government to

track down its witness. Id. The defendant then argued that the Double Jeopardy

Clause barred his retrial, and the Supreme Court agreed. Id. Downum explained 29

that when the government proceeded with jury selection, despite not being able to

locate its witness, it “took a chance,” and even if the witness’s absence left the

government “without sufficient evidence to convict,” that did not amount to manifest

necessity. Id. at 737 (quoting Cornero v. United States, 48 F.2d 69, 71 (9th Cir.

1931), and explaining Cornero espoused “the correct [view]”); see also Seay v.

Cannon, 927 F.3d 776, 782 (4th Cir. 2019) (granting habeas relief because mistrial

could not be granted over defendant’s objections just because government’s key

witness, without whom there was insufficient evidence to convict, went missing).

Similarly, in Hubbard v. State, what is now the Supreme Court of Maryland

announced a seemingly rigid rule that where the government “created the

conundrum,” it could not “be the beneficiary of a manifest necessity analysis” so

that a mistrial could not be granted over defense objection. 909 A.2d 270, 282 (Md.

2006). That case involved a joint trial for attempted murder, and a government

witness made pretrial identifications of both co-defendants, but that identification

had been suppressed as to one of the two co-defendants. Id. at 272-73. After the

jury had been empaneled, but before the jury had heard any evidence, the

government raised a concern that in eliciting the admissible identification it could

not help but elicit the suppressed one as well. Id. The trial court declared a mistrial

to permit the government to sever the two cases and alleviate the issue. Id. at 278.

The Supreme Court of Maryland reversed. Id. It concluded that the government’s 30

own negligence was not a basis to abridge the defendant’s double jeopardy rights,

and it explained that the proper remedy was to simply exclude the eyewitness’s

testimony entirely. Id. at 282-83.

Likewise, in State v. Rowe, the court held that “[t]he State cannot rely on a

problem created by its own neglect to establish the existence of manifest necessity.”

480 A.2d 778, 782-83 (Me. 1984) (interpreting double jeopardy clause of Maine’s

constitution). In Rowe, the government waited until the third day of a joint trial to

seek a ruling from the trial court regarding a Bruton issue, see Bruton v. United

States, 391 U.S. 123 (1968), which concerns whether the state can admit one

co-defendant’s confession when it clearly implicates another co-defendant (raising

a Confrontation Clause issue where the confessing co-defendant chooses not to

testify and so cannot be confronted), Rowe, 480 A.2d at 780-81. When the trial court

ultimately ruled that the confession could not be admitted during the joint trial, it

declared a mistrial so that the government could sever the defendants and try them

individually. Id. at 781. The Supreme Judicial Court of Maine reversed, reasoning

that the first trial should have simply proceeded without the confession, because the

government’s failure to seek a ruling on the Bruton issue before trial was no

justification for infringing on the defendant’s double jeopardy rights. Id. at 782-83.

A retrial was thus barred. Id. at 783. 31

Also consider Vega, a case where this court exalted the importance of a

defendant’s choice when some trial mishap prejudices only them, even if through no

fault of the prosecutors. In Vega, near the end of a nine-day trial, defense counsel

“failed to appear in court to give closing argument.” 709 A.2d at 1168; id. at 1170

(defense counsel “had a breakdown”). The trial court declared a mistrial without

consulting the defendants. Id. at 1171. We reversed and barred any retrial, stressing

that the trial court’s failure to consult the defendants about whether they wished to

proceed was constitutional error because it was “at least conceivable they would

have chosen” to deliver closing arguments pro se, or to continue without closing

argument at all. Id. at 1173-75. Even if a mistrial were necessary to “prevent a

debacle” like that, we reasoned that the defendants had the right to opt for the debacle

rather than forgoing their constitutional right to be tried just the once. Id. at 1174.

The government seeks to cabin Downum, Hubbard, and Rowe (it ignores

Vega) to cases involving “protracted [prosecutorial] negligence,” whereas it chalks

its error in this case up to a “brief” slip-up, i.e., the short time it took to introduce the

improper hearsay. We doubt that matters, 8 but even if it did, this case involves

8 This is not a matter of punishing the prosecution for its errors, but a question of whether there is sufficient justification to strip a defendant of his constitutional right to see their trial through to its end. Whenever the problem is created by the prosecution—no matter how protracted their negligence—this consideration will strongly counsel against a mistrial over defense objection. Also, the government 32

prosecutorial negligence that is just as protracted as in any of those cases. As the

trial court stressed, the parties spent “countless hours” pretrial “litigating everything

that was anticipated to come out of Ms. Borum’s mouth” in the months before the

trial began, and yet the government conspicuously failed to alert the court or defense

counsel that it intended to elicit the highly prejudicial statements at issue here. 9 Had

it flagged the issue earlier, the trial court made clear that it never would have

permitted the inadmissible hearsay to come before the jury.

does not articulate any basis for its assertion that the negligence in Downum was protracted. The government had done its part to subpoena its missing witness in that case. 372 U.S. at 735 (“Subpoenas for all of [the witnesses], including [the missing one], had been delivered to the marshal for service.”). 9 The trial prosecutors offered just one explanation for why they did not preview the prejudicial hearsay at issue here, but it was facially implausible and the trial court discredited it. Prosecutors argued that the pretrial litigation was about hearsay exceptions, while positing that the problematic statements were non- hearsay. That is indeed a legal distinction, but it is beyond us how anybody could think it a relevant one when it comes to deciding which of various statements a party might want to seek a pretrial ruling on. If there is one out-of-court statement that a prosecutor would think to preclear before introducing it in court—whether under the umbrella of a hearsay exception or as non-hearsay—it is an out-of-court declaration to the effect that the defendant is guilty of the crimes charged, like the ones at issue here. Also, the record undermines the trial prosecutors’ explanation because the government in fact repeatedly argued in the pretrial litigation that various statements it sought to admit were non-hearsay, so the government on appeal understandably makes no attempt to justify the prosecutorial tactics. While we do not disturb the trial court’s finding that the government did not intentionally goad a mistrial, its related finding that the prosecutors “knew [what they] were doing” when they kept these statements in their back pocket was well founded. 33

To take a step back from the doctrine for a moment, simply consider the

respective interests at stake in a case like this one, where mid-trial prosecutorial error

prompts a mistrial over defense objection. If Walker was correct that he was en

route to acquittal—as he had good reason to believe, given how the trial was going—

then the trial court deprived him of a reasonable chance of walking free. No

comparable harm could possibly have befallen the government if it had simply been

required to see its prosecution through to its end. As we stressed in Vega, this is not

a case where the government had been prejudiced by defense error, so there was

simply no need to vindicate “society’s interest in giving the prosecutor one complete

opportunity to convict those who have violated its laws.” 709 A.2d at 1173 (quoting

Washington, 434 U.S. at 509). The prosecution would get its fair shot at convicting

Walker regardless. If trial had simply proceeded and Walker was acquitted despite

the unfair handicap against him, the prosecution could have no complaint about that.

Whereas if the prosecution secured a conviction with its unfair advantage in

hand, at worst it would have been unable to defend that conviction on appeal—which

was far from certain—a temporary setback that would have been entirely of its own

making. Even in that worst-case scenario for the government, it would have been

right back in the same place of retrying Walker after an appellate reversal, with the

only salient difference being that Walker probably would have spent some years in

prison under an infirm conviction while waiting for an uncertain appellate process 34

to play out (and the parties would have squandered some resources along the way).

So when a defendant prefers to proceed against the headwinds of unfair prejudice

introduced against them at trial, and opts to risk years of imprisonment rather than

immediately starting anew, we see no adequate justification for stripping them of

that choice. Id. After all, in this scenario it is the defendant who is faced with the

“‘Hobson’s choice’ between giving up his first jury and continuing a trial tainted by

prejudicial judicial or prosecutorial error.” Dinitz, 424 U.S. at 609. Whatever choice

they make must be paramount. Cf. Curry v. Superior Court, 470 P.2d 345, 351 (Cal.

1970) (courts “must avoid depriving the defendant of his constitutionally protected

freedom of choice in the name of a paternalistic concern for his welfare” under

California constitution’s double jeopardy clause).

The government cites just one case where prosecutorial error was held to be

adequate grounds to declare a mistrial over a defendant’s objection, and it is readily

distinguishable: Illinois v. Somerville, 410 U.S. 458 (1973). 10 Somerville concerned

10 The government’s brief relies largely on cases that either (1) did not implicate the manifest necessity standard at all because they simply did not concern mistrials granted over defense objection, see, e.g., Coleman v. United States, 779 A.2d 297, 299-300 (D.C. 2001) (concluding that mistrial at defendant’s request was warranted with no mention of manifest necessity), and Fletcher, 569 A.2d at 597-98 (rejecting argument that double jeopardy barred a retrial after appellate reversal without mention of manifest necessity); or (2) involved defense error that prejudiced the government, see Washington, 434 U.S. at 511; Green, 992 N.W.2d at 59; Moussa 35

a defective indictment that “as a matter of Illinois law [was] not curable by

amendment.” Id. at 468. The government caught its error and sought to remedy it

after a jury had been empaneled—so that jeopardy had technically attached—but

“before any evidence had been presented” in the case. Id. at 459. The Supreme

Court ruled that the government’s error in procuring a defective indictment gave rise

to manifest necessity to declare a mistrial because, quite unlike this case, the

prosecutorial error at issue (1) was caught before the introduction of any evidence,

(2) it did not “lend itself to prosecutorial manipulation,” and (3) reversal would have

been “a certainty,” i.e., “automatic[],” had the trial proceeded. Id. at 464.

Gouleed, 589 F.3d at 980, a scenario we have already explained involves a very different balancing of interests.

Two other cases the government does not rely upon merit brief discussion: Gori v. United States, 367 U.S. 364 (1961), and United States v. Sedgwick, 345 A.2d 465 (D.C. 1975). Both cases upheld manifest necessity findings prompted by prosecutorial error, and both share two common distinguishing features: (1) they both predate the Supreme Court’s opinion in Dinitz, and are quite difficult to square with its central principle that a defendant must retain control of the course to be followed in the event of judicial or prosecutorial error, and (2) the defendant in neither case expressed any interest in proceeding with his first trial. Gori, 367 U.S. at 365 (mistrial declared “with neither approval nor objection by petitioner’s counsel”); Sedgwick, 345 A.2d at 473 (“while the defendant did not expressly consent to a mistrial . . . at no time did his counsel” express any interest in “completion of the trial by the first jury”). While we have serious doubts about the extent to which Gori and Sedgwick remain good law—we elaborate on Gori in footnote 12—they are easily distinguished on the basis that Walker expressed his preference to proceed with his first trial if the charges were not dismissed outright. 36

There is some tension between Somerville and Dinitz’s later pronouncement

that defendants retain the option to proceed in the face of prosecutorial error.

Commentators have reconciled them on the quite reasonable bases that the

evidentiary phase of trial was not underway when a mistrial was declared in

Somerville, and the error in Somerville was truly of a structural nature; that is to say

the prosecutorial error was not conceivably made for any strategic advantage and

there was no hope of curing it. “Until testimony opens, the defendant’s interest in

avoiding reprosecution ordinarily will be rather weak.” Stephen J. Schulhofer,

Jeopardy and Mistrials, 125 U. Pa. L. Rev. 449, 507 (1977). But “[a]s soon as the

taking of evidence begins, mistrial involves a radically increased potential for

prejudice to the defendant.” Id. at 508; Klarman, supra, 34 Stan. L. Rev. at 1078 &

n.96 (reconciling Somerville and Dinitz on the basis that Somerville did not involve

an error “that prejudice[d] the jury”). As we now explain, those reconciliations are

sound and align with our precedents.

2. It is Particularly Hard to Justify Aborting a Trial in Its Evidentiary Phase

Once a trial has entered its evidentiary phase, it is tough to see how

prosecutorial error could ever justify stripping the defendant of a jury that has heard

opening statements and some evidence in his case. Douglas, 488 A.2d at 130

(defendant’s interests “in retaining a chosen jury may intensify as the trial

proceeds”). A defendant has a particularly strong interest in “being able, once and 37

for all, to conclude his confrontation with society” once evidence has been presented

and he has any reason to believe that the jury is “favorably disposed to his fate,” i.e.,

“if the defendant senses the trial is going well and the factfinder is leaning toward

acquittal.” Id. (quoting Jorn, 400 U.S. at 486 (plurality opinion)).

This case is a vivid illustration of why that is. Consider several aspects of

what had unfolded at Walker’s trial before the trial court declared a mistrial: (1) the

evidence presented suggested that the government had a weak case against Walker,

(2) the defense had previewed its tactics and evidence during the aborted trial, (3) the

jurors had heard enough to formulate tentative opinions, and (4) the prosecutors’

introduction of prejudicial hearsay against Walker could have been part of an effort

to save a floundering case. 11 LaFave, supra, at § 25.2(c) (listing each of these four

factors as among those that courts regularly count against a manifest necessity

finding). We now elaborate on each of those four points.

First, by any standard this was a shaky government case that depended on

extremely problematic witnesses: One lead government witness (White Boy) was an

obvious suspect himself, another (Borum) would have been substantially undercut

11 The government never requested a mistrial without prejudice, but it also never objected to one when the trial court floated that possibility. The defense was the only party who objected to starting anew, and while we cannot say the prosecution expressly welcomed it, it did not seem particularly opposed to or aggrieved by the trial court’s ultimate ruling. 38

by GPS evidence, and the third (Taper) had been convicted of perjury under

remarkably similar circumstances to the case at bar. Plus the prosecutors did not

foresee that Taper would be impeached with her prior perjury conviction because

they mistakenly believed that the conviction was too old to be admitted. At any

retrial, prosecutors could better anticipate that major flaw in their case and try to

shore it up, or at least take the sting out of the perjury conviction by eliciting it in

the government’s own case rather than leaving the defense to batter Taper with it.

Douglas, 488 A.2d at 130 (“The opportunity to present the government’s case a

second time may permit the prosecutor to compensate for weaknesses exposed

during the first trial.”); id. (“[G]overnment witnesses frequently become more

definite in their testimony and more favorable to the prosecution with successive

trips to the stand.”); Washington, 434 U.S. at 504 n.14 (“[S]ubtle changes in the

State’s testimony, initially favorable to the defendant, may occur during the course

of successive prosecutions.” (citation omitted)).

Second, the government got a substantial preview of the defense case, ranging

from how it was going to undercut the government’s witnesses to a third-party

perpetrator defense that it had not revealed before trial. See United States v.

DiFrancesco, 449 U.S. 117, 128 (1980) (“[I]f the Government may reprosecute, it

gains an advantage from what it learns at the first trial about the strengths of the

defense case and the weaknesses of its own.”); Ashe v. Swenson, 397 U.S. 436, 447 39

(1970) (double jeopardy precludes the government from “treat[ing] the first trial as

no more than a dry run for the second prosecution”); Schulhofer, supra, 125 U. Pa.

L. Rev. at 506 (unlike the defense, which learns a great deal about the government’s

case due to pre-trial disclosure obligations, “the prosecution generally lacks the

opportunity to learn much prior to trial about defense tactics”); see also Moghalu v.

United States, 263 A.3d 462, 472 (D.C. 2021) (defense does not need to disclose

third-party perpetrator defense to government before trial). To simply start things

over from scratch would be a sizable boon for the government, because it would

walk into any retrial with Walker’s playbook, depriving Walker of an element of

surprise that is far more critical to criminal defendants than it is to the prosecution.

May v. Shinn, 954 F.3d 1194, 1206 (9th Cir. 2020) (“Due to asymmetries in

disclosure obligations,” the defense is typically “able to learn more about the

prosecution’s case before trial beg[ins] than the other way around.”); Huss v. Graves,

252 F.3d 952, 957 (8th Cir. 2001) (“The prohibition of a retrial prevents the

government from ‘gain[ing] an advantage from what it learns at the first.’” (quoting

DiFrancesco, 449 U.S. at 128)).

Third, the jury had heard more than enough to form tentative opinions in this

case. The government was through twenty-six witnesses—the bulk of its case—and

yet, by its own telling, had produced “scant evidence that Walker had committed the

murders.” A jury that has sat through two-and-a-half weeks of trial and has heard 40

little evidence of a defendant’s guilt is sure to have some tentative views on the case,

likely favoring acquittal.

Fourth, while we do not disturb or doubt the trial court’s finding that the

government did not intentionally goad a mistrial, this was nonetheless prosecutorial

error that “len[t] itself to prosecutorial manipulation,” Somerville, 410 U.S. at 464,

and “could have been intentionally created” by the government in order “to

strengthen its case,” LaFave, supra, at § 25.2(c) (emphasis added). 12 Our precedents

12 LaFave lists a total of twelve factors that courts regularly consider as relevant to the manifest necessity inquiry. We discuss the bulk of them in this opinion, though several of them are irrelevant on these facts (such as whether “the composition of the jury was unusual”). Only one factor even superficially favors a manifest necessity finding here: the fact that the trial court seemed to grant a mistrial “solely” for Walker’s benefit. LaFave, supra, at § 25.2(c). Contrary to LaFave, we do not understand this to be any point in favor of a mistrial granted over defendant’s objections. The notion that a trial court’s paternalistic motives are a point in favor of a mistrial is an outdated view grounded in Gori, 367 U.S. at 369, where the Supreme Court upheld a manifest necessity ruling because the trial court was acting “in favor of the accused,” and the defendant lodged no objection at all. Id. at 367. Gori was substantially undermined two years later in Downum, then undermined further the following decade in Jorn, 400 U.S. at 483 (plurality opinion) (“[A]n appellate court’s assessment of which side benefited from the mistrial ruling does not adequately satisfy the policies underpinning the double jeopardy provision.”). The Supreme Court’s still subsequent holding in Dinitz was the death knell for any notion that acting for the defendant’s benefit is a point in favor of declaring a mistrial over their objection. See Schulhofer, supra, 125 U. Pa. L. Rev at 463, 481, 535 (discussing how, two years after Gori, “the Supreme Court radically transformed the jurisprudence of mistrials in Downum,” then later “inter[red]” Gori in Jorn, and then Dinitz enshrined the now-prevailing principle of “honor[ing] the defendant’s preference” in the event of prosecutorial or judicial error); Klarman, supra, 34 Stan. L. Rev. at 1066 (describing Gori as a case where “the mistrial was manifestly 41

stress that we apply the “strictest scrutiny” to manifest necessity determinations

when there is a mere “reason to believe that the prosecutor” sought “to achieve a

tactical advantage over the accused” when creating a problem that led to a mistrial.

Routh v. United States, 483 A.2d 638, 642, 645 (D.C. 1984) (quoting Washington,

434 U.S. at 508). 13 And here, informing jurors that “everyone was saying” Walker

committed the murders was certainly strategically advantageous to the prosecution.

unnecessary”). Vega demonstrates just how out-of-step Gori is with modern double jeopardy jurisprudence. In Vega, the trial court declared a mistrial because defense counsel was a no-show for closing arguments, but we firmly rejected the notion that the trial court’s paternalistic motive in declaring a mistrial for the defendants’ benefit was any point in its favor. 709 A.2d at 1174-75. Just the opposite. We stressed that where the defendants are the only party prejudiced by the “debacle,” it was most essential that they be permitted to decide whether or not to proceed. Id. (citing Dinitz, 424 U.S. at 609). 13 The government counters that this “strictest scrutiny” standard applies only when there is intentional goading, but that cannot be. First, Routh applied the strictest scrutiny to a mistrial that was declared due to the unavailability of a government witness—holding that a retrial was barred—despite there being no suggestion that the government intentionally made its witness unavailable for the sake of tanking the first trial. 483 A.2d at 642. Second, when the government intentionally goads a mistrial, that ends the double jeopardy analysis; there is no scrutiny at all of whether manifest necessity existed, because a retrial is precluded regardless, even if the mistrial was granted at the defendant’s request and even if it was manifestly necessary. Kennedy, 456 U.S. at 679. While our cases have not always been clear on the matter, we view Routh’s “strictest scrutiny” test as a mere application of the Dinitz principle that defendants must retain primary control over the course to be followed in the event that prosecutorial error infects a trial. 42

Consider an analogy. Imagine you are eight rounds into a twelve round

boxing match when your opponent punches you a couple of times below the belt—

prototypical “low blows.” The illegal punches hurt, giving your opponent something

of an unfair advantage over you, but you are immediately willing and able to resume

the fight. It would make sense for the referee to (1) issue a warning to your opponent

or (2) to deduct some points from your opponent’s score. And it might even make

sense, if the illegal punches were particularly egregious or intentional, (3) to declare

a forfeit in your favor, so that you are declared the winner. 14 But it would make no

sense at all for the referee to terminate the fight and schedule a rematch at a future

date over your objections that you would rather settle the bout right then and there.

If you think you are winning the fight despite the unfair advantage your opponent

has created for themselves—or even if you just like your odds better in the present

match than in any rematch, or would prefer not to repeat the ordeal—then what

salient objection could your opponent conceivably have to proceeding? Their

14 We look to boxing because, similar to jurors in a trial, there are judges keeping a running score but the combatants themselves will not learn of those scores until the match concludes. While participants in both trials and boxing matches do not know exactly how their adjudicators are scoring the match, they will often have informed assessments of how things are going. It is illegal to punch “below the belt,” and violations of that rule can result in warnings, the deduction of points, or outright disqualification, at the referee’s discretion. See Association of Professional Boxing, Rules § 1.38(a) (prohibiting “hitting below the belt”), § 1.38.1 (referee may “caution or disqualify” for violating that rule), § 1.38.2 (referee “may also deduct points”), available at https://apbcboxing.com/?page_id=28; https://perma.cc/UMS7-T49X. 43

infractions are simply not a good reason to rob you of your chance to wrap up the

match, or to grant them a do over, particularly if you think victory is at hand.

And it would not resonate in the slightest if your opponent argued that they

should not have to finish the fight because their own violations were so egregious

that they would surely be stripped of any victory should you lodge a protest after the

fight. That is an utterly hollow complaint: Even if that were true, a rematch is most

certainly not inevitable because you might win the first match and settle the matter

for good. Your opponent’s complaint amounts to nothing more than that they may

have to duplicate some efforts due to their own infractions, which is a minor harm

when compared to stripping you, the blameless party, of a potential victory. If

somebody’s ox needs to be gored, it should be your opponent’s, not yours. Making

you start from scratch because of your opponent’s infractions accords with no sense

of fairness or justice. 15

15 To run with the analogy, your opponent would have a more tenable argument for restarting the match if they had committed a mere technical infraction that they sought to remedy before the match was meaningfully underway—say they entered the ring with non-regulation gloves and noticed the impropriety just after the opening bell, before any punches were thrown. That maps onto the circumstances in Somerville, 410 U.S. at 459, and there is at least a colorable argument that the match should simply start anew—permitting your opponent to lace up with the proper gloves—rather than requiring your opponent to continue the match with illegal gloves that would taint any result. But once the match is truly underway, with tactics revealed and punches thrown, it offends basic notions of fairness to allow your opponent’s violations to strip you of your shot at winning the first match. 44

Of course, criminal trials are not mere sport, but that only highlights why a

defendant should not be stripped of a jury because of a prosecutorial blunder. As

defense counsel aptly put it, Walker was “fighting for his life,” not merely a

championship belt or a hefty prize, so to rip a potential acquittal from him because

of prosecutorial error is particularly hard to countenance. Contrary to the trial

court’s reasoning, declaring a mistrial after the prosecution “g[a]ve that left/right

punch across the bow” did not “put[] everybody back at the starting point” as if

nothing had happened. Walker had strong and constitutionally backed reasons for

preferring that his initial trial proceed to a verdict, and he cogently explained why a

mistrial would have put him “in a far worse place than” simply proceeding with trial

sans remedial action. The trial court stripped Walker of his right to receive a verdict

from a jury that he had reason to believe was favorably disposed toward acquittal.

There was no countervailing interest that justified depriving Walker of his right to

see his trial through to its end, even if no alternative measure could have remedied

the error that the prosecution introduced into the trial.

In sum, there was no high degree of necessitating a mistrial in this case. We

therefore conclude that it is immaterial whether any alternatives to a mistrial would

have been effective, and that Walker’s trial should have proceeded regardless of

whether any alternative measures could have remedied the prejudice against him.

Nonetheless, for the sake of completeness, and mindful that this area of the law 45

“abjures the application of any mechanical formula,” Somerville, 410 U.S. at 462,

we now discuss the perfectly viable alternatives to a mistrial that existed.

C. There Were Reasonable Alternatives to a Mistrial

There was no manifest necessity for a mistrial in this case for the additional

reason that, at the second step of our inquiry, there were viable alternatives to a

mistrial. Specifically, the trial court considered two alternatives to a mistrial that

would have been far better than aborting Walker’s trial over his objection, either one

of which might have sufficiently cured the prejudice against him. The court could

have stricken Borum’s offending testimony along with a curative instruction, as the

government preferred and as the trial court steadfastly opined would “cure the

problem.” And if, counterfactually, the trial court concluded that was too weak

medicine, it could have stricken Borum as a witness entirely, as the defense

preferred. Either alternative was reasonable.

First, the trial court could have simply stricken Borum’s testimony about the

phone call with Wobbie, along with a curative instruction that would have told jurors

(1) to disregard it, (2) that it was unsupported by any evidence, and (3) to approach 46

Borum’s testimony “with caution” and to “scrutinize[]” it “with care.” 16 That would

have been an atypically strong curative instruction for the defense, as the last piece

of it incorporated the type of instruction that is usually reserved for the testimony of

known perjurers. See Green v. United States, 231 A.3d 398, 415 n.57 (D.C. 2020)

(recounting jury instruction that “the testimony of an admitted perjurer . . . should

be considered with caution and scrutinized with care”).

“Our cases recognize that an instruction not to consider stricken testimony,”

even without the gloss from the perjurer’s instruction, “is usually a sufficient remedy

where a jury has heard damaging testimony it should not have been permitted to

hear.” Foote v. United States, 108 A.3d 1227, 1238 (D.C. 2015). “We ordinarily

16 The government asserts that Walker has waived any argument that a curative instruction would have worked, because he argued before the trial court that it was an inadequate remedy. Of course, the government has also reversed course on this topic—it argued before the trial court that a curative instruction would have been effective—so its waiver argument comes with a hefty dose of chutzpah. That aside, there has been no waiver and nothing precludes Walker’s course reversal in this context, as we previously explained in Sanchez, 919 A.2d at 1152. In Sanchez, the defendant had agreed in the trial court that there was no jurisdiction to proceed with trial, but opposed a mistrial without prejudice and argued that a dismissal with prejudice was the proper remedy. Id. On appeal from a mistrial without prejudice, the defendant then “disavow[ed]” his prior concession and claimed the court had jurisdiction to proceed, an argument we ultimately agreed with when reversing the finding of manifest necessity and barring a retrial. Id. We rejected the government’s argument that the defendant had waived his newfound position, explaining that the defendant’s “objection to a mistrial preserved his double jeopardy claim” and that he was “not limited to the precise arguments [he] made below.” Id. (citing Yee v. City of Escondido, 503 U.S. 519, 534-35 (1992), and quoting Salmon v. United States, 719 A.2d 949, 953 (D.C. 1997)). The same is true here. 47

presume that the jury understands and obeys the trial judge’s instructions,” and that

an instruction to disregard some prejudicial testimony will therefore be effective.

Atkins v. United States, 290 A.3d 474, 485 (D.C. 2023) (quoting Holloway v. United

States, 25 A.3d 898, 903 (D.C. 2011)). That presumption extends even to a co-

defendant’s confession that implicates the defendant in all but name. See Samia v.

United States, 599 U.S. 635, 646 (2023) (“[O]ur legal system presumes that jurors

will ‘attend closely the particular language of [curative] instructions in a criminal

case and strive to understand, make sense of, and follow’ them.” (quoting United

States v. Olano, 507 U.S. 725, 740 (1993))). Our caselaw abounds with similar cases

where curative instructions were found to have sufficiently mitigated extremely

damning evidence introduced against a defendant. 17

17 Harris v. New York, 401 U.S. 222, 223-25 (1971) (jurors can follow instruction to consider otherwise inadmissible confession not for its truth, but only for its impeachment value); Allen v. United States, 579 A.2d 225, 228-29 & n.7 (D.C. 1990) (jury could follow limiting instruction, in manslaughter prosecution, that defendant was “coming to kill” the decedent for its limited probative value of explaining its effect on the decedent, but not for its truth), adopted on reh’g en banc, 603 A.2d 1219, 1228 n.20 (D.C. 1992) (en banc); Johnson v. United States, 883 A.2d 135, 139-40, 141 (D.C. 2005) (admission of highly prejudicial confession by co-defendant implicating defendant and introduced through hearsay testimony deemed harmless, and no mistrial required, where substantial other evidence of defendant’s guilt existed and court gave a limiting instruction); Peyton v. United States, 709 A.2d 65, 68, 72 (D.C. 1998) (prejudice from prosecution witness declaring he had taken a lie detector test was cured by judge’s instruction to disregard that testimony); Edelen v. United States, 627 A.2d 968, 972-73 (D.C. 1993) (no reversible error after prosecution elicited testimony that a non-witness 48

What’s more, the trial court in this case was of the view that a curative

instruction would have worked. Just before declaring a mistrial, the judge said she

had “crafted an instruction that the Court believes will cure the problem” and

“offered . . . what the Court believes is an appropriate corrective instruction to fix

this mess.” The court never walked that assessment back, but instead declared a

mistrial on the erroneous basis that the defense refused to accept, or “work with,”

“that instruction.” Washington, 434 U.S. at 510, n.28 (in assessing “manifest

necessity” ruling, “[i]f the record reveals that the trial judge has failed to exercise

the ‘sound discretion’ entrusted to him, the reason for . . . deference by an appellate

court disappears”). It should go without saying that trial courts should not declare

mistrials over defense objection simply because the defense will not concede a

potentially potent appellate issue. See Braxton v. United States, 395 A.2d 759, 771-

72 (D.C. 1978) (no manifest necessity where trial court’s “exhaustion” with defense

counsel “more than any other factor, appear[ed] to be the motivating factor for the”

mistrial).

That is not to say that such a curative instruction would have fully cured the

prejudice against Walker—we doubt it would have, and the improper prejudice here

identified the defendant out of a photo array because the court intervened and told the jury “the last comment you heard . . . you should disregard that. It should have no impact at all on your deliberations in this case.”). 49

quite possibly could have justified a mistrial at Walker’s request. But we need not

definitively opine on whether such a curative instruction would have ultimately

rendered the trial court’s error in admitting the testimony in the first place harmless.

We could not even venture a guess about that because it depends heavily on how the

rest of trial would have unfolded. In this setting, it is enough for us to say—and we

think it is inarguably true—that striking the offending testimony along with a

curative instruction might have worked. See Somerville, 410 U.S. at 459 (finding

manifest necessity where “reversal on appeal [was] a certainty” and “automatic[]”);

LaFave, supra, § 25.2(c) (one factor favoring manifest necessity finding is “whether

any conviction resulting from the trial would inevitably be subject to reversal on

appeal” (emphasis added)).

Second, if the curative instruction were not sufficient, then striking Borum as

a witness entirely was also a viable alternative that Walker had acquiesced to. Even

assuming that Borum could be fairly described as an “essential” government

witness—as the government argues in its brief—courts have pretty routinely held

that the government simply has to forgo even essential witnesses when a

prosecutorial error creates a conundrum where the alternative is declaring a mistrial

over defense objection. Downum, 372 U.S. at 737 (noting that absent witness was

“essential” for “two of the six counts,” but that was no justification for a mistrial

because the government had simply “entered upon the trial of the case without 50

sufficient evidence to convict”); Seay, 927 F.3d at 782 (defendant could not be

retried where mistrial was declared to permit government to track down essential

witness, without whom the evidence was insufficient to convict); see also Hubbard,

909 A.2d at 282 (prosecution having to forgo eyewitness identification testimony in

attempted murder case was insufficient cause to grant mistrial over defense

objection); Rowe, 480 A.2d at 782-83 (prosecution having to forgo evidence of

defendant’s confession insufficient cause to grant mistrial over defense objection).

And again, as with the curative instruction, the trial court here never opined

that striking Borum as a witness entirely would have been an ineffective remedy.

While the court did opine that it would have been “overly punitive” to the

government, that view was expressly predicated on the court’s belief that “this group

of 15 [jurors] took an oath to follow the Court’s instructions,” i.e., the jurors would

abide the less severe alternative of a curative instruction. If the court had in fact

rejected the curative instruction route—as the government now argues it must have

implicitly done—then the court offered no explanation whatsoever for why striking

Borum entirely was not a reasonable remedy to the conundrum that the government

created. Our discussion above and the quartet of cases just cited (Downum, Seay,

Hubbard, and Rowe) explain why this remedy was not overly punitive in light of 51

Walker’s significant interest and constitutional right to see his trial through to its

end.

At bottom, the trial court’s ruling leads us to the inescapable conclusion that

it “declared the mistrial without evincing [sufficient] concern for the double

jeopardy consequences of its actions,” “and without giving adequate consideration

to the less drastic alternative[s] . . . available.” Douglas, 488 A.2d at 145. It is no

small thing, contrary to the trial court’s reasoning, to simply start anew over a

defendant’s objections once their criminal trial is underway. It is an intrusion into

the defendant’s constitutional rights and should be done only as a last resort in only

the most extraordinary of circumstances. Because there was no manifest necessity

for declaring a mistrial over Walker’s objections, his retrial is now barred by the

Fifth Amendment’s Double Jeopardy Clause.

III. Conclusion

For those reasons, we hold that the trial court failed to exercise its sound

discretion when it “foreclose[ed] [Walker’s] option to go to the first jury.” Vega,

709 A.2d at 1175 (quoting Douglas, 488 A.2d at 145). Double jeopardy therefore

bars Walker’s retrial and the indictment against him is dismissed.