United States v. Chavez

976 F.3d 1178

• Court: Court of Appeals for the Tenth Circuit
• Decided: September 30, 2020
• Docket: 17-8096
• Status: Published

Opinion

FILED United States Court of Appeals Tenth Circuit

PUBLISH September 30, 2020 Christopher M. Wolpert UNITED STATES COURT OF APPEALS Clerk of Court

TENTH CIRCUIT

UNITED STATES OF AMERICA,

Plaintiff - Appellee,

v. No. 17-8096

RANDOLFO LIBRADO CHAVEZ, JR.,

Defendant - Appellant.

Appeal from the United States District Court for the District of Wyoming (D.C. No. 2:17-CR-00111-ABJ-2)

Neil D. Van Dalsem, Assistant Federal Public Defender (Julia L. O’Connell, Federal Public Defender, with him on the briefs), Office of the Federal Public Defender, Northern & Eastern Districts of Oklahoma, Tulsa, Oklahoma, for Defendant-Appellant.

Timothy Forwood, Assistant United States Attorney (Mark A. Klaassen, United States Attorney, with him on the brief), Office of the United States Attorney, Cheyenne, Wyoming, for Plaintiff-Appellee.

Before HARTZ, SEYMOUR, and HOLMES, Circuit Judges.

HOLMES, Circuit Judge. Defendant-Appellant Randolfo Librado Chavez, Jr., was convicted, after a

jury trial in the federal district court in Wyoming, of two counts of distributing

methamphetamine in violation of 21 U.S.C. § 841(a)(1) and (b)(1)(B). His appeal

from those convictions centers on the district court’s admission into evidence of

three transcripts, which purportedly were of audio recordings of incriminating

conversations that he had had in Spanish and in English. The district court did

not admit into evidence the audio recordings themselves or play the recordings for

the jury. Furthermore, the district court instructed the jury that they could not

question the accuracy of the English-language translations in the transcripts.

On appeal, Mr. Chavez challenges both the district court’s admission of the

transcripts and the jury instructions relating to them. He argues that the district

court (1) violated the best-evidence rule by admitting the transcripts into evidence

without admitting the recordings themselves, and (2) committed plain error by

instructing the jury to accept the accuracy of the translations in the transcripts.

As to Mr. Chavez’s first contention, we conclude that the district court’s

admission of the transcripts, without also admitting the recordings they purported

to transcribe, violated the best-evidence rule and constitutes reversible error.

Accordingly, exercising jurisdiction under 28 U.S.C. § 1291, we reverse and

remand the case to the district court with instructions to vacate Mr. Chavez’s

2 convictions and grant him a new trial. In light of this disposition, we do not

reach the merits of Mr. Chavez’s jury-instruction challenge.

I

We start by surveying (A) the events on which the audio recordings and

transcripts are predicated, (B) the contents of the recordings and the transcripts,

and (C) the district court’s admission of the transcripts at trial.

A

The relevant events trace back to July 2016, when Bryan Salas agreed to

serve the Wyoming Division of Criminal Investigation (“WDCI”) as a

confidential informant in exchange for the removal of a misdemeanor theft charge

from his record. Mr. Salas advised WDCI Special Agent Jason Ruby that he knew

of people in the Gillette, Wyoming, area involved in the sale of methamphetamine

and that he could arrange controlled purchases of the drug from them. To that

end, working with several WDCI agents, Mr. Salas arranged to purchase four

ounces of methamphetamine from a former co-worker named McKleen Miranda 1

on July 22. The agents gave Mr. Salas $4,000 to buy the four ounces of

methamphetamine, and they outfitted him with a wire that would record audio and

transmit it live to the agents.

1 Mr. Miranda was indicted with Mr. Chavez in this case but testified at trial for the government.

3 On July 22, the day of the controlled purchase, the wire recorded a call that

Mr. Salas allegedly received from Mr. Miranda. Because the conversation was

largely in Spanish, the agents did not understand what they were saying, so Mr.

Salas informed them later about the conversation’s contents. See R., Vol. III, at

102, 111–12 (Tr. Jessie Lile’s Test., dated Oct. 3, 2017). Mr. Salas told the

agents that Mr. Miranda had informed him that “they [i.e., Mr. Miranda and one

or more unidentified individual(s)] had the four ounces of methamphetamine” that

he had requested, and that “they wanted to meet at Walmart in Gillette.” Id. at

102. At the last minute, the location of the drug purchase changed from Walmart

to the parking lot of a nearby Old Chicago pizza restaurant.

The trial testimony presented slightly conflicting stories about precisely

what took place in the Old Chicago parking lot. Mr. Salas testified that he was

the first to arrive and that a white four-door car occupied by Mr. Miranda, Mr.

Chavez, and a person he did not know eventually pulled up next to his car. He

recounted that one of the white car’s windows rolled down and that Mr. Chavez

then handed him the methamphetamine in exchange for cash. Mr. Miranda, on the

other hand, testified that he, Mr. Chavez, and Mr. Miranda’s brother-in-law,

Carlos Dominguez, were the first to arrive in the white four-door car and that Mr.

Salas subsequently showed up and pulled up next to their car. As for the drug

handoff, Mr. Miranda testified that Mr. Chavez initially rolled down a window in

4 their white car, but when Mr. Salas’s car window failed to open, Mr. Salas opened

his car door instead and then exchanged Mr. Chavez’s methamphetamine for cash.

None of the law enforcement officers who were at the scene were able to

confirm either specific account. At most, one of them testified that he “could see

that there were definitely two occupants in the white car,” id. at 183 (Tr. Eric

Vos’s Test., dated Oct. 3, 2017), while another testified that he could hear “other

people talking” over the live wire transmission, id. at 106. But that was the full

extent of what they could ascertain: none of them could see who was inside the

white car, much less identify Mr. Chavez as one of its occupants. The testimony

is consistent, however, in conveying that the entire interaction between occupants

of the two cars was brief, lasting somewhere between approximately “15 or 30

seconds” and one minute. Id. at 142 (Tr. Bryan Salas’s Test., dated Oct. 3, 2017);

see id. at 106 (officer testifying that the meeting lasted “right around the one-

minute mark”).

After the cars parted, Mr. Salas and the agents returned to the WDCI office,

where the agents collected from Mr. Salas a bag containing methamphetamine.

The methamphetamine weighed 3.55 ounces, falling short of the four ounces that

Mr. Salas had been instructed to purchase. Accordingly, at the agents’ request,

Mr. Salas contacted Mr. Miranda about the methamphetamine shortage. Then,

pertaining to a separate matter, Special Agent Ruby asked Mr. Salas also to

5 contact Mr. Miranda to inquire into “another individual that [Special Agent Ruby]

was trying to activate a case on.” Id. at 231–32 (Tr. Jason Ruby’s Test., dated

Oct. 3, 2017).

Three days later, on July 25, Mr. Salas visited Mr. Miranda’s home.

Special Agent Ruby indicated that the purpose of the visit was for Mr. Salas to

inquire into the aforementioned person of interest. But while Mr. Salas was at

Mr. Miranda’s home, Mr. Salas acquired—without Special Agent Ruby’s

authorization—the approximately half-ounce of methamphetamine that had been

missing from the July 22 controlled purchase. Shortly after leaving Mr.

Miranda’s home with the approximately half-ounce of methamphetamine, Mr.

Salas contacted Special Agent Ruby and told him that he had just acquired from

Mr. Miranda “the methamphetamine that [had been] shorted from the deal on the

22nd [of July].” Id. at 231.

Mr. Salas then met Special Agent Ruby at the WDCI office at the special

agent’s request. Special Agent Ruby took from Mr. Salas what he had obtained

from Mr. Miranda, which Special Agent Ruby described as “a potato chip bag”

containing “a paper towel or napkin-type substance,” inside of which “was a

plastic bag that contained a crystalline substance similar to methamphetamine.”

Id. at 232. Special Agent Ruby “admonish[ed]” Mr. Salas for acquiring the

remaining methampetamine from Mr. Miranda without the participation of WDCI

6 agents. Id. at 234–35. Two days later, Mr. Salas was arrested for driving under

the influence (“DUI”).

Despite Mr. Salas’s unauthorized acquisition of the remaining

methamphetamine as well as his DUI arrest, WDCI agents continued to work with

him. Sometime shortly after July 22, Special Agent Ruby contacted Mr. Salas and

asked him to set up another controlled purchase of methamphetamine. Mr. Salas

told Special Agent Ruby that Mr. Chavez had already contacted him about

whether he needed more methamphetamine, and Special Agent Ruby asked Mr.

Salas to arrange to have Mr. Chavez sell him four more ounces of

methamphetamine. Mr. Salas and Mr. Chavez allegedly arranged for this second

purchase to take place on August 3 in the Walmart parking lot.

On August 3, WDCI agents again provided Mr. Salas with cash and

outfitted him with a wire. At Special Agent Ruby’s request, Mr. Salas called Mr.

Chavez to advise him that he was heading to Walmart. Mr. Salas arrived first at

Walmart, where he waited inside for Mr. Chavez. Eventually, a white car pulled

up to the store, and Mr. Salas got in.

Mr. Salas testified that there were two people in the car: Mr. Chavez and

Mr. Dominguez. According to Mr. Salas, a conversation took place in the car

between Mr. Chavez and himself, during which Mr. Chavez said that he did not

have all four ounces of methamphetamine that Mr. Salas had requested. Mr.

7 Chavez also allegedly said, at some point during this conversation, “I am the boss.

I can get you anything that you want.” Id. at 152. Mr. Salas further testified that

Mr. Chavez handed him the lesser amount of methamphetamine he had been able

to procure, and that Mr. Salas handed Mr. Chavez a corresponding amount of cash

in exchange.

Agents at the scene of the August 3 transaction presented a less definitive

picture of what transpired, testifying that they were unable to ascertain exactly

how many people other than Mr. Salas were in the white car, nor could they

identify precisely who those other people were. In terms of what actually

happened in the car, agents testified that they listened in on an approximately six-

minute-long conversation between the vehicle’s occupants, which was mainly

carried out in Spanish, but that they were unable to understand the Spanish-

speaking portions of it. Moreover, the agent charged with surveilling the white

car while Mr. Salas was inside it testified that he could not see into the car and

thus did not observe any “drugs being handed off and exchanged for money.” Id.

at 299, 302 (Tr. Chris McDonald’s Test., dated Oct. 4, 2017).

The agents eventually observed Mr. Salas exit the white car, which then

drove off. Mr. Salas walked back into Walmart. Shortly thereafter, Special

Agent Ruby picked him up and returned with him to the WDCI office. There, Mr.

8 Salas provided Special Agent Ruby with the methamphetamine that he had

purportedly received during the controlled purchase.

Meanwhile, a Gillette Police Department officer, who had been charged

with stopping the white car and identifying its occupants, pulled the car over a

short distance away from Walmart. The white car contained two people. The

person in its passenger seat identified himself as “Randolfo Chavez,” but he told

the officer that he did not have an ID on him. No one was arrested, nor were any

citations issued, and the officer ultimately let the car go. At trial, the officer

identified Mr. Chavez as the person who was sitting in the white car’s passenger

seat during the traffic stop. Id. at 305, 308 (Tr. Ryan Warney’s Test., dated Oct.

4, 2017).

B

It was against this backdrop that the government charged Mr. Chavez with

two counts of distributing methamphetamine in violation of 21 U.S.C. § 841(a)(1)

and (b)(1)(B). Count One pertained to the July 22 controlled buy, and Count Two

pertained to the August 3 controlled buy. At trial, the government’s case against

Mr. Chavez rested in substantial part on three transcripts: Trial Exhibits 15, 16,

and 17. These transcripts purportedly reflect audio recordings of conversations

that Mr. Chavez had in both Spanish and English in connection with the July 22

and August 3 controlled buys, except that the Spanish-language portions have

9 been translated into English. The district court’s admission of these transcripts is

at the center of this appeal. Accordingly, before turning to our analysis, we first

describe the content of these transcripts.

Exhibit 15 is a transcript that purportedly reflects the contents of an audio

recording from the July 22 controlled drug purchase. See id., Vol. IV, Ex. 15 at 1

(entitling exhibit, “Transcript of audio 7-22 Controlled Buy”). The transcript is

formatted as a table, consisting of nine rows and four columns. Each individual

row contains information pertaining to a statement 2 purportedly captured on the

recording. The first column identifies each statement by number, in ascending

numerical order. The second column identifies the purported speaker of the

statement: the speakers here are “V1” and “V2,” with V1 representing an

unnamed “Law Enforcement” officer and “V2” representing Mr. Salas. Id. The

third column purports to be a transcription of the statement in the language in

which the statement was originally made. And the fourth column purports to be a

translation of that statement into English, if the statement was originally made in

2 Certain rows in the exhibit display a single sentence fragment, while others contain a series of sentences. As a shorthand and for simplicity’s sake, we refer to the content contained in each row of the transcript—whether it is a sentence fragment or a full-blown monologue of sorts—as a “statement.” We follow a like approach with respect to Exhibits 16 and 17.

10 Spanish. Below is an example of one of the rows in the table capturing one of the

purported statements:

Id. at 2 (Statement 6).

Exhibit 15 features a total of eight such statements. The first row in the

table consists of a statement by an unnamed law-enforcement officer introducing

the recording. The officer describes it as the “recording of a control[led]

purchase of methamphetamine from Karin [sic] Miranda” on “July 22, 2016.” Id.

at 1 (Statement 1). The remaining seven statements correspond to the substance

of the recording, which purports to reflect Mr. Salas’s side of his conversations

with one or more unidentified person(s). According to the transcript, Mr. Salas

first tells his interlocutor(s) in Spanish that he is en route, then that he has arrived

and is near Office Depot, and finally that he will go to “Chicago” instead. Id.

(Statements 2, 4). He then says the following in English: “. . . the white car . . .

approaching.” Id. at 2 (Statement 5) (omissions in original). The transcript picks

up again when Mr. Salas says in English (presumably to the agents listening in),

“I got the product, I’m leaving . . . leaving that white Suzuki.” Id. (Statement 7)

11 (omission in original). Mr. Salas lastly describes in English the route that he is

taking as he drives toward the “office.” Id. (Statement 8).

Several features of Exhibit 15 merit comment. First, it does not define or

otherwise explain certain key, recurring words in the transcript—namely, the

terms “[u]nintelligible” and “[n]oisy.” In eight instances, the transcript indicates

that the voice on the recording is “[u]nintelligible.” Id. at 1–2 (Statements 2–4, 6,

8). But whether each marking of “[u]nintelligible” replaces one word, a single

sentence, an entire paragraph, or some other portion of the conversation is

altogether unclear. Furthermore, in four instances, the recording is dubbed

“[n]oisy.” Id. (Statements 2–3, 5, 8). But the transcript does not identify

precisely what the noises were, or whether they affected the quality of the

recording—and if so, to what extent. Moreover, in one particular instance, the

recording is characterized as “[u]nintelligible–noisy,” but, again, the transcript

provides no information as to what this term means. Id. at 1 (Statement 2).

Finally, as is the case with the other two transcripts at issue here (discussed

infra), this transcript is devoid of information regarding its authorship and other

aspects of its creation. The transcript contains no information addressing who

prepared it, how much time elapsed between the statements in each row, what

process its preparer used to create it, or how and why the statements were broken

up in the manner in which they were, among other missing contextual details.

12 2

Exhibit 16 is a transcript that purportedly reflects the contents of an audio

recording from the August 3 controlled drug buy. See id., Ex. 16 at 1 (entitling

exhibit, “Translation of 8-3-2016 Recorded Purchase”). Like the transcript

constituting Exhibit 15, the Exhibit 16 transcript is formatted as a table with rows

of information pertaining to each statement allegedly made on the recording. The

first column contains numbers in ascending order that enumerate each statement. 3

The second column identifies the voice making the statement: rows marked “B”

denote Mr. Salas as the purported speaker, rows marked “R” denote Mr. Chavez

as the purported speaker, and rows marked “C” denote Mr. Dominguez as the

purported speaker. Id. The third column reproduces what was purportedly said

on the recording, most of which is in Spanish. And the fourth column purports to

be an English translation of the text in column three. The row for Statement 30,

by way of example, is as follows:

3 Directly beneath the numbers listing each statement are various number ranges, denoting what might be time stamps, signifying the time that elapsed on the recording between the start and end of each statement. But we are unable to say definitively that these number ranges represent time stamps because neither the transcript itself, nor the testimony or other record materials pertaining to this exhibit, provide any clarity to this effect. In any event, what these number ranges represent is not material: that is to say, our disposition is unaffected by them. In offering a description of the exhibit, we simply make note of this feature that is unique to Exhibit 16.

13 Id. at 5 (Statement 30). 4

There are fifty-four such statements in Exhibit 16. Of these, eight are

marked as being completely “unintelligible.” Id. at 1–3, 5, 8, 11 (Statements 2, 8,

14, 16, 26, 28, 41, 53). The remaining statements supposedly reflect Mr. Salas’s

side of several seriatim phone calls that took place just before the August 3

controlled drug purchase, as well as a purported recording of the controlled

purchase itself involving Mr. Salas, Mr. Chavez, and Mr. Dominguez.

The first approximately dozen statements with an identifiable speaker are

attributed to Mr. Salas, and those statements appear to reflect his side of several

phone calls with another person about the meeting at Walmart. See id. at 1–4

(Statements 1, 3–7, 9–13, 15, 17–18). The remaining statements purportedly

reflect the conversation that took place in the white car among Mr. Salas, Mr.

Chavez, and Mr. Dominguez. See id. at 4–11 (Statements 19–54). According to

the transcript, Mr. Chavez tells Mr. Salas that he does not have everything that

4 We presume that the term “ininteligible” that appears in this row—as well as in various other parts of the transcript containing text in Spanish—is the Spanish-language equivalent of the English word “unintelligible.” See, e.g., Aplt.’s Opening Br. at 14–15 (pointing out examples in Exhibit 16 where “the source recording” is “unintelligible”).

14 Mr. Salas requested. See id. at 4 (Statement 19, 21, 23) (indicating Mr. Chavez

said that there was “a problem” because “[s]omebody from here screwed us” and

“gave us ‘base’ with half”). They engage in a back and forth on this issue, with

the transcript attributing the bulk of the statements to Mr. Chavez.

Toward the end of this discussion, Mr. Chavez purportedly says, inter alia,

“you did a really good job, you’re making some deals happen,” and, “[i]f you

keep being smart and if the more we get rid of and the more we can give you, we

will give you more. I can get pills.” Id. at 8 (Statements 43–44). The transcript

also attributes to Mr. Chavez the statement, “I’m like the boss. I’m your boss. I

can get you anything you want.” Id. at 10 (Statement 51). No one in the

conversation expressly mentions methamphetamine. The transcript shows that the

speakers conversed predominantly in Spanish, albeit with an occasional phrase or

sentence in English.

As was the case with Exhibit 15, certain aspects of Exhibit 16 merit

comment. In particular, in certain instances, the Spanish transcription in column

three and English translation in column four reflect what appear to be discrepant

accounts of what was said on the recording. Take, for example, Statement 51:

15 Id. Column three contains four words in Spanish, along with an “unintelligible”

word or phrase. Meanwhile, column four contains thirty-eight words in

English—over nine times as many words as the Spanish-language original in

column three. It is unclear how four words in Spanish plus an “unintelligible”

word or phrase translate into thirty-eight intelligible words in English.

Exhibit 16 embodies another kind of discrepancy between columns three

and four: the English words that respectively appear in the two columns do not

match up. Although witness testimony and the transcript itself suggest that most

of the statements captured on the recording were originally spoken in Spanish, in

a handful of instances, the speakers spoke in English. In such instances, column

three (i.e., the column capturing the statements as they were originally made) and

column four (i.e., the column reflecting the English translation of those

statements) both contain statements in English. And it would stand to reason that

the English statements in the two columns should be identical to one another.

The transcript, however, proves otherwise, as Statement 46 exemplifies:

16 Id. at 9 (Statement 46). 5 Column three includes, inter alia, the following

language attributed to Mr. Chavez: “you keep on doing this shit we gave you more

money because we fucked up.” Id. Column four, on the other hand, shows Mr.

Chavez allegedly stating, in relevant part, “[k]eep doing what we’re doing and

that will be more money for you.” Id. Query which of these two accounts

represents what was actually said by the voice identified as Mr. Chavez’s. Did he

say, “you keep on doing this shit,” or did he say, “[k]eep doing what we’re

doing”? And in that same vein, did he say, “we gave you more money because we

fucked up,” or did he say, “and that will be more money for you”? The transcript

leaves these questions unanswered. Such discrepancies pervade Exhibit 16.

Finally, Exhibit 17 purports to be a transcript of a phone call between Mr.

Chavez and Mr. Salas on August 3 before they arrived at the Walmart parking lot

where they allegedly engaged in a drug transaction that day. See id., Ex. 17 at 1

(entitling exhibit, “Translated Recorded Call on 8-3-2016”). Like Exhibits 15 and

16, Exhibit 17 is formatted in the style of a table containing rows for each

statement allegedly made on the recording. The first column enumerates each

statement by number, in ascending order, and the second column identifies the

5 The single bracket (i.e., “[”) in the fourth column for Statement 46 is apparently a typographical error.

17 speaker, with “V1” and “V2” corresponding, respectively, to what are purportedly

Mr. Chavez’s and Mr. Salas’s voices on the recording. Id. A third voice (“V3”)

is unidentified. Id. A fourth voice, which is identified only as “Officer,”

introduces the recording as a call “in reference to arranging the purchase of

approximately 4 ounces of methamphetamine.” Id. (Statement 1). Column three

purportedly reproduces what was said on the recording in its original language.

And the fourth column purports to be an English-language translation of the text

in column three. An illustrative example is the entry for Statement 10:

Id. (Statement 10).

Exhibit 17 features thirty-six such statements. Of those statements, sixteen

are attributed to Mr. Chavez. Id. at 1–2 (Statements 4, 6, 8, 10, 12, 14, 16, 18,

20, 22, 24, 26–27, 29, 31, 33). Thirteen statements are attributed to Mr. Salas.

See id. (Statements 5, 7, 9, 13, 15, 17, 19, 21, 23, 25, 28, 30, 32). Only an

“[u]nintelligible” statement is attributed to the unidentified third participant in the

conversation, “V3.” Id. at 1 (Statement 11). And four statements are attributed

to the unidentified “Officer.” Id. at 1–3 (Statements 1, 3, 34–35). A majority of

the statements—as reflected in column three of the transcript—are entirely or

partially in Spanish.

18 As for the substance of the transcript, the officer’s introductory statement

indicates that it is of “a recorded call to Randy . . . . in reference to arranging the

purchase of approximately 4 ounces of methamphetamine.” Id. at 1 (Statement 1).

The transcript then purports to record an oral exchange between Mr. Chavez and

Mr. Salas, wherein they discuss meeting and the price “for the four,” although

what “the four” refers to is not made clear. Id. at 2 (Statement 17).

C

At trial, the three above-described transcripts were introduced by the

government as Exhibits 15, 16, and 17, and various parts of the transcripts were

read into the record. This process of securing admission of the exhibits began

with the government calling Trooper Joseph Scimone “to testify as to the veracity

of the translations” in the transcripts. Id., Vol. III, at 175. Working through the

exhibits in seriatim fashion, the government asked Trooper Scimone in front of

the jury whether he had reviewed each of them and whether, based on his training

and experience, they constituted accurate translations of their respective

recordings. As to all three exhibits, Trooper Scimone answered in the

affirmative. At no point did the government ask whether he had prepared the

transcripts himself, nor did he indicate that this was the case. Indeed, at no point

19 during trial did the government disclose (through witness testimony or otherwise)

the identity of the preparer of the transcripts. 6

After concluding its examination of Trooper Scimone and while the jury

was in recess, the government moved the district court to find that Trooper

Scimone had offered (presumably sufficient) proof that the transcripts constituted

“accurate translations of the controlled purchases and the recorded phone calls.”

Id. at 94–95. Mr. Chavez said that he had “no objection” to the accuracy of the

translations, but he objected on the ground that “[t]here has yet to be foundation

and things of that sort.” Id. at 95. The district court did not rule on Mr. Chavez’s

objection.

Later in the proceedings, in the presence of the jury, the government called

Mr. Salas and Special Agent Ruby to authenticate the exhibits. Starting with

Exhibit 16, the government asked Mr. Salas whether he had previously reviewed

the transcript while listening to the concomitant audio recording; Mr. Salas

6 At oral argument, the government confirmed that Trooper Scimone did not prepare the transcripts and acknowledged that the record did not identify who did. Oral Arg. at 24:00–24:17. It then went on to note, without elaboration, that “interpreters” at a “firm” had prepared the transcripts’ Spanish-to-English translations but that it proved too logistically difficult to call those interpreters to testify at trial. Id. at 24:17–24:35. Accordingly, the government explained, it called Trooper Scimone—who, according to the government’s representations, had previously conducted an “interpretation” for the government—to review the recordings, make “a couple” of corrections to the transcripts, and testify at trial to their accuracy. Id. at 24:35–25:17.

20 testified that he had. Id. at 153. He further attested that based on his review of

the recording, the voice labeled “B” on the transcript was his own voice, and that

the voice labeled “R” on the transcript was Mr. Chavez’s. Id. at 153–54. Then,

the government elicited similar testimony from Mr. Salas as to Exhibit 17, with

Mr. Salas again testifying that he had reviewed the transcript and the recording

and that the transcript “accurately reflect[s] the voices [he] heard in” the

recording. Id. at 155. Specifically, he confirmed that the voices on the recording

corresponding to the letters “R” and “B” on the transcript were Mr. Chavez’s

voice and his own voice, respectively.

Following his cross-examination of Mr. Salas, in the jury’s presence, Mr.

Chavez lodged the following objection: “I realize[] that there ha[ve] been at least

three exhibits talked about -- transcription and interpretation of a document; this

witness also testified that according to him, it is what he heard, saw and this is an

accurate translation. I think that offends the Best Evidence Rule.” Id. at 171.

Accordingly, he requested that if the government was “going to introduce

conversations [from] during the buys, that the actual audio be played and

interpreted for the jury.” Id. Shortly thereafter, during a brief jury recess, Mr.

Chavez reiterated his objection as follows:

21 I think the Best Evidence Rule applies. We are talking about Rule 1002 and those transcripts -- those audio transcripts [sic] [7] are available. I just got a copy a little while ago. If this individual -- we are talking about the person that just testified [i.e., Mr. Salas], he is the one that says, “Well, you know, this is what I am hearing and this is what it was,” he is not the one who produced those written transcripts. That was another person. I think it was Trooper Scimone. I don’t think they worked in concert when they did that, and so it seems to me that due process would call that the defendant should be allowed to have the jury listen to the transcripts [sic]; let them actually hear what is being said, even though it may be in Spanish, then we have -- the State can always have it interpreted as they go piece by piece as to what is being said, or the defendant -- I’m sorry -- the witness that just testified.

Id. at 174–75. The government rejoined that the transcripts were admissible

because Trooper Scimone had testified about his Spanish expertise and attested to

the accuracy of the translations, and Mr. Salas had authenticated the identity of

the speakers and was present at the original conversations.

The district court, still outside of the jury’s presence, overruled Mr.

Chavez’s objection and admitted Exhibits 16 and 17 into evidence without playing

or admitting their underlying audio recordings. The court reasoned that the

7 During oral argument, the government represented that it had provided the audio recordings of the two controlled purchases to Mr. Chavez and had done so “well in advance of trial.” Oral Arg. at 20:45–20:48, 21:25–21:26. Given the context of his objection and his reference to the best-evidence rule and specifically Federal Rule of Evidence 1002, we believe that, in the quoted passage, when Mr. Chavez mentioned “audio transcripts” and subsequently urged “hav[ing] the jury listen to the transcripts,” he actually was referring to the audio recordings.

22 government had laid a sufficient foundation and that there was no dispute that the

transcripts accurately represented the contents of the audio recordings, such that it

could properly “receive those documents, and they, in fact, reflect and are the

evidence in this case.” Id. at 176. In so ruling, the district court went on to note

as follows:

Usually, a transcript is not evidence in and of itself, but the tape would be the evidence. To the extent that a transcript would differ in the minds of the jury from the transcript [sic], the jury would consider what they heard when the tape was played for them in the English language. This is not so with Spanish conversations that have been introduced.

The translated transcripts are the evidence you should rely on, and the jury is not free to reject the translation contained in the transcripts of the tape recordings given in this case and agreed to be as accurate translations.

Id. at 176–77.

Later on in the trial proceedings before the jury, the government examined

Special Agent Ruby in connection with authenticating Exhibit 15. Special Agent

Ruby confirmed that he had reviewed the Exhibit 15 transcript and listened to the

corresponding audio recording. He further attested that, based on his familiarity

with Mr. Salas’s voice from multiple conversations with him, he recognized the

primary voice on the recording as Mr. Salas’s. The government then moved to

admit Exhibit 15 into evidence. In response, Mr. Chavez objected, “Best

evidence rule. [Special Agent Ruby] didn’t produce this document. Half of it is

23 in Spanish. I don’t think he is qualified in any way, shape or form to speak or

understand Spanish.” Id. at 230. The government seemed to concede that Special

Agent Ruby was not qualified to understand Spanish. See id. But, the

government explained, Special Agent Ruby was authenticating the transcript

instead by attesting to his recognition of Mr. Salas’s voice on the recording (and

not by verifying the translation’s accuracy). Without explanation, the district

court overruled Mr. Chavez’s objection and admitted Exhibit 15 into evidence.

Having at that point moved all three exhibits into evidence, the government

turned its attention to asking Special Agent Ruby questions about Exhibit 17.

Preliminarily, during a sidebar, the government noted its intention to read certain

parts of the exhibit’s transcribed phone call to the jury. In response, Mr. Chavez

asserted that his “objection still stands that [the call’s conversation was] in

Spanish.” Id. at 241. The government advised the district court that it would

“just read the English.” Id. Overruling Mr. Chavez’s “continuing objection,” the

court concluded the sidebar by explaining, “I just don’t see how you would

require the jurors to become experts in the Spanish language.” Id. Resuming its

examination in front of the jury, the government arranged for selected portions of

the transcript to be read into the record, with Special Agent Ruby reading aloud

24 the statements in the transcript attributed to “V1” and counsel for the government

reading aloud the statements attributed to “V2.” 8

The government subsequently took up Exhibit 16 in similar fashion. See id.

at 259–61. At the beginning of his cross-examination of Special Agent Ruby, Mr.

Chavez mounted an objection to the government’s direct examination, arguing in

front of the jury that the government “should be required to comply with the Best

Evidence Rule” and noting that he was “particularly concerned about Exhibit 16.”

Id. at 266. He pointed out, by way of example, that the transcript for Statement

34 “has a few lines in Spanish” in the third column, but then it “has a long

paragraph in English” in the fourth (i.e., English-translation) column. Id. at

266–67. He then rhetorically asked, “[W]here are they getting this information?

It is not . . . a direct translation from the Spanish to the English.” Id. at 267. In

light of such issues, Mr. Chavez posited, “I think due process requires that the

defendant be allowed -- to allow the jury to actually hear that Spanish and have

the State provide a line by line translation about who is saying what, when.” Id.

When asked by the district court whether he was changing his position after

agreeing earlier in the proceedings that the transcript contained “an accurate

translation,” Mr. Chavez responded as follows: “[T]hat’s been one of my

8 Recall that during his examination, Mr. Salas indicated that, based on his review of the audio recording, “V1” and “V2” reflected the voices of Mr. Chavez and himself, respectively.

25 concerns, but my first -- Best Evidence Rule objection was pretty much the same

thing. I think the best evidence would be the actual audio recording being played

to the jury, and then interpreted line by line to the jury as it is being played.” Id.

The district court overruled this objection.

Later in the proceedings, after the government rested and before the jury

began deliberating, the district court provided final instructions to the jury. As

relevant here, the court issued Instruction 26, which read, in relevant part, as

follows:

The United States has introduced Spanish language transcripts . . . . While in a case involving English recorded conversations and transcripts, the jury is routinely instructed that they are not bound by the transcript, that is because every juror is just as capable as the person preparing the transcript to tell what is being said on the recording. This is not so with Spanish conversations . . . . The translated transcripts are the evidence you should rely on. You are not free to reject the translation contained in the transcripts of the tape recordings . . . . You are, however, free to give this evidence whatever weight or consideration you deem to be justified . . . .

Id., Vol. I, at 112 (Jury Instr. No. 26, dated Oct. 5, 2017) (emphasis added). Mr.

Chavez did not object to this instruction. See Aplt.’s Opening Br. at 35.

The jury convicted Mr. Chavez of both counts of distributing

methamphetamine. For each count, the district court sentenced him to ninety-five

months of imprisonment followed by five years of supervised release, to be served

concurrently. R., Vol. I, at 152–53 (Am. J., entered Dec. 18, 2017).

26 II

On appeal, Mr. Chavez advances two main challenges to his convictions.

First, he argues that the district court’s admission of the transcripts of the

recordings, without admitting the recordings themselves, violated the “best

evidence rule” under Federal Rule of Evidence 1002. Aplt.’s Opening Br. at 19.

This rule, Mr. Chavez maintains, “required the government to admit the actual

audio recordings,” with the transcripts serving “only as an aid in understanding

the audio recordings”—and not, as the district court ruled, as the exclusive

evidence of the conversations captured on those recordings. Id.; see also id. at

20–34. Second, Mr. Chavez challenges as plain error the district court’s

instruction to the jury that it was required to accept as accurate the translations

contained in the transcripts of the recordings; in his view, the jurors “should have

been instructed that they were free to reject the accuracy” of those translations.

Id. at 19–20; see also id. at 35–42.

The government disagrees on both scores. With respect to Mr. Chavez’s

best-evidence-rule challenge, the government rejoins that “[a]n audio recording of

a transaction entirely conducted in a language that jurors do not understand is not

the best evidence.” Aplee.’s Resp. Br. at 7; see also id. at 8–14. The government

contends that once “[t]he proper foundation for the translation[s] and

transcription[s] was provided,” “those properly authenticated transcripts were the

27 ‘best evidence’ of the transactions or conversations to which they related.” Id. at

7. As to Mr. Chavez’s jury-instruction challenge, the government argues that the

district court did not commit plain error because the instruction “was a correct

statement of the relevant law” that “properly informed the jurors to give the

transcripts the weight they deemed appropriate.” Id. at 8; see also id. at 14–22.

For reasons set forth below, we agree with Mr. Chavez that the district

court violated the best-evidence rule by admitting the transcripts of the audio

recordings without admitting the recordings themselves. In so doing, the district

court committed reversible error, and we thus reverse and remand the case to the

district court with instructions to vacate Mr. Chavez’s convictions and grant him a

new trial. As noted, because we are able to resolve this appeal on the basis of Mr.

Chavez’s best-evidence-rule challenge, we have no need to reach his jury-

instruction challenge and, accordingly, decline to do so.

A district court’s evidentiary decisions are reviewed for abuse of discretion.

United States v. Iverson, 818 F.3d 1015, 1019 (10th Cir. 2016); see also United

States v. Phillips, 543 F.3d 1197, 1203–04 (10th Cir. 2008) (reviewing appellants’

challenge to district court’s admission of evidence, brought pursuant to best-

evidence rule, for abuse of discretion). “Because evidentiary rulings are within

the sound discretion of the district court, this court will reverse only upon a

28 ‘definite and firm conviction that the lower court made a clear error of judgment

or exceeded the bounds of permissible choice in the circumstances.’” United

States v. Samaniego, 187 F.3d 1222, 1223 (10th Cir. 1999) (quoting Gilbert v.

Cosco, Inc., 989 F.2d 399, 402 (10th Cir. 1993)). Notably, “[a] district court

abuses its discretion if its decision is based upon an error of law.” United States

v. Tan, 254 F.3d 1204, 1207 (10th Cir. 2001) (quoting United States v. Cherry,

217 F.3d 811, 814 (10th Cir. 2000)).

The best-evidence rule, embodied in Federal Rule of Evidence 1002,

codifies a doctrine with a “long and venerable history” rooted in English common

law. 5 C HRISTOPHER B. M UELLER & L AIRD C. K IRKPATRICK , F EDERAL E VIDENCE

§ 10:17, Westlaw (database updated May 2020). Dating back to the 1700s, this

foundational doctrine requires a party seeking to prove the contents of any

writing, recording, or photograph to produce the originals. 9 See Seiler v.

9 The phrase “best-evidence rule” is something of a misnomer, as the rule does not demand that litigants furnish only the evidence that is categorically the “best” in a qualitative sense of that term. See, e.g., Best, WEBSTER ’ S T HIRD N EW I NTERNATIONAL D ICTIONARY (2002) (defining “best” as, inter alia, “excelling or surpassing all others of its kind in inherent quality”). Rather, the “origins and subsequent history of the doctrine make clear that it was always narrowly applied to regulate just evidence of the contents of writings” and, in later times, recordings, requiring that the contents of an available writing or (continued...)

29 Lucasfilm, Ltd., 808 F.2d 1316, 1318 (9th Cir. 1986); 2 K ENNETH S. B ROUN ET

AL ., M C C ORMICK ON E VIDENCE §§ 230–31, Westlaw (database updated Jan. 2020).

Stated differently, under this rule, evidence offered to prove the contents of an

original writing, recording, or photograph is not admissible, unless the original

itself is also admitted. See 2 B ROUN ET AL ., supra, § 231.

The animating purpose of this rule is to “promote accurate fact-finding.”

31 C HARLES A LAN WRIGHT ET AL ., F EDERAL P RACTICE & P ROCEDURE § 7162,

Westlaw (database updated Apr. 2020). And the rule fulfills this role in three

main ways. First, it alleviates “[t]he danger of mistransmission of critical facts

through the use of written copies or recollection.” 2 B ROUN ET AL ., supra, § 232.

In other words, it minimizes the risk of human error that inheres from the

production of so-called “secondary” evidence—that is, evidence other than the

original offered to prove all or part of the original’s contents; the original, of

course, constitutes “primary evidence” of its own contents. United States v.

9 (...continued) recording be proved by introducing into evidence the original itself. 31 C HARLES A LAN WRIGHT ET AL ., F EDERAL P RACTICE & P ROCEDURE § 7181, Westlaw (database updated Apr. 2020); see F ED . R. E VID . 1002 advisory committee’s note to 1972 Proposed Rules (noting that “[t]he rule is the familiar one requiring production of the original of a document to prove its contents” and that “[a]pplication of the rule requires a resolution of the question whether contents are sought to be proved” (emphasis added)). Thus, the best-evidence rule might be more aptly called “the ‘original document rule.’” 3 BARBARA E. BERGMAN ET AL., WHARTON’S CRIMINAL EVIDENCE § 15:1, Westlaw (database updated Nov. 2019).

30 Nunez, 532 F.3d 645, 651 (7th Cir. 2008) (using the term “primary evidence” to

describe an audio recording and to contrast it with a transcript of the recording);

see United States v. Alexander, 326 F.2d 736, 742 (4th Cir. 1964) (“As between a

supposed literal copy and the original, the copy is always liable to errors on the

part of the copyist, whether by wilfulness or by inadvertence”—a problem that

“wholly disappears when the original is produced.” (quoting 4 J OHN H ENRY

WIGMORE , E VIDENCE § 1179 (3d ed. 1940))); 2 B ROUN ET AL ., supra, § 232 & n.4

(“The burden on litigants of requiring them to introduce the writing if available is

outweighed by the increased accuracy of the factfinding process.” (quoting

C HRISTOPHER B. M UELLER ET AL ., E VIDENCE § 10.1 (6th ed. 2018))).

Second, the best-evidence rule guards against fraud. In the absence of the

original of a given source, witnesses may be tempted to “lie with impunity about

the original’s contents because the risk of detection is small”—a temptation that

the best-evidence rule removes by requiring the originals to be produced. 31

WRIGHT ET AL ., supra, § 7182; see United States v. Howard, 953 F.2d 610, 613

(11th Cir. 1992) (per curiam) (“The best evidence rule presupposes the existence

at one time of a decipherable original, and is intended to prevent fraud in proving

the contents of documents and/or recordings.”). Finally, even when there is no

human error or outright fraud, secondary evidence “may leave out crucial details

31 the omission of which may be difficult to discern”; that is, it raises the risk of

“incomplete[ness].” 31 WRIGHT ET AL ., supra, § 7182.

Enter the best-evidence rule, which implements the “elementary wisdom”

that an original is “a more reliable, complete and accurate source of information

as to its contents and meaning than anyone’s description.” Gordon v. United

States, 344 U.S. 414, 421 (1953); Seiler, 808 F.2d at 1319 (“[T]he hazards of

inaccurate or incomplete duplication are [among] the concerns addressed by the

best evidence rule.”).

Congress codified this common-law wisdom in the Federal Rules of

Evidence, specifically in Rule 1002. 5 M UELLER & K IRKPATRICK , supra, § 10:17

(“Rule 1002 represents the ‘teeth’ of the Best Evidence doctrine.”). In doing so,

Congress left it to the courts to read and apply Rule 1002’s terms according to

“their plain meaning.” Warger v. Shauers, 574 U.S. 40, 44 (2014) (explaining

that in construing a disputed Federal Rule of Evidence, the Court would “simply

accord [the rule’s] terms their plain meaning”). And here, Rule 1002’s meaning

is plain as day: it demands that courts exclude secondary evidence of an original’s

contents unless the original is in evidence. That bar is absolute, other than where

“[the Federal Rules of Evidence] or a federal statute provides otherwise.” F ED . R.

E VID . 1002. Consistent with this command, Congress has approved of specific

exceptions to the best-evidence rule, see F ED . R. E VID . 1003–07—but an

32 exception for foreign-language recordings is not among them. Cf. Antonin Scalia

& Bryan A. Garner, R EADING L AW : T HE I NTERPRETATION OF L EGAL T EXTS 93

(2012) (“[A] matter not covered is to be treated as not covered.”). Put simply,

under the plain meaning of Rule 1002, the best-evidence rule does not permit

courts to admit English-translation transcripts of foreign-language recordings

when the recordings themselves are not also in evidence.

Our circuit’s unbroken practice comports with the best-evidence rule’s

plain meaning. For over forty years, in cases in which we have affirmed the use

of a transcript of a recording at trial, we have done so only where the recording

itself was admitted. That is to say, we have consistently and unreservedly said

that a party seeking to introduce a transcript must furnish that transcript in

addition to—not in lieu of—the underlying recording. 10 And even then, we have

10 See, e.g., United States v. Hooks, 551 F.3d 1205, 1215–16 (10th Cir. 2009) (holding that the district court did not plainly err in allowing jury to use transcripts as an aid in listening to a recording without a limiting instruction); United States v. Davis, 929 F.2d 554, 559 (10th Cir. 1991) (holding that the district court did not abuse its discretion in admitting transcripts with “a proper cautionary instruction . . . that the tapes were the true evidence and that the transcripts were to be used only for clarification”); United States v. Mayes, 917 F.2d 457, 463 (10th Cir. 1990) (same); United States v. McIntyre, 836 F.2d 467, 471 (10th Cir. 1987) (holding that the district court did not abuse its discretion in allowing the jury to view transcripts of tape recordings in evidence in part because it instructed the jury “to disregard the transcript if particular words were indeterminable from the tapes”); United States v. Mittleider, 835 F.2d 769, 773 (continued...)

33 indicated that the transcript must serve merely “to assist the trier of fact” in

listening to the recording—as opposed to replacing the recording entirely. See

United States v. Devous, 764 F.2d 1349, 1354 (10th Cir. 1985) (emphasis added);

see, e.g., United States v. Hooks, 551 F.3d 1205, 1215–16 (10th Cir. 2009)

(holding that the district court did not plainly err in allowing jury to use

transcripts as an aid in listening to a recording without a limiting instruction);

United States v. Mittleider, 835 F.2d 769, 773 (10th Cir. 1987) (“[T]he court [did

not] err in allowing the jury to view the written transcripts of the tapes. The

transcripts were not admitted into evidence. The court preliminarily instructed

the jury that the transcripts were not evidence and had been furnished solely for

the purpose of guidance.”), abrogated on other grounds by Staples v. United

States, 511 U.S. 600 (1994).

10 (...continued) (10th Cir. 1987) (holding that the district court did not abuse its discretion in permitting the jury to view transcripts of tape recordings in evidence “for the purpose of guidance”), abrogated on other grounds by Staples v. United States, 511 U.S. 600 (1994); United States v. Devous, 764 F.2d 1349, 1354–55 (10th Cir. 1985) (holding that the district court did not abuse its discretion in admitting transcripts of “tapes of marginal quality” in evidence “to assist the trier of fact”); United States v. Lucero, 601 F.2d 1147, 1149–50 (10th Cir. 1979) (holding that the district court did not reversibly err in allowing jury to use transcripts as aids in listening to tapes in evidence); United States v. Watson, 594 F.2d 1330, 1336 (10th Cir. 1979) (holding that the district court did not abuse its discretion in permitting the jury “to use the transcripts only to assist them in listening to the tapes” in evidence and instructing them “not to consider the transcripts as evidence”).

34 More pointedly, we have followed this practice in cases involving foreign-

language recordings. Specifically, we have allowed English-translation

transcripts of foreign-language recordings only as aids in understanding the

admitted recordings themselves (i.e., the primary evidence). In other words,

under our practice, the English-translation transcript is permitted for use only in

conjunction with the foreign-language audio recording: it is the recording

itself—not the transcript of the recording—that constitutes the primary evidence.

See, e.g., United States v. Verdin-Garcia, 516 F.3d 884, 892 (10th Cir. 2008)

(“Appellants argue next that the district court should have held a Daubert hearing

on the admission of translations of intercepted [Spanish-language] phone calls,

and should have excluded them. . . . The recordings themselves were admitted as

substantive evidence; the translations were shown to the jury for demonstrative

purposes only.”); United States v. Gomez, 67 F.3d 1515, 1526–27 (10th Cir. 1995)

(same); see also United States v. Rivera, 778 F.2d 591, 600 (10th Cir. 1985)

(holding that district court did not abuse its discretion in admitting foreign-

language recordings with English-translation transcripts). Thus, the unmistakable

practice of this circuit comports with the best-evidence rule, as understood at

English common law and as embodied in Federal Rule of Evidence 1002. 11

11 Our pattern jury instructions, although not binding, provide telling confirmation of the state of our existing practice, in particular as that practice (continued...)

35 3

Our practice is also consistent with that of many of our sister circuits

regarding the admissibility of foreign-language recordings. Time and again, these

courts have likewise established that English-translation transcripts may be

admitted only where the underlying foreign-language recordings are themselves in

evidence. The First Circuit, for instance, has unambiguously held that insofar as

English-translation transcripts of foreign-language recordings are at issue, “[t]he

best evidence rule requires that the [foreign-language] recordings themselves

must be furnished.” United States v. Morales-Madera, 352 F.3d 1, 9 (1st Cir.

2003). The Second Circuit views the matter similarly. For example, in United

11 (...continued) pertains to foreign-language recordings. Those instructions persuasively confirm the clear import of our caselaw (as just surveyed supra): viz., to the extent that a party makes use of a transcript of a recording at trial, the recording itself must be in evidence. Indeed, under our pattern jury instructions, jurors are to be instructed that “transcripts are not evidence” and that rather “[t]he recordings themselves are the evidence.” 10 TH C IR . P ATTERN J URY I NSTR . § 1.40 (2018); see also id. (citing Gomez, 67 F.3d at 1527 n.15, a case involving recordings of partially Spanish-language conversations, for the proposition that “a cautionary instruction that the transcripts are only an aid in understanding the sound recording is preferred”). Stated otherwise, our pattern jury instructions counsel, consistent with the practice uniformly reflected in our caselaw, that it is the underlying recordings themselves—not the transcripts that purportedly reflect the recordings’ contents—that constitute the best evidence of their contents. And, therefore, where a litigant seeks to introduce proof of a recording’s contents, the litigant must admit the recording itself; this is a necessary condition for seeking to aid the jury’s understanding of the recording’s contents through the admission of a transcript of the recording.

36 States v. Ben-Shimon, 249 F.3d 98 (2d Cir. 2001) (per curiam), that court stated

that when “the recorded conversation is conducted in a foreign language, an

English language transcript may be submitted to permit the jury to understand and

evaluate the evidence [i.e., the recording],” id. at 101 (emphases added)—with

the clear implication being that the recording constitutes the best evidence and,

insofar as the concomitant transcript was furnished to the jury, the recording itself

must be admitted.

The Seventh Circuit has articulated largely the same understanding of the

role of foreign-language recordings in trial proceedings: “[t]he jury should be

instructed that it is the tape recording itself which is the primary evidence, [and]

that the transcript is to assist the jury in evaluating the primary evidence . . . .”

Nunez, 532 F.3d at 651; see id. (holding that where English-translation transcripts

had been prepared of recorded Spanish-language conversations purportedly

reflecting defendant’s participation in methamphetamine transactions, the district

court erred in instructing the jury “that it could afford as much weight as it felt

proper to the transcripts of the intercepted conversations” because “transcripts

should not ordinarily be given independent weight”). Likewise, the Third Circuit

has upheld the use of English-translation transcripts to perform the “limited role”

of aiding the jury in understanding foreign-language audio recordings, which were

themselves admitted into evidence. Gov’t of V.I. v. Martinez, 847 F.2d 125, 128

37 (3d Cir. 1988). So, too, the Eighth Circuit. See, e.g., United States v. Martinez,

951 F.2d 887, 888 & n.2 (8th Cir. 1991) (rejecting defendant’s challenge to the

jury’s use of transcripts where “the prosecution introduced three taped

conversations [some of which were in Spanish] between [defendant] and a police

informant recorded while the informant purchased cocaine from [defendant]” and

where English-translation “[t]ranscripts of the tapes were given to the jurors with

which to follow along while the tapes were played”).

The Eleventh Circuit, moreover, has affirmed the admission of English-

translation transcripts for the purpose of assisting the jury in understanding

admitted foreign-language recordings. See, e.g., United States v. Garcia, 854

F.2d 1280, 1283 (11th Cir. 1988) (noting, in a case involving Spanish-language

recordings, that “[o]ur case law provides district courts with the authority to allow

juries to read properly authenticated transcripts while listening to taped

conversations” and that the district court did not abuse its discretion in allowing

“the admission of the transcript . . . as an aid to the jurors” (emphasis added)).

The Ninth Circuit, for its part, has in certain cases affirmed the admission

of English-translation transcripts as substantive evidence—as opposed to merely

for assistive purposes (i.e., as jury aids); however, that court has not allowed a

trial court to admit such transcripts if the trial court has not also admitted into

evidence the foreign-language recordings themselves. Compare, e.g., United

38 States v. Abonce-Barrera, 257 F.3d 959, 962, 964 (9th Cir. 2001) (discerning no

abuse of discretion by the trial court where “the Spanish-language tapes were

played for the jury, and the English translations were read to the jury”), with

United States v. Franco, 136 F.3d 622, 625–28 (9th Cir. 1998) (approving of the

trial court’s admission of English-translation transcripts as substantive evidence,

along with the Spanish-language recordings). And more generally, other courts

have expressly held that “the [trial] court must ensure that the transcript is used

only in conjunction with the tape recording.” United States v. Holton, 116 F.3d

1536, 1543 (D.C. Cir. 1997).

As this survey reveals, many of our sister circuits, in analyzing the

admissibility of foreign-language recordings, are guided by the following

touchstone principle: English-translation transcripts may be admitted only in

addition to the underlying foreign-language recordings, not in lieu of them. 12 In

12 To be sure, the caselaw across—and even within—the foregoing circuits diverges on whether English-translation transcripts are to be admitted only for assistive purposes or as substantive evidence regarding the contents of the foreign-language recordings. But that distinction does not alter the unmistakable message that these cases uniformly convey: English-translation transcripts may be admitted only in addition to the foreign-language recordings, not in lieu of them. Whatever the evidentiary status of the transcripts, there is no question that the audio recordings must be admitted—as the primary evidence of the content of the foreign-language conversations.

39 other words, there is no question in these circuits that the audio recordings must

be admitted as evidence of the content of the foreign-language conversations. 13

13 Our research has unearthed several decisions of our sister circuits that appear to hold to the contrary. See, e.g., United States v. Estrada, 256 F.3d 466, 473 (7th Cir. 2001) (noting that “the district court saw no value in allowing a presumably English speaking jury to hear tapes that were recorded in Spanish”); United States v. Grajales-Montoya, 117 F.3d 356, 367 (8th Cir. 1997) (finding no abuse of discretion where the trial court denied defendant’s request to admit the tapes as well as the transcripts, on the view that the defendant “has suggested no reliable means of enabling people who do not speak Spanish to interpret inflections and tone, and we cannot think of any, either”); United States v. Valencia, 957 F.2d 1189, 1194 (5th Cir. 1992) (concluding that “an English translation transcript can be introduced into evidence without admitting or playing the underlying foreign language tape for the jury”), abrogated on other grounds by Apprendi v. United States, 530 U.S. 466 (2000). These cases do not give us pause. For one, these decisions espouse a view that, as clearly demonstrated supra, is distinctly in the minority among our sister circuits. Indeed, some of these decisions—like that of the Seventh Circuit panel in Estrada—appear to be outliers in their own circuits. Compare, e.g., Estrada, 256 F.3d at 473, with Nunez, 532 F.3d at 651 (where Spanish-language recordings were admitted into evidence, stressing that “[t]he jury should be instructed that it is the tape recording itself which is the primary evidence, [and] that the transcript is to assist the jury in evaluating the primary evidence”), and United States v. Magana, 118 F.3d 1173, 1181, 1184 (7th Cir. 1997) (noting, where “tapes played were of conversations in either English, Spanish, or Assyrian,” “that ‘[c]ourts possess wide discretion in determining whether to permit the jury to use written transcripts as aids in listening to tape recordings’” (alteration in original) (quoting United States v. Zambrana, 841 F.2d 1320, 1335 (7th Cir. 1988))), and 7 TH C IR . P ATTERN C RIM . J URY I NSTR . 3.15 (2012 ed., last updated Apr. 2019) (providing in the Seventh Circuit pattern jury instructions the following: “During the trial, [list name of language] language recordings were admitted in evidence. You were also given English transcripts of those recordings . . . .”). Furthermore, these decisions are thin on analysis and, in some instances, their language regarding the lack of necessity for admitting the foreign-language recordings themselves amounts to no more than dicta. For example, insofar as the analysis in Valencia, supra, speaks to the actual exclusion of foreign-language recordings from (continued...)

40 In sum, the best-evidence rule’s plain meaning, centuries of common-law

wisdom, over forty years of this court’s unbroken practice, and the decisions

aplenty of our sister circuits, counsel that transcripts of recordings—including

English-translation transcripts of foreign-language recordings—are inadmissible

unless the recordings, themselves, have been admitted. As such, that approach

drives our analysis here, to which we now proceed.

Applying the foregoing principles, we hold that the district court violated

the best-evidence rule and in so doing committed reversible error. We explain

our holding in two parts. First, we assess whether the district court abused its

discretion by admitting the transcripts of the recordings without admitting the

recordings themselves in violation of the best-evidence rule. We conclude that

the district court did so err. 14 Second, we assess whether this error was

13 (...continued) evidence, it is dictum because the foreign-language recording there was admitted into evidence, even though it was not played for the jury. See 957 F.2d at 1193. In sum, these contrary decisions lack the power to persuade and, consequently, do not give us pause, let alone alter the substance of our analysis. 14 For clarity’s sake, we pause to underscore the limited scope of our holding. What we address here is the question of whether a district court—facing the same or similar circumstances involving at least partly foreign-language recordings and English-translation transcripts—would be obliged to admit into evidence the foreign-language recordings themselves under the best-evidence rule as the primary evidence of the contents of those recordings. Mr. Chavez’s core relief for his first claim hinges on the answer to that question. See, e.g., Aplt.’s (continued...)

41 14 (...continued) Opening Br. at 19 (“It was error for the district court to allow the government to admit the transcripts instead of the recordings.”). And we answer it in the affirmative. What we do not address is how a district court, in circumstances such as these, may properly regulate the use of such foreign-language audio recordings once they are admitted into evidence. And it follows perforce that we do not hold that, in the same or similar circumstances, district courts must routinely play the foreign-language audio recordings in their entirety for the jury. The use-of-audio-recordings question is not before us because the district court here did not admit the recordings into evidence at all. And we recognize that the permissible uses to which parties may put such evidence, once admitted, is a matter that lies in the first instance within the sound discretion of the district court. See F ED . R. E VID . 611(a)(1), (2) (“The court should exercise reasonable control over the mode and order of . . . presenting evidence so as to . . . make those procedures effective for determining the truth” and to “avoid wasting time . . . .”); Silver v. Cormier, 529 F.2d 161, 164 (10th Cir. 1976) (“The extent, by repetition or otherwise, of the admission of competent evidence is a matter peculiarly within the discretion of the trial judge and will not be disturbed by this court absent a clear abuse of such discretion.”); see also F ED . R. E VID . 403 (providing that “[t]he court may exclude relevant evidence if its probative value is substantially outweighed by a danger of,” inter alia, “confusing the issues, . . . undue delay, wasting time, or needlessly presenting cumulative evidence”); F ED . R. E VID . 611 advisory committee’s note to 1972 Proposed Rules (“The ultimate responsibility for the effective working of the adversary system rests with the judge.”); cf. United States v. Lujan, 603 F.3d 850, 859 (10th Cir. 2010) (“[T]he district court can use its discretion to ensure the jury receives the evidence that is highly probative . . . , while at the same time excluding details carrying lesser probative value that . . . poses a real risk of causing unfair prejudice.”). And there could be a range of appropriate uses for such audio-recording evidence depending on the circumstances of the case, the parties’ theories, and the settled practices in the particular circuit. Compare Franco, 136 F.3d at 628 (“We find no abuse of discretion in the district court’s decision not to play the tapes in full for the jury. The tapes were lengthy and the conversations were conducted in a language [i.e., Spanish] that most or all of the jury did not understand.”), with United States v. Cruz, 765 F.2d 1020, 1024 (11th Cir. 1985) (“The district court played the tape recording for the jury and had an interpreter signal the jury when it was appropriate to turn the pages of the transcript. This procedure enabled the (continued...)

42 harmless. We ultimately hold that it was not harmless, compelling us to reverse

and remand this case to the district court. We undertake this analysis in turn.

Our analysis begins with determining, first, whether the district court

abused its discretion in admitting the transcripts while excluding the recordings in

violation of the best-evidence rule. Before conducting this inquiry, however, we

must determine, as a threshold matter, whether the best-evidence rule applies here

at all. As noted, Rule 1002 generally requires the admission of the “original . . .

recording” where the proponent seeks “to prove its content.” F ED . R. E VID . 1002.

For purposes of this rule, a “‘recording’ consists of letters, words, numbers, or

their equivalent recorded in any manner.” F ED . R. E VID . 1001(b). The “content”

of an audio recording, in turn, refers to “the spoken words or other audible sounds

it contains.” 31 WRIGHT ET AL ., supra, § 7184. Finally, under this rule, “[a]n

14 (...continued) jury to detect changes in voice modulation and note any hesitancies or other characteristics which might give meaning to the tape recording.”). Notably, in Franco, the Ninth Circuit discerned no reversible error where (1) “the district court made clear that it would entertain requests to play specific portions of tape when proffered at the appropriate point in the proceedings; [but] no such proffer was made”; and (2) the foreign-language “tapes were not sent into the jury room, but the jury was advised that it could listen to tapes upon request; [and] no request was made.” 136 F.3d at 628–29. On remand, the district court here should determine in the first instance—after considering the parties’ arguments and the settled practice of our circuit—the appropriate use of the audio recordings during Mr. Chavez’s new trial.

43 ‘original’ of a . . . recording means the . . . recording itself or any counterpart

intended to have the same effect by the person who executed or issued it.” F ED .

R. E VID . 1001(d).

Here, the government unquestionably sought to prove the contents of the

recordings. That is, it sought to prove the words purportedly spoken by Mr.

Chavez and others in connection with the controlled methamphetamine purchases,

as those words were captured on the recordings; indeed, as explicated below,

these recordings were central to the government’s case at trial. See, e.g., R., Vol.

III, at 39–43 (government grounding its opening statement repeatedly on the

recorded conversations that allegedly took place between Mr. Chavez and others

in connection with the controlled purchases); see also id. (government reiterating

throughout opening statement that these purportedly incriminating conversations

were captured on audio recordings, stressing, for instance, the following: Law

enforcement agents “are listening to the wire. They can hear it over real time.

They can hear what is happening.”). Thus, in seeking to prove the recordings’

contents, the government triggered the best-evidence rule. F ED . R. E VID . 1002

advisory committee’s note to 1972 Proposed Rules (noting that “[t]he rule is the

familiar one requiring production of the original of a document to prove its

contents” and that “[a]pplication of the rule requires a resolution of the question

whether contents are sought to be proved”). Furthermore, none of the best-evidence

44 rule’s exceptions apply here. See, e.g., FED. R. EVID. 1003 (allowing the use of

duplicates); FED. R. EVID. 1004 (enumerating exceptions to the requirement of an

original). At bottom, the best-evidence rule is controlling here.

Having determined that the best-evidence rule applies, we must determine whether

it was violated. We conclude that it was. Because the government sought to prove the

contents of the three recordings concerning the alleged drug buys from Mr. Chavez, it

was required under the best-evidence rule to secure admission of the “original” recordings

themselves; that is, the three audio recordings that the translation transcripts purported to

reflect. This, the government did not do. At trial, it proffered instead the transcripts, and

the transcripts only, as proof of the recordings’ contents. See, e.g., R., Vol. III, at 175–76

(moving to admit Exhibits 16 and 17 as substantive evidence of what was allegedly said

and heard on the recordings); see also, e.g., Aplee.’s Resp. Br. at 3 (noting that “the

United States offered translated transcripts of the recorded controlled purchases as

evidence”). As the government itself has recognized, at no point did it even seek the

admission of the underlying recordings. See Oral Arg. at 20:32–20:41 (“The United

States did not admit the tapes into evidence. The United States could have; we had the

tapes; we had the tapes available . . . , but we did not introduce those tapes at trial.”). And

in turn, the district court, over Mr. Chavez’s repeated objections on best-evidence-rule

45 grounds,15 admitted all three transcripts as the sole proof of the recordings’

contents—without admitting (or even listening to) the recordings. See id. at 23:44–23:49

(“The district court never reviewed those recordings . . . . We did not [proffer them].”).

The district court’s decision to admit the transcripts—transcripts that were

proffered to prove the contents of the original recordings—without admitting the original

recordings themselves plainly flies in the face of the best-evidence rule and was an abuse

of the court’s discretion. See Wright v. Farmers Co-Op of Ark. & Okla., 681 F.2d 549,

553 n.3 (8th Cir. 1982) (“The transcription offered as evidence in this case clearly was

not the ‘original’ statement, nor could it be considered a ‘duplicate.’ Moreover, plaintiffs

clearly intended it as proof of the original statement’s content. Thus, absent the

applicability of an alternative provision for admissibility, Rule 1002 would mandate

exclusion of the transcribed statement.” (citations omitted) (first quoting FED. R. EVID.

1001(3), and then quoting FED. R. EVID. 1001(4), 1003)).

In a seeming attempt to elude the best-evidence rule’s ambit, the government

makes much of the fact that the recordings “were all in Spanish,” contending that this

“alters the analysis under the Best Evidence Rule.” Aplee.’s Resp. Br. at 10–11. Under

15 See, e.g., R., Vol. III, at 171 (objecting to the government’s prospective introduction of the three transcripts without the recordings because “that offends the Best Evidence Rule”); id. at 174–75 (lodging objection to the government’s introduction of the transcripts constituting Exhibits 16 and 17 on the ground that “the Best Evidence Rule applies”); id. at 230 (mounting objection to government’s introduction of Exhibit 15 in light of the “[b]est evidence rule”).

46 its view, recordings “in a language that jurors do not understand is not the best evidence,”

id. at 7; rather, “[w]hen a recording captures a foreign language conversation the

[English-translation] transcript controls,” id. at 12 (quoting United States v. Ramirez, 576

F. App’x 385, 388 (5th Cir. 2014) (per curiam) (unpublished)); see also id. at 13 (warning

that construing the best-evidence rule to require admission of the underlying foreign-

language recordings would have the “inevitable effect of . . . impos[ing] on future juries

the impossible duty to endure—and to try and make sense of—what to them will surely be

nothing more than indecipherable noise”). The government attempts to apply these

principles here, arguing that “the ‘best evidence’ in this case was actually not an audio

recording of [a] Spanish language conversation of the two controlled buys and set up call,

none of which the jurors would have understood.” Id. at 12–13; see also id. at 11

(reasoning that it would have “made no sense whatsoever to play those recordings to the

jurors, who almost certainly would not have understood any of it[,]” because “the audio

recordings of the controlled buys and the set up call were all in Spanish”). “Rather,” it

posits, “the ‘best evidence’ was the translated transcripts that [Mr. Chavez] agreed w[ere]

accurate.” Id. at 13.

The government’s argument, however, holds no water. For starters, it evinces a

fundamental misunderstanding of the legal import of the best-evidence rule. More

specifically, the tacit premise of the government’s argument is that the best-evidence rule

calls for the admission of the evidence of the recordings’ contents that is the “best” in a

47 qualitative sense. And, thus, the government reasons that because the recordings were in

a foreign language that the jury did not understand, the English-translation transcripts

actually were the best evidence, not the recordings themselves. But the government’s

premise is fatally flawed. Despite its somewhat misleading name, the best-evidence

rule’s concern is not with the admission of the best evidence in a qualitative sense. See

United States v. Smith, 566 F.3d 410, 413 (4th Cir. 2009) (“The [best-evidence] Rule

exists to afford guarantees against inaccuracies and fraud by requiring that the original of

the document be offered, subject to exceptions . . . . Thus it is more accurate to refer to

Rule 1002 as the ‘original document rule,’ not the ‘best evidence rule.’”); see also FED. R.

EVID. 1001 advisory committee’s note to 1972 Proposed Rules (referring to the

“misleading[ly] named ‘best evidence rule’”); supra note 9. Rather, the “origins and

subsequent history of the doctrine make clear that it was always narrowly applied to

regulate just evidence of the contents of writings” and, in later times, recordings. 31

WRIGHT ET AL., supra, § 7181. Thus, whether the transcripts here are the “best” evidence

of the contents of the recordings in a qualitative sense—as the government appears to

believe—has no bearing on our analysis under the best-evidence rule, and we must

therefore reject the government’s argument in this respect.

The government’s argument also misconstrues the best-evidence rule’s scope. In

contending that “the translation-foreign language element alters the analysis under the

Best Evidence Rule,” see Aplee.’s Resp. Br. at 10, the government is, in effect, arguing

48 that there is an exception to the clear dictates of the best-evidence rule for recordings in a

foreign language. Under the plain language of the Federal Rules of Evidence, however,

no such exception exists, nor has one been carved out in our caselaw. To the contrary, as

we have explicated at length, supra, the plain meaning of the Rules of Evidence, our

abiding common-law principles, and a long line of cases in this and other circuits make

pellucid that “evidence concerning the contents of the recording of [a] foreign-language

conversation is subject to the best evidence rule.” 31 WRIGHT ET AL., supra, § 7184

n.17.1. Put simply and directly, for purposes of our best-evidence-rule analysis, it is

immaterial whether the recordings were in Spanish, and the government’s contentions to

the contrary must be rejected.

In addition to its faulty legal underpinnings, the government’s argument is also

predicated on a false factual premise. The government has repeatedly represented and

implied on appeal that “the recordings were completely in Spanish.” Oral Arg. at

17:55–17:57; see, e.g., Aplee.’s Resp. Br. at 11 (“[T]he audio recordings . . . were all in

Spanish.”); see also, e.g., id. at 2 (claiming that “[b]oth [Mr. Salas] and [Mr. Chavez]

spoke Spanish during the entirety of th[e] [July 22] transaction,” the audio recording of

which “was translated and transcribed into Exhibit 15”). But the transcripts themselves

squarely belie the government’s representations,16 as they show that the contents of each

16 We accept the contents of the transcripts at face value strictly for purposes of the instant discussion. As explicated below, however, we ultimately (continued...)

49 recording were at least partly in English. For example, of the forty-six intelligible

statements in Exhibit 16, no fewer than seven of them—about fifteen percent of the

transcript’s content—are either partly or entirely in English. See R., Vol. IV, Ex. 16 at

1–3, 6, 8–10 (Statements 6, 12, 35, 44–47). But see Aplee.’s Resp. Br. at 2–3

(representing that “[Mr. Salas] and [Mr. Chavez] both spoke Spanish during the entirety

of the [August 3] transaction,” the audio recording of which was purportedly evinced in

the Exhibit 16 transcript).

The best-evidence rule’s command is clear: a court may not admit a transcript of a

recording to prove the contents of that recording unless the recording itself is admitted or

an express exception to the best-evidence rule applies. The district court plainly violated

that command and, as such, abused its discretion. Tan, 254 F.3d at 1207 (noting that “[a]

district court abuses its discretion if its decision is based upon an error of law” (quoting

Cherry, 217 F.3d at 814)).17

16 (...continued) find them to be saddled with serious problems that substantially call into question their integrity. 17 Notably, Federal Rules of Evidence 702 and 403—the rules upon which the Dissent predicates its position—play no role whatsoever in the government’s contrary arguments for affirmance. See F ED . R. E VID . 702 (enumerating the requirements for expert testimony); F ED . R. E VID . 403 (providing that “[t]he court may exclude relevant evidence if its probative value is substantially outweighed by a danger of,” inter alia, “unfair prejudice, confusing the issues, [or] misleading the jury”). Though, as we have seen, the government vigorously advocates for affirmance, it mentions nary a word in its briefing concerning the expert-testimony principles of Rule 702 or the various concerns, (continued...)

50 17 (...continued) such as juror-confusion, that animate Rule 403. Indeed, the government’s briefing never cites those rules at all. And with respect to Rule 702 specifically, neither party even intimates in its briefs that the transcripts at issue here were admissible as expert testimony. All this is not entirely surprising, given the circumstances of this case. As the Dissent concedes—albeit only in its closing five pages—the government did not present Trooper Scimone as an expert, nor did the district court qualify him as such. Indeed, the government never gave formal notice that Trooper Scimone would be offering expert testimony, much less move to admit the transcripts as Trooper Scimone’s expert testimony. And, as for Rule 403, the district court here was never positioned to conduct the multi- factor balancing that Rule 403 requires because, as the government acknowledged during oral argument, it never even provided the audio recordings to the district court for its consideration. See Oral Arg. at 23:44–23:49 (“The district court never reviewed those recordings . . . . We did not [proffer the recordings].”); see also United States v. Mangiameli, 668 F.2d 1172, 1176 (10th Cir. 1982) (noting that “the trial court, under Rule 403, may exclude otherwise admissible evidence after balancing its probative value against certain competing considerations set forth in the rule and concluding that the costs of the evidence outweigh its benefits” (emphasis added)); accord United States v. Espinosa, 771 F.2d 1382, 1406 (10th Cir. 1985); cf. United States v. Willis, 826 F.3d 1265, 1271, 1273–74 (10th Cir. 2016) (articulating the factors that must be balanced in a sexual-assault case to which Rule 413 applies).

Undaunted, the Dissent crafts arguments for affirmance completely sua sponte and, more specifically, without the benefit of the parties’ adversarial exchange. As a jurisprudential matter, this is imprudent, and—under these circumstances, where we have demonstrated (with reference to the parties’ arguments) that Mr. Chavez’s conviction rests unjustly on legal error—it is troubling. “[T]he adversary system is a cornerstone of our jurisprudence.” Bryan A. Garner et al., T HE L AW OF J UDICIAL P RECEDENT 226 (2016). Indeed, the very notion of “judicial precedent implies that the point to the decision . . . should have been argued by opposing counsel.” Id. (quoting John F. Dillon, T HE L AWS AND J URISPRUDENCE OF E NGLAND AND A MERICA 233 (1894)). This adversarial testing serves important ends: notably, it increases the odds that the court will “reach the ‘correct’ decision because the advocates will uncover and present more useful information and arguments . . . than the court would develop on its own.” (continued...)

51 2

That the district court abused its discretion does not end our inquiry, however.

Even where we have determined that the district court has so erred, “we still may not

grant relief if the district court’s error was harmless.” United States v. Washington, 653

F.3d 1251, 1270 (10th Cir. 2011); United States v. Blechman, 657 F.3d 1052, 1067 (10th

Cir. 2011) (“We will not reverse a defendant’s conviction on the basis of the district

court’s erroneous admission of . . . evidence if the error was harmless to the defendant.”);

see FED. R. CRIM. P. 52(a) (“Any error, defect, irregularity, or variance that does not

17 (...continued) Adam A. Milani & Michael R. Smith, Playing God: A Critical Look at Sua Sponte Decisions by Appellate Courts, 69 T ENN . L. R EV . 245, 273 (2002). After all, “[c]ounsel almost always know a great deal more about their cases than we do.” Greenlaw v. United States, 554 U.S. 237, 244 (2008) (quoting United States v. Samuels, 808 F.2d 1298, 1301 (8th Cir. 1987) (R. Arnold, J., concurring in denial of reh’g en banc)). To be sure, appellate courts have the “discretion to affirm on any ground adequately supported by the record.” Elkins v. Comfort, 392 F.3d 1159, 1162 (10th Cir. 2004). But ordinarily, in exercising that discretion, we have been—as a matter of basic fairness—“guided” by whether the parties have “fully briefed and argued” the alternate ground, and whether they have had “a fair opportunity to develop the factual record.” Brown v. Perez, 835 F.3d 1223, 1236 (10th Cir. 2016) (quoting Elkins, 392 F.3d at 1162). These circumstances are conspicuously absent here. Accordingly, we deem the Dissent’s sua sponte handiwork to be, again, not only imprudent, but—under these circumstances, where Mr. Chavez’s demonstrably infirm conviction hangs in the balance—troubling. Because the parties have not had an opportunity to engage with the merits of the Dissent’s arguments, neither will we.

52 affect substantial rights must be disregarded”). “A non-constitutional error, such as a

decision whether to admit or exclude evidence, is considered harmless ‘unless a

substantial right of [a] party is affected.’” United States v. Charley, 189 F.3d 1251, 1270

(10th Cir. 1999) (alteration in original) (quoting FED. R. EVID. 103(a) (1999)). An error

affecting a substantial right of a party is an error that “had a substantial influence on the

outcome or leaves one in grave doubt as to whether it had such effect.” United States v.

Roach, 582 F.3d 1192, 1207 (10th Cir. 2009) (quoting United States v. Bornfield, 145

F.3d 1123, 1131 (10th Cir. 1998)). In other words, “reversal is appropriate where an error

has a substantial influence on the outcome of a trial or leaves one in grave doubt as to

whether it had such effect.” United States v. Richter, 796 F.3d 1173, 1197 (10th Cir.

2015). “By ‘grave doubt’ we mean that, in the judge’s mind, the matter is so evenly

balanced that he feels himself in virtual equipoise as to the harmlessness of the error.”

O’Neal v. McAninch, 513 U.S. 432, 435 (1995).

In determining whether an error had such an effect on the outcome of a trial, we

“should not consider the error in isolation, but rather should consider it in the context of

the entire record.” United States v. Irving, 665 F.3d 1184, 1209 (10th Cir. 2011) (quoting

28 MOORE’S FEDERAL PRACTICE § 652.03[1], at 652–8 (3d ed. 2011)). To that end, our

analysis entails an examination of the following: (1) “the context, timing and use of the

erroneously admitted evidence at trial,” United States v. Hanzlicek, 187 F.3d 1228, 1237

(10th Cir. 1999); (2) “how [that evidence] compares to properly admitted evidence,” id.;

53 and (3) “whether the district court’s instructions to the jury mitigated [the] error,” Richter,

796 F.3d at 1197; see United States v. Hands, 184 F.3d 1322, 1332 (11th Cir. 1999).

“The burden is on the government to establish the harmlessness of any error.”

Washington, 653 F.3d at 1270; United States v. Holly, 488 F.3d 1298, 1307 (10th Cir.

2007) (“It is well-established that the burden of proving harmless error is on the

government.”).18

18 The government includes some harmless-error statements in its brief. But these statements amount to little more than a few cursory sentences baldly insisting that any error here “would necessarily have been harmless on this record,” accompanied by a single case citation, providing a generic recitation of the harmless-error standard. Aplee.’s Resp. Br. at 13; see also id. at 13–14 (asserting, without elaboration, that “it would be purely fanciful to even imagine that such an error could have had any discernible effect on the outcome of this case, never mind the kind of ‘substantial influence’ on the jury’s verdict that would be required for it to have been anything other than harmless” (quoting United States v. Nance, 767 F.3d 1037, 1043 n.5 (10th Cir. 2014))). Simply claiming that an error is “harmless,” however, does not make it so; more is required from the government to secure harmless-error review. Seeing as the government has not meaningfully engaged with the issue of harmlessness, we question whether it has even made a harmless-error argument adequate to warrant our review. See, e.g., United States v. Faust, 795 F.3d 1243, 1248 n.4 (10th Cir. 2015) (noting that, where a party “does not elaborate on [its] bare assertion[,] . . . we may rightly deem any argument that could be based on it to be waived”); United States v. Pursley, 577 F.3d 1204, 1231 n.17 (10th Cir. 2009) (“Under our precedent, this skeletal reference is insufficient to raise the [appellant’s] concern as a discrete appellate issue.”). We need not definitively opine on this waiver question, however, because even if we exercised our discretion to conclude that the government’s skeletal statements are sufficient to raise the issue of harmlessness for our review, they lack the power to persuade us in light of our analysis below and thus do not alter our conclusion. See United States v. Lente, 647 F.3d 1021, 1037 (10th Cir. 2011) (“Harmlessness must be proven by a preponderance of the evidence, and the burden of making this showing falls on (continued...)

54 Viewing the district court’s erroneous admission of the transcripts (without the

audio recordings) through this lens, we hold that the government has failed to carry its

burden. That is to say, the error here was not harmless. Several reasons underlie our

conclusion: (1) the transcripts assumed a central role in the government’s case, which is

especially problematic because their defects and omissions make their integrity and

soundness questionable, (2) the government’s other evidence was far from overwhelming,

and (3) the district court’s jury instructions failed to mitigate the error and, indeed,

probably exacerbated it. We expound on these reasons below.

In performing our harmless-error analysis, we begin by considering the nature of

the erroneously admitted evidence, with an eye toward “the context, timing and use of the

erroneously admitted evidence at trial.” Hanzlicek, 187 F.3d at 1237. As part of this

inquiry, we consider “the strength, importance, and pervasiveness of the erroneously

admitted [evidence].” Richter, 796 F.3d at 1197. That is to say, a determination that the

wrongly admitted evidence figured prominently in the government’s case militates in

favor of holding that the error was not harmless. See, e.g., United States v. Tome, 61 F.3d

18 (...continued) the beneficiary of the error—in this case, the government.” (quoting United States v. Cerno, 529 F.3d 926, 939 (10th Cir. 2008))); see also, e.g., id. at 1038 (“The government bears the burden to show that the error did not affect the sentence. Its argument consists of assertions that the district court considered the § 3553(a) factors and that a different sentence on remand is ‘highly unlikely.’ That is not enough.”).

55 1446, 1455 (10th Cir. 1995) (concluding that the improper admission of evidence was not

harmless where that evidence “was extremely compelling”); Specht v. Jensen, 853 F.2d

805, 808 (10th Cir. 1988) (en banc) (holding that the erroneous admission of expert

testimony was prejudicial where the testimony was “pervasive”).

Here, in prosecuting Mr. Chavez for methamphetamine distribution, the

government’s case was predicated on two events: the July 22 and August 3 controlled

purchases. See, e.g., R., Vol. III, at 39–40 (the government, in previewing its case during

opening statements, informing the jurors that throughout trial they would “hear evidence

about two incidents of distribution of a controlled substance, specifically

methamphetamine”). The government argued that Mr. Chavez was the so-called “boss”

or lynchpin in connection with the July 22 and August 3 transactions, acting as “the one

that made the deliveries [of methamphetamine]” and as “the one [Mr. Salas] purchased

methamphetamine from on those two dates.” Id. at 44–45.

The transcripts played a significant role in establishing this theory at trial—doing

so, not infrequently, in a problematic and potentially misleading manner. To begin,

Exhibits 16 and 17 explicitly identify Mr. Chavez as one of the individuals who was

present at and involved in the August 3 controlled buy. Specifically, in the legends at the

top of the two transcripts, Mr. Chavez’s name is expressly listed as one of the speakers on

the recording. See id., Vol. IV, Ex. 16 at 1 (identifying “Randolfo Librado Chavez” as

one of the speakers on the recording); id., Ex. 17 at 1 (same). In effect, the transcripts

56 treat Mr. Chavez’s participation in the drug transactions as an established (even

stipulated) fact; this is particularly significant when we consider that not a single one of

the testifying agents who surveilled the scene on August 3 was able to identify Mr.

Chavez as a participant in the drug transactions, much less the orchestrator of those

transactions (as discussed infra). Stated otherwise, viewed from a perspective of

harmfulness, the evidence reasonably could be deemed significant because it squarely

undercut a central leg of Mr. Chavez’s defense, which relied on pointing out the lack of

government proof of his participation in the controlled drug buys. Id., Vol. III, at 45–46

(highlighting in his opening statement the purported paucity of government evidence

linking him to the events of July 22 and August 3). Thus, from the outset, the transcripts

expressly imbued the jury (at a minimum) with the understanding—without sightings

from the agents to support it—that Mr. Chavez participated, in some capacity, in at least

the August 3 transaction and—because of the common nexus of Mr. Salas—possibly the

July 22 one as well.

Adding further color to the jury’s understanding of Mr. Chavez’s supposed

involvement in the drug-trafficking events of July 22 and August 3, the transcripts purport

to reflect recordings of “controlled buys” involving methamphetamine. Take Exhibit 15,

for example. It purports to capture a conversation pertaining to the July 22 controlled

purchase. See id., Vol. IV, Ex. 15 at 1 (entitled “Transcript of audio 7-22 Controlled

Buy”). The transcript’s only explicit reference to a methamphetamine transaction,

57 however, is the introductory, internally made statement by an unidentified law-

enforcement officer that the transcript relates to “a control[led] purchase of

methamphetamine.” Id. (Statement 1). In a similar vein, Exhibit 17 begins with the

following internal statement by an unidentified officer: “[t]his will be a recorded call to

Randy . . . . to arrang[e] the purchase of approximately 4 ounces of methamphetamine.”

Id., Ex. 17 at 1 (Statement 1). This statement unambiguously links Mr. Chavez to the sale

of methamphetamine, despite the absence of any reference to “methamphetamine”

whatsoever within the transcribed conversations themselves. In other words, it is notable

that not once does the word “meth” or “methamphetamine” appear within the text of the

transcribed statements of the alleged participants in the July 22 and August 3 transactions,

including Mr. Chavez. Yet, what effectively amounts to editorial commentary in these

transcripts—commentary made by unidentified speakers not subject to cross-

examination—directly implicates Mr. Chavez in the sale of methamphetamine. Cf.

United States v. Grinage, 390 F.3d 746, 751–52 (2d Cir. 2004) (upon review of

defendant’s conviction for possession with intent to distribute the drug PCP, concluding

that improper admission of case agent’s interpretations of calls involving defendant was

not harmless where “[t]here was no reference to PCP or drugs on the calls [themselves]”).

Further, the transcribed conversations themselves constitute strong—albeit, again,

problematic and potentially misleading—evidence of Mr. Chavez’s supposed leading role

in effectuating the controlled purchases. Exhibit 16, for example, attributes various

58 statements to Mr. Chavez, including “I can get pills,” R., Vol. IV, Ex. 16 at 8 (Statement

44), and “I’m like the boss. I’m your boss. I can get you anything you want,” id. at 10

(Statement 51). And Exhibit 17 reflects a back-and-forth allegedly between Mr. Chavez

and Mr. Salas about whether it is “[t]hree thousand dollars . . . for the four,” with Mr.

Chavez stating that “[i]t was thirty-six, at least.” Id., Ex. 17 at 2 (Statements 17, 20).

When those statements are then viewed through the lens of the unidentified officer’s

introductory commentary, they strongly imply that their subject matter was the price of

methamphetamine. Put simply, these statements deeply implicated Mr. Chavez in the

very methamphetamine-related offenses he was charged with, despite the lack of any

explicit indication within the operative text of the transcripts themselves that the

conversations concerned methamphetamine at all, and despite Mr. Chavez’s insistence

that the government lacked proof that he participated in the controlled buys, let alone

coordinated them.

Thus, the transcribed conversations inexorably pointed toward Mr. Chavez’s

culpability—indeed, his leading role—in these two transactions and consequently had the

power to prejudice him. See Holton, 116 F.3d at 1540 (noting risk of prejudice “when a

transcript attributes incriminating statements to a defendant that the defendant does not

admit making”); cf. United States v. Berry, 64 F.3d 305, 309 (7th Cir. 1995) (Rovner, J.,

concurring) (“[I]t is difficult to imagine evidence that would be more prejudicial to the

defendant than a transcription of a conversation in which he allegedly sold crack cocaine

59 to a government informant, where the transcription identifies him as the speaker and the

sole issue in the case is who was speaking during the transaction.”).

It cannot be gainsaid, moreover, that the transcripts played a significant role during

the trial. Indeed, at essentially every turn of the proceedings, the transcripts figured into

the government’s case. In its opening statement alone, the government made four explicit

references to “the wire” that was used to generate the audio recordings—the contents of

which the government contended were reflected in the transcripts. See R., Vol. III, at

40–43. And throughout the government’s presentation of its case, references to—as well

as witness examinations regarding—the transcripts were constant and extensive. For

instance, the government engaged in a lengthy examination of Special Agent Ruby

concerning the transcripts, having him read substantial portions of Exhibits 16 and 17 into

the record for the jurors. See, e.g., id. at 242–44, 258–61.

Finally, in closing statements, the government not only placed the transcripts front

and center at multiple points, but in so doing underscored some of the most incriminating

statements in them: it highlighted, for instance, that Mr. Chavez (supposedly) had said, “I

am like the boss.” Id. at 344 (referring to id., Vol. IV, Ex. 16 at 10). Even in its

concluding line at trial the government again referenced the transcripts, stating that Mr.

Chavez “was the person who said he was the boss on the wire, and the boss is the one that

did this. We ask that you find him guilty.” Id. at 361; see Hands, 184 F.3d at 1332

60 (“When assessing the effect of the evidentiary error upon the case as a whole, we also

must consider the prosecutor’s closing argument.”).

Based on the foregoing, it seems clear to us that the government’s reliance on the

transcripts at trial was repeated and extensive, which militates in favor of a conclusion

that the (improperly admitted) transcripts may have had a substantial influence over the

jury’s guilty verdict. See, e.g., United States v. Taylor, 210 F.3d 311, 316 (5th Cir. 2000)

(holding that defendant’s “substantial rights were affected by the presence of and

repeated references to the government’s [erroneously admitted evidence]”); United States

v. DeSantis, 134 F.3d 760, 770 (6th Cir. 1998) (holding that error was not harmless “in

light of[, inter alia,] the repeated references the prosecution made” to improper evidence

that was materially suggestive of the defendant’s guilt).

That the transcripts occupied a central role at trial is especially problematic

because defects and omissions inhering in them make their integrity and soundness

questionable. As an initial matter, the transcripts offer no indication of who created them,

when or how the respective transcribers did so, how much time elapsed between the

making of each successive statement within the three transcripts, and whether these

exhibits reflect a complete transcription of the underlying recordings, to name a few

problematic aspects of them. Indeed, at oral argument, the government admitted that

Trooper Scimone did not prepare the transcripts and that it did not call as witnesses the

actual transcribers for logistical reasons. See Oral Arg. at 24:00–24:35; see also supra

61 note 6. Put simply, there was no information before the district court and the jury

regarding the identities of the transcribers, nor the methodology that they employed. And

we have no such information before us now.

Furthermore, as to Exhibit 15 specifically, it is likely not a complete transcript of

the recording of the July 22 controlled purchase: one agent testified that he and his fellow

law enforcement agent could hear “other people talking” to Mr. Salas, R., Vol. III, at 106,

but the transcript is silent as to what those other people may have said, see id., Vol. IV,

Ex. 15 at 2 (Statements 5–7), to say nothing of who the speakers may have been.

Exhibit 16 is beset by several such problematic issues. For one, numerous words

or phrases labeled “unintelligible” in the Spanish column are not deemed as such in the

English-translation column; i.e., they are effectively deemed intelligible in the English

column. Take Statement 36, for example:

Id., Ex. 16 at 7 (Statement 36). With respect to this statement attributed to Mr. Salas, the

term “unintelligible”19 appears twice in column three (i.e., Mr. Salas’s statement as it was

19 See supra note 4, concerning the occasional presence of the word “ininteligible” in the Spanish-language text of the transcript.

62 purportedly originally heard on the recording), but it does not appear anywhere in column

four (i.e., the purported English translation of Mr. Salas’s statement found in column

three). This discrepancy raises a question: how are these words or phrases that are

denoted “unintelligible” in Spanish somehow deemed to be intelligible in the English

translation? It is unclear. And this discrepancy is only the tip of the iceberg: while thirty-

nine words or phrases are designated as “unintelligible” in column three of this transcript,

see id. at 1–11 (Statements 2, 4–6, 8–10, 12, 14, 16–18, 25–28, 30, 34–38, 40–41, 43–45,

47, 50–51, 53), fourteen of those words or phrases are not designated as such in column

four, see id. at 4, 6–10 (Statements 18, 34–38, 43–44, 47, 50–51).

Exhibit 16 is saddled by another discrepancy. At least eighteen statements, as they

appear in columns three and four of the exhibit, respectively, appear to reflect conflicting

accounts of what was said on the recording. Id. at 1, 3–10 (Statements 6, 12, 18, 25, 27,

29, 31, 34–36, 38, 44–47, 50–52). Statement 52 exemplifies this issue:

Id. at 10 (Statement 52). The Spanish text in column three consists of six words in

Spanish, followed by an ellipsis, and attributed to a single speaker (i.e., Mr. Chavez). The

English-translation text in column four, by contrast, consists of a full fourteen words,

63 comprising a fully punctuated sentence—viz., without an ellipsis—that is attributed to one

speaker (i.e., Mr. Chavez), as well as a question attributed to a second speaker (i.e., Mr.

Salas). We are hard pressed to discern how a purported translation could generate such a

sharp disparity between the foreign-language and English-language versions.

In a similar vein, Exhibit 17 contains perplexing discrepancies such as the one

below:

Id., Ex. 17 at 2 (Statement 13). The text in column three reflects an “unintelligible”

portion of the recording, while the text in column four bears no indication of

unintelligibility. How is this so? Again, we are left with nary an explanation—exposing

further problematic cracks in what is essentially a critical part of the government’s case.

In short, the transcripts went to the heart of the government’s case: not only did

they purport to show Mr. Chavez’s involvement in the July 22 and August 3 controlled

purchases, but they also put in high relief Mr. Chavez’s supposedly leading role in the

exchange of methamphetamine for cash. They did so, however, in a tenuous and

problematic manner—their contents repeatedly raising serious questions regarding

whether they truly reflected the contents of the recordings in the first place, viz., serious

questions concerning the integrity of the transcripts.

64 Accordingly, given the central role that the transcripts played in the government’s

case, and their inherently problematic and potentially misleading nature, our first inquiry

strongly inclines us to at least have grave doubt about whether the transcripts had a

substantial influence on the outcome of the trial; if we ultimately have such doubt, the

district court’s admission of Exhibits 15, 16, and 17 would be non-harmless. United

States v. Medina-Copete, 757 F.3d 1092, 1109 (10th Cir. 2014) (“The harmless-error

doctrine does not require us to precisely measure the damage caused by [the

impermissible] testimony. It is sufficient for us to say that we have ‘grave doubt’ about

the outcome of this trial . . . .”).

As part of our harmless-error analysis, in addition to assessing the nature of the

erroneously admitted evidence itself, we also consider “the overall strength of the other

evidence against the defendant.” Blechman, 657 F.3d at 1069. The salient inquiry here is

this: removing the erroneously admitted evidence from the equation, how strong—or

not—was the government’s case at trial? In that regard, we have held that where the

other (properly admitted) evidence that the government adduced was “overwhelming,”

the error stemming from the improperly admitted evidence was harmless. United States

v. Perez, 989 F.2d 1574, 1584 (10th Cir.1993) (en banc); see also United States v. Turner,

285 F.3d 909, 914–15 (10th Cir. 2002) (holding that other evidence properly admitted at

65 trial was “sufficiently strong” to permit the conclusion that the erroneous admission of

unreliable testimony was harmless).

The other evidence that the government adduced at Mr. Chavez’s trial, however,

was far from overwhelming. Indeed, putting aside the inadmissible transcripts, the

government’s evidence establishing that Mr. Chavez participated in the two controlled

buys, much less orchestrated them, was not strong.

More specifically, in evaluating the strength of the government’s case, the

evidence that it did not produce is telling. For instance, the government did not produce

any photographic evidence of Mr. Chavez participating in the controlled buys. See R.,

Vol. III, at 299. Nor did the government proffer any forensic evidence linking Mr.

Chavez to the methamphetamine that Mr. Salas allegedly procured from him. Id. at

281–82 (“Q. Was there any effort at least to attempt to try and lift fingerprints off of those

[sic] packaging? A. I do not recall. . . . Q. Most certainly that is not in the evidence, is it?

Fingerprints lifted from the packaging is not [in] evidence; is that right? A. No, sir.”).

Nor did the government present evidence corroborating that the individual with whom

Mr. Salas spoke by telephone in advance of the August 3 controlled purchase was—as the

government represented at trial—Mr. Chavez. Id. at 279–80 (upon questioning about the

identity of the individual on the other end of Mr. Salas’s calls, a law-enforcement agent

explaining that he (i.e., the agent) “did not identify the defendant,” given that the

66 communications were placed from “prepaid” phones that are “commonly used by people

who don’t want to have phone numbers or phones associated with them”).

Finally, the government did not adduce a shred of law-enforcement testimony

directly linking Mr. Chavez to the controlled buys. More specifically, none of the law

enforcement agents were able to observe whether drugs even changed hands during the

controlled purchases, nor were they able to hear on the recordings any express mention of

drugs. See, e.g., id. at 106 (Agent Lile’s testimony) (testifying that “[m]yself and [another

agent] were not obviously parked . . . right on top of where the buy went down, so we

could not see everything that happened”); id. at 112 (“Q. From what you observed what

you actually saw, you didn’t see the actual transaction take place; is that correct? A. It

is.”); id. at 123 (Agent McDonald’s testimony) (“Q. Were you able to see any particular

detail of what happened between the two cars? A. I wasn’t able to see a lot of detail. I

could see the two cars meet, but I didn’t have eyes on the informant. That wasn’t my

role.”).

Indeed, of the law enforcement officers examined at trial—whom, the government

assured the jury, had eyes on the scene the entire time—not a single one among them was

able to confirm that Mr. Chavez was present at the scene of the two controlled buys on

July 22 and August 3, much less participated in them. Indeed, in one witness examination

after the other, the agents acknowledged that they were unable to make out exactly who

was in the white car from which Mr. Salas had purportedly procured the

67 methamphetamine. See, e.g., id. at 106 (Agent Lile’s testimony) (explaining that in light

of where he was parked at the scene, what he “definitely could understand” was that “Mr.

Salas had met with another party, because there were other people talking”); id. at 133

(Agent Preston’s testimony) (“Q. [W]ere you able to observe [Mr. Salas] while he was [in

the Old Chicago parking lot]? A. No. . . . Q. [Y]ou weren’t able to see his car or any

other vehicles at Old Chicago? A. I -- I can’t -- no, I do not believe that we had eyes on

him the whole time. Q. Or at least you did not? A. I did not.”); id. at 188 (Officer Vos’s

testimony) (“Q. To this day, you don’t know . . . who was in that [white] vehicle from

observing the vehicle -- the people at the time, on July 22nd, 2016, you didn’t know who

was in that vehicle, correct? A. Correct.”).

To be sure, an officer did positively identify Mr. Chavez as one of the individuals

in the white car when he stopped it shortly after the August 3 controlled purchase. See id.

at 305. However, the “mere presence at the scene of a drug transaction, without more, is

insufficient to support a conviction for aiding and abetting.” United States v. Williamson,

53 F.3d 1500, 1515 (10th Cir. 1995); accord United States v. Ibarra-Diaz, 805 F.3d 908,

932 (10th Cir. 2015); cf. United States v. Anderson, 189 F.3d 1201, 1207 n.3 (10th Cir.

1999) (distinguishing “those cases [where] we have held that a defendant’s mere presence

at the scene of a crime, even with knowledge that the crime is being committed, is

insufficient to support an aiding and abetting conviction”). Therefore, it logically follows

perforce that such presence would hardly constitute strong evidence that a defendant was

68 the principal-cum-boss of the drug transaction. Cf. United States v. Verners, 53 F.3d 291,

294–95 (10th Cir. 1995) (“The case against [the defendant] for possession with intent to

distribute is far weaker [than that against her co-defendant] and presents a close question.

Although there is ample evidence that she had knowledge of the presence of the drugs in

the kitchen and sufficient evidence to show constructive possession of these drugs, there

is little, if any, evidence of her intent to distribute. The government did not, for example,

offer any evidence that she actively participated in either the manufacture or distribution

of the drugs.”). Therefore, though the officer’s identification of Mr. Chavez in the white

vehicle shortly after the August 3 controlled purchase was “certainly probative evidence”

bearing on Mr. Chavez’s guilt, it did not add much to the strength of the government’s

case. Williamson, 53 F.3d at 1516.

At bottom, the only evidence the government adduced that affirmatively linked Mr.

Chavez to the controlled purchases was the testimony of Mr. Miranda and Mr. Salas. But

their credibility was open to question. Hands, 184 F.3d at 1330 n.23 (explaining that

“[h]armless error review, unlike a determination of the sufficiency of the evidence, does

not require us to view witnesses’ credibility in the light most favorable to the

government” (citing United States v. Marshall, 173 F.3d 1312, 1318 n.15 (11th Cir.

1999))). More specifically, Mr. Miranda, who had been indicted along with Mr. Chavez,

pleaded guilty to one count of distribution of methamphetamine and testified at trial in the

hopes of receiving a reduced sentence. See R., Vol. III, at 193–94 (Mr. Miranda attesting

69 to his awareness of the portion of his plea agreement providing that the government had

the discretion to recommend a downward departure from his otherwise applicable

sentencing range upon determining that he had fully cooperated with the government);

see also U.S.S.G. § 5K1.1 (providing for downward departures from the otherwise

applicable U.S. Sentencing Guidelines range for offenders who have provided

“substantial assistance” to the government); Hands, 184 F.3d at 1330–32 (in connection

with harmless-error analysis, finding that government’s other evidence adduced at trial

was insufficient in part because its main witnesses “[e]ach had pled guilty to narcotics

offenses, and each testified in the hope of receiving a reduced sentence or reduced

charges,” thereby calling into question their credibility).

Mr. Salas’s credibility was open to question for a similar reason: as he himself

averred at trial, he agreed to serve as a confidential informant for the express purpose of

having a misdemeanor theft charge removed from his record. See R., Vol. III, at 136–38.

Calling his credibility into further question was his unauthorized methamphetamine

acquisition from Mr. Miranda following the July 22 controlled purchase, as well as his

DUI arrest. See id. at 149, 236–37.

To be sure, it could be reasonably argued that the aforementioned evidence that the

government adduced was “sufficient to sustain the jury’s verdict . . . as a purely legal

matter.” United States v. Irvin, 682 F.3d 1254, 1264 (10th Cir. 2012). But that would be

answering the wrong question. See Richter, 796 F.3d at 1197. The controlling question

70 here is whether the improperly admitted transcripts had “a substantial influence on the

outcome of [the] trial or leaves one in grave doubt as to whether it had such effect.” Id.

And, in light of the foregoing analysis, we feel constrained—absent some mitigating

effect of the jury instructions, discussed infra—to answer this question in the affirmative

because we would harbor at least grave doubt concerning whether the transcripts had a

substantial influence on the outcome. See Medina-Copete, 757 F.3d at 1107, 1109

(concluding that the “properly admitted evidence” was sufficient to support the jury’s

guilty verdict, but concluding that the admission of the improper evidence was “not

harmless” because the court had “‘grave doubt’ about the outcome of this trial”).

We consider here whether the district court’s instructions to the jury may have

mitigated the substantial prejudicial effects of the improperly admitted transcripts.20 See

Richter, 796 F.3d at 1197; United States v. Sanders, 928 F.2d 940, 942 (10th Cir. 1991)

(recognizing that a limiting instruction has the potential to cure any prejudice from the

erroneous admission of evidence); see also Hands, 184 F.3d at 1329–35 (in assessing

20 To be clear, our analysis here regarding Jury Instruction 26 strictly concerns the issue of harmless error in connection with Mr. Chavez’s first claim of error under the best-evidence rule (i.e., the district court’s allegedly erroneous admission of the English-translation transcripts without the recordings themselves). This analysis is distinct from any analysis we might have conducted concerning Mr. Chavez’s second claim of error, which asserts instructional error as to Jury Instruction 26. As we have previously stated, in light of our resolution of Mr. Chavez’s first claim of error, we need not—and thus do not—reach his second claim.

71 whether an error is harmless, observing that “a limiting instruction regarding the

[erroneously admitted] evidence” may help to “diminish[] its prejudicial impact”). In this

regard, we determine that not only did the court’s jury instructions not mitigate the

prejudice arising from the improper admission of the transcripts, but they probably

exacerbated that prejudice—leaving us confident that the district court’s error must be

deemed non-harmless. We expound on these reasons below.

The salient instruction here is Jury Instruction 26; therein, the district court

instructed the jurors as follows:

While in a case involving English recorded conversations and transcripts, the jury is routinely instructed that they are not bound by the transcript, that is because every juror is just as capable as the person preparing the transcript to tell what is being said on the recording. This is not so with Spanish conversations that have been introduced in this case. The translated transcripts are the evidence you should rely on. You are not free to reject the translation contained in the transcripts of the tape recordings given by Officer Scimone. You are, however, free to give this evidence whatever weight or consideration you deem to be justified . . . .

R., Vol. I, at 112 (emphasis added). Far from neutralizing the prejudicial effect of the

improperly admitted transcripts, these instructions did precisely the opposite: the court

required the jury to rely on the transcripts as the primary evidence of the conversations

that the audio recordings purportedly captured and, notably, imposed none of the kinds of

limitations on the use of such evidence that courts have determined mitigates prejudice.

See United States v. Garza, 435 F.3d 73, 77 (1st Cir. 2006) (holding that error was

harmless where “[t]he court instructed the jury that the transcripts were not evidence to be

72 considered in deciding th[e] issue [of whether the defendant was the participant in the

drug deal that the transcripts described]”). Notably, the court did not instruct the jurors to

determine whether the transcripts accurately reflected the speakers’ identities—including,

as important here, the identity of Mr. Chavez. Rather, the court said nothing at all on the

matter, thereby leaving the jury free to accept the transcripts’ identification of Mr.

Chavez, as well as their attribution to him of statements that (in the context of the

transcripts as a whole) could hardly be anything other than significantly incriminating.

In short, the district court’s jury instructions did nothing to diminish the substantial

prejudicial effects of the improperly admitted transcripts; instead, the court’s instructions

probably exacerbated those effects. As such, we are constrained to conclude that the

district court’s error in admitting those transcripts is non-harmless and reversible.

* * *

To recapitulate, the best-evidence rule embodied in Rule 1002 applied to Exhibits

15, 16, and 17—that is, English-translation transcripts that purportedly reflected the

contents of audio recordings pertaining to alleged methamphetamine purchases on July 22

and August 3. The district court erred under the best-evidence rule by admitting those

transcripts without the audio recordings themselves. And, because we harbor at least

grave doubt concerning whether the transcripts had a substantial influence on the outcome

of Mr. Chavez’s trial, we deem that error non-harmless.

73 III

For the foregoing reasons, we REVERSE and REMAND to the district court with

instructions to VACATE Mr. Chavez’s convictions and grant him a new trial.

74 17-8096, United States v. Chavez

HARTZ, J., Circuit Judge, dissenting.

The sin of the trial judge was to use his common sense. He ruled that a tape

recording of a conversation in a language incomprehensible to the jury was inadmissible.

And he ruled that the jury could receive an English-language translation of that recording

when the accuracy of the translation was not disputed. If the Federal Rules of Evidence

prohibited what the judge did, they would be obtuse. Fortunately, they do not. The

advisory committee notes to the Rules, as well as academic commentary, support the

judge’s rulings. No federal appeals court has reversed a district court for admitting into

evidence a translation without also admitting the original-language document or

recording. And several federal appeals courts have endorsed exactly what the judge did

at Defendant’s trial.

Although the merits of the trial judge’s logic seems obvious, this dissent will need

to explore in some depth various rules-of-evidence matters that receive little attention

from the judiciary or commentators. After that discussion the dissent will turn to the

specifics of the case at hand.

But before explaining why the trial judge’s rulings were correct—in particular,

why they did not violate the Best Evidence Rule—I need to address some side issues that

should not distract from the central error of the majority opinion. First, to say that the

original recordings were properly excluded from evidence is not to say that they should

have been withheld from Defendant. Obviously, the defense should have access to the

originals so it can be sure that the translation is correct. But that is not an issue here. Defense counsel told the district court that he had no complaints about timely access to

the recordings and the translation.

Second, to say that the original recordings were properly excluded in this case is

not to say that such recordings should always be excluded. There are a variety of reasons

why a jury may benefit from hearing a recording even if the members do not understand

the language. The recording may be helpful in identifying the speaker, assessing the tone

of the conversation, etc. But Defendant did not adequately present such grounds to the

judge. And the translation is admissible regardless of whether the recording is admitted.

Third, the majority opinion is correct in raising questions about inconsistencies in

the transcripts in this case. For example, it is hard to see how six words in Spanish could

be translated into 14 words of English. But Defendant did not adequately alert the trial

judge to such flaws. Nor has he provided any basis for believing that the inconsistencies

are material. They may have nothing to do with the accuracy of the translation (as

opposed to errors in the transcription of what was said in Spanish). Moreover, defense

counsel said that he had no concerns about the accuracy of the translation, never

suggested an alternative translation, and to this day has not pointed to any error in the

translation that could have influenced the jury. Defendant’s position is that the

translation should not have been admitted without the original recordings also being

admitted. Yet he has failed to offer a single reason why the jury might have reached a

different verdict if the recordings had also been admitted. (He has not even included the

recordings in the record on appeal, which ordinarily would preclude him from arguing

that exclusion of the recordings was reversible error.) In closing argument the prosecutor

2 quoted a few lines from the translated transcripts; but Defendant has not asserted, much

less established, that those lines were mistranslations. When this case is retried, the

recordings will be admitted into evidence and the translation will be provided to the jury.

The only change in the evidence allowed will be that the jurors may be permitted to listen

to a recording in a language they do not understand. Neither the majority opinion nor

Defendant has explained why Defendant will be better off as a result. I need not address

harmless error, but the majority opinion is singularly unpersuasive on that issue.

Further, it should not be surprising that Defendant did not attack the translation in

district court. First, defense counsel conceded at trial that he was not aware of any

material error in the translation. And second, contrary to the presentation by the majority

opinion, the prosecution had a compelling case (which this dissent will summarize when

it turns to the specifics of the trial) that informant Bryan Salas had purchased drugs from

an occupant of specific cars on two occasions, and the recordings were largely irrelevant

to the only real issue at trial—whether Salas had lied under oath about whom he had dealt

with (that is, whom he was talking with during the recorded conversations) in the two

cars. Nitpicking the translation would have accomplished nothing.

Finally, I should respond to footnote 17 of the majority opinion, which was added

in response to my dissent. It eloquently explains why resolving this case under my

analysis would be contrary to all principles of proper adjudication. Unfortunately, the

footnote is also contrary to this circuit’s precedent. That precedent emphasizes our duty

to examine whether the lower-court decision can be affirmed. Twenty-five years ago we

responded as follows to a suggestion from an appellant that we apply to the appellee our

3 rule that an appellant’s contention is ordinarily considered waived if not raised in the

opening brief:

[The appellant’s] argument misconceives the roles played by the appellant, the appellee, and the court of appeals when a district court judgment is appealed. While the appellant challenges the district court’s ruling, the appellee is only interested in maintaining the status quo, i.e., an affirmance. Because the appellant comes to the court of appeals as the challenger, he bears the burden of demonstrating the alleged error and the precise relief sought. A court of appeals is not required to manufacture an appellant’s argument on appeal when it has failed in its burden to draw our attention to the error below. In the event of such a failure, the court will ordinarily consider the appellant’s point waived. Appellees bear no such burden. Though Fed.R.App.P. 28(b) requires the appellee’s brief to contain arguments addressing the issues raised by the appellant, we have never characterized the appellee’s obligation in terms of a categorical imperative. The distinction between appellant’s and appellees’ obligations under Rule 28 grows out of the court of appeals’ freedom to affirm a district court decision on any grounds for which there is a record sufficient to permit conclusions of law, even grounds not relied upon by the district court. This broad power to affirm extends beyond the counter-arguments raised by the appellee; it includes any ground for which there is record to support conclusions of law. Once the appellant alleges the district court erred, we have a duty to assess the validity of the appellant’s allegations. This duty arises in part out of our relationship with the district court, and we may not neglect it simply because an appellee fails to defend adequately the district court’s decision. To do so would open the door to a perverse jurisprudence by which properly decided district court decisions could be reversed.

Hernandez v. Starbuck, 69 F.3d 1089, 1093–94 (10th Cir. 1995) (footnote, citations,

brackets, and internal quotation marks omitted). More recently, now-Justice Gorsuch

expressed the point as follows:

[Our] reluctance to command do-overs in the district court is also why we treat arguments for affirming the district court differently than arguments for reversing it. We have long said that we may affirm on any basis supported by the record, even if it requires ruling on arguments not reached by the district court or even presented to us on appeal. See United States v. Davis, 339 F.3d 1223, 1227 (10th Cir.2003); Griess v. State of Colo., 841 F.2d 1042, 1047 (10th Cir.1988); see also S.E.C. v. Chenery

4 Corp., 318 U.S. 80, 88, 63 S.Ct. 454, 87 L.Ed. 626 (1943) (“[I]n reviewing the decision of a lower court, it must be affirmed if the result is correct although the lower court relied upon a wrong ground or gave a wrong reason.”) (internal quotation omitted). This preference for affirmance no doubt follows from the deference we owe to the district courts and the judgments they reach, many times only after years of involved and expensive proceedings. Because of the cost and risk involved anytime we upset a court’s reasoned judgment, we are ready to affirm whenever the record allows it. So it is that appellants must always shoulder a heavy burden—they must come ready both to show the district court's error and, when necessary, to explain why no other grounds can support affirmance of the district court's decision.

Richison v. Ernest Group, Inc., 634 F.3d 1123, 1130 (10th Cir. 2011). Of course, before

affirming on an unargued ground, it is often appropriate to request supplemental briefing

from the parties. Such a request, however, is not a prerogative of a dissenter.1

In this case, there is an additional institutional reason for conducting the analysis

addressed in this dissent. The analysis can provide essential guidance for courts in the

future. I have expressed my view that the district court acted in accordance with common

sense. Indeed, to my knowledge this is the first appellate decision reversing a district

court for admitting a translation of a foreign-language recording while excluding the

1 The majority opinion’s other reasons for not addressing the dissent’s arguments are unsound. There are no relevant factual disputes. There was nothing for the district court to weigh under Rule 403 because Defendant did not suggest how it would be helpful for the jury to hear the tape recordings. As for treating the translation as expert opinion and Lt. Scimone as a qualified expert, Defendant did not challenge the accuracy of the translation and Scimone’s testimony (described later) undoubtedly qualified him as an expert. The majority opinion’s footnote will unfortunately encourage lawyers to waste everyone’s time with unnecessary requests for rulings on undisputed matters. As for notice to Defendant that Scimone would be giving expert testimony, the government provided notice of and access to the recordings and the translations in ample time to prepare for trial. Indeed, as a result of that notice and access, defense counsel stated that he had no concerns about the accuracy of the translation. 5 recording itself. Several circuit courts have affirmed the practice; the government’s brief

cites two such circuit opinions. But apparently no appellate opinions have explored how

the Rules of Evidence operate in this context. I hope that this dissent provides a solid

start for that exploration.

At least the majority opinion’s footnote 17 serves one useful function. It makes

clear that the opinion is not considering (and therefore not rejecting) the view of the

Rules of Evidence taken in this dissent. Therefore, trial courts in this circuit, as well as

panels of this court, are free to adopt that view if they find it persuasive. The only

caution I would suggest is that to avoid reversal, trial courts explicitly cite the applicable

Rules of Evidence (perhaps in the future it would be wise to cite Rule 802 when

excluding hearsay) and make a determination of the expert qualifications of the

translator, even when the parties stipulate that the translation is accurate.

I. ANALYSIS

A. The Nature of Testimony Translating Documents or Recordings A proper analysis of this case must begin with an examination of the nature of

evidence translating material that originated before trial. A witness who testifies to the

English meaning of something expressed in a foreign language is undoubtedly an expert.

Not just anyone is permitted to provide such testimony. The translator must be qualified

“by knowledge, skill, experience, training, or education.” Fed. R. Evid. 702. The

witness is expressing an opinion about the English meaning of what was said in a foreign

language. What the witness testifies to is an opinion even though there may be broad

agreement about the meaning by those fluent in both English and the foreign language.

6 Such agreement is not that unusual with expert opinion and does not mean that the

opinion is less than expert. If a suspect leaves a good DNA sample at the scene of a

crime, there will be little disagreement among experts about whether the sample matches

the DNA of the defendant. There will be plenty of x-rays that every competent

radiologist interprets as showing a particular type of fracture.

Experts are called as witnesses because they are capable of drawing inferences

from information that the lay people on a jury are incapable of properly analyzing. If the

jurors were capable of properly analyzing the information, then the expert testimony

would be unnecessary and inadmissible, because it would not “help the trier of fact to

understand the evidence or to determine a fact in issue.” Fed. R. Evid. 702(a). If jurors

were required to draw relevant inferences from the information on which the expert relies

for his or her opinion, they might be left at sea and likely to reach unsubstantiated or even

irrational conclusions. That would certainly be the case if the jurors were left on their

own to determine the meaning of a document or conversation in a language they did not

understand.

The information on which an expert relies may be admissible evidence that is

presented to the jury, but that is not required. “If experts in the particular field would

reasonably rely on those kinds of facts or data in forming an opinion on the subject, they

need not be admissible for the opinion to be admitted.” Fed. R. Evid. 703. Indeed, the

evidence may be inadmissible precisely because jurors are incapable of evaluating it,

would have to speculate about its importance, and could draw highly inappropriate

inferences. In that circumstance the trial court could properly say that the information

7 upon which the expert relies is not relevant under Federal Rule of Evidence 402 or is

inadmissible under Federal Rule of Evidence 403 because its “probative value is

substantially outweighed by a danger of . . . unfair prejudice, confusing the issues, [or]

misleading the jury.” Yet it would be eminently reasonable for the expert to be permitted

to rely on the identical information.

The interplay of Rules 402 and 403 with Rule 703 has particular impact when a

statement or conversation in a foreign language is relevant to issues being litigated. The

trial court may reasonably decide that Rule 402 or 403 requires exclusion of a foreign-

language document or recording. If the issue is the meaning of the foreign words, the

document or recording will not be helpful to an English-speaking jury. On the contrary,

the jury may be misled or confused if it tries on its own to translate the words or sounds.

But under Rule 703 the inadmissibility of the document or recording does not foreclose

the admissibility of a translation of it into English by a qualified expert.

What about the Best Evidence Rule? On its face it seems to require that the

original in the foreign language be admitted if the translation is to be presented to the

jury. The core of the Rule is: “An original writing, recording, or photograph is required

in order to prove its content . . . .” Fed. R. Evid. 1002. I will assume that a translation of

a document or recording “prove[s] its content.”2 So it might seem that before a party can

prove the content of the original document or recording by providing a translation, the

2 But cf. Romo v. State, 941 N.E.2d 504, 508 (Ind. 2011) (admission of transcripts of translations of foreign-language recordings is not governed by Indiana’s Best Evidence Rule, Ind. R. Evid. 1002, since playing the recordings “would not serve the purpose of the rule because it could not prove any content to the jury.”). 8 Rule requires that the original be admitted. That presents quite a conundrum. Why

should the Rules of Evidence require the admission of an incomprehensible original when

the translation is what the jury must rely on? The answer is that the Rules have no such

requirement.

Rule 703 solves the conundrum by, as previously noted, permitting admission of

the expert translation without requiring admission, or even admissibility, of the document

or recording on which the translation is based. And it does so without violating the Best

Evidence Rule, because the Rule recognizes exceptions. The full text of Rule 1002 is:

“An original writing, recording, or photograph is required in order to prove its content

unless these rules or a federal statute provides otherwise.” Fed. R. Evid. 1002 (emphasis

added). And Rule 703 provides otherwise. In fact, the advisory committee note to Rule

1002 recognizes that Rule 703 limits the scope of the Rule. It states: “It should be noted

. . . that Rule 703 . . . allows an expert to give an opinion based on matters not in

evidence, and the present rule [Rule 1002] must be read as being limited accordingly in

its application.” To say that the Rule is limited in its application by Rule 703 is to say

that Rule 703 overrides it; that is, Rule 1002 does not apply when the writing, recording,

or photograph serves only as the basis for expert testimony. The expert can base an

opinion on, for example, an original writing even if the writing is not admitted. As

observed by the author of a leading treatise on evidence, “For all purposes, Rule 703

creates an exception to the original writing rule, Rule 1002.” Michael H. Graham, Expert

Witness Testimony and the Federal Rules of Evidence: Insuring Adequate Assurance of

Trustworthiness, 1986 U. Ill. L. Rev. 43, 66 (1986).

9 In short, and perhaps unsurprisingly, the treatment of foreign texts and recordings

by the Rules of Evidence comports with common sense. Presenting the original foreign-

language document or recording to the jury may only (1) waste the time of those jurors

who rationally decide to ignore what is provided or (2) create a substantial risk of

confusing or misleading those jurors who feel a responsibility to try to extract

information from that evidence. The trial court should be allowed to exercise its

discretion in determining whether the original should be presented to the jury.

Sometimes it makes sense to admit the original; sometimes it does not. But the decision

on whether to admit the original should not affect the admissibility of the translation.

The translation by a properly qualified expert will be helpful to the jurors regardless of

whether there is anything about the original document or recording that could be of use to

them. The Best Evidence Rule “should never be applied so rigidly as to interfere with

good trial sense.” 6 Jack B. Weinstein & Margaret A. Berger, Weinstein’s Federal

Evidence, § 1001.02 (Mark S. Brodin, ed., Matthew Bender 2d ed. 2018) (“Weinstein &

Berger”).

An error that infects the majority opinion’s analysis is that it treats a translation as

not being true evidence but only an aid to the jury in assessing evidence (the foreign-

language recording). But that cannot be. Consider a defendant being prosecuted for

fraud based on false statements in a document written in a foreign language. If the

translation of the document is not evidence, then the jury verdict cannot be based on it.

The jury would have to base its verdict on the foreign-language document that no juror

could understand. How is that possible? How could the jury know that the defendant

10 uttered a falsehood when it does not know the meaning of what the defendant said?

Certainly, if no translation were provided during the trial, the appellate court could not

affirm a conviction predicated on the meaning of a foreign-language document. The

same would be true if a foreign-language oral conversation was essential to the verdict.

Say, the only evidence that the defendant participated in a conspiracy consists of tape

recordings of the defendant’s conversations, all of which were in a language not

understandable by the jury. How is the jury to find the defendant guilty without

considering the meaning of the recorded conversations? It would not be enough to prove

that the defendant (whose voice may be recognizable on the recording) had a

conversation with a member of the conspiracy (whose voice is also recognizable). The

translation is not an aid to understanding evidence, it is the evidence on which the jury’s

verdict must rest. In short, the translation of a document or conversation is undoubtedly

evidence.

In a related error, the majority opinion suggests that a court can admit into

evidence the original recording of a foreign-language conversation but refuse to allow the

jury to listen to it. Again, how can that be? What in the world does it mean, then, to

admit something into evidence? Surely it has something to do with consideration by the

jury. But if the jury is barred from listening to the recording, how can it consider that

recording (as opposed to considering a translation or transcript) in reaching its verdict?

The majority opinion is forced into its absurd position by a combination of its erroneous

view that the Best Evidence Rule requires that the recording be admitted and its

recognition that the trial court can properly decide that allowing the jury to listen to the

11 recording could create confusion and misunderstanding. The majority may have a proper

concern that a translation not be admitted into evidence unless the opposing party has the

opportunity to examine the original foreign-language recording. But the trial court can,

and should, require that such an opportunity be afforded without having to admit the

original recording into evidence.

One source of the majority opinion’s error is its alluring, but misguided, treatment

of a translation of a foreign-language recording as being essentially the same as a

transcript of an English-language recording. But ordinarily there is no expertise—no

need for expert opinion—in preparing a transcript of a conversation in the language that

all jurors understand. Rule 703 does not apply. Having a transcript in hand, a juror may

find it easier to follow an English-language conversation being played from a recording.

But the court will generally instruct the jurors that they must rely on what was heard

rather than what was printed out for them. The court does not want the jury to be unduly

influenced by the transcript. In other words, the transcript is not admissible into evidence

but may be a demonstrative exhibit that can assist jurors in processing the evidence

presented to them.

A translation from a foreign language is a different animal altogether. The jurors

must rely on the expertise of the translator and could not possibly rely on their individual

understandings of a recording in a language they do not know. If the translation is not

admissible in evidence, then the recording may have no probative value whatsoever. As

the Ninth Circuit has explained: “When tapes are in English, they normally constitute the

actual evidence and transcripts are used only as aids to understanding the tapes; the jury

12 is instructed that if the tape and transcript vary, the tape is controlling. When the tape is

in a foreign language, however, such an instruction is not only nonsensical, it has the

potential for harm where the jury includes bilingual jurors.” United States v. Franco, 136

F.3d 622, 626 (9th Cir. 1998) (emphasis added) (citation and internal quotation marks

omitted); see United States v. Morales-Madera, 352 F.3d 1, 9 (1st Cir. 2003) (“[A]n

instruction that the jury should consider only what is on the tape and not what is in the

[translated] English transcript would not be appropriate.”); United States v. Valencia,

957 F.2d 1189, 1195 & n.16 (5th Cir. 1992) (similar), overruling on other grounds

recognized in United States v. Keith, 230 F.3d 784, 786 (5th Cir. 2000).

A number of scholars and commentators (cited below) have distinguished between

transcripts of English-language recordings and transcripts of foreign-language recordings

in just this way. Transcripts of English recordings are presented to the jury not as

evidence, but as demonstrative exhibits. Translations of foreign-language recordings, in

contrast, are admissible as evidence, and the recordings themselves may well be

inadmissible.

A leading treatise, after explaining why some courts treat transcripts of English

recordings as only inadmissible demonstrative exhibits rather than as evidence, then

discusses the contrary view and endorses it only for transcripts of recordings in a foreign

language:

Other courts dispute the notion that a transcript is merely demonstrative evidence that illustrates a recording. Under their approach, a transcript is admitted under the theory it is opinion testimony from the transcriber as to the meaning of those parts of the recording that are difficult to understand. Most commonly this theory is employed to justify the admission of a

13 transcript that translates a recording of a conversation in a foreign language. This theory deals with the best-evidence doctrine by reasoning that the transcript is not offered to prove the contents of the recording but, rather, to explain what inferences should be derived from it. Thus, courts adhering to this approach conclude that the transcript is no more objectionable under best evidence law than a physician’s testimony concerning the meaning of an x-ray or a handwriting expert’s testimony concerning the genuineness of handwriting. This approach makes sense where the transcript is a translation of a conversation in a foreign language since the author of the transcript clearly contributes expertise in the form of specialized knowledge concerning the language in question. However, where a transcript is prepared just to clarify difficult-to-understand portions of a recording in English, it is hard to justify viewing the transcript as expert opinion testimony because the author of the transcript may not bring any special expertise to this task . . . .

Charles Alan Wright and Victor James Gold, 31 Fed. Prac. & Proc. Evid. § 7167 at 360–

61 (emphasis added) (footnotes omitted).

A useful guide to the Federal Rules of Evidence states: (6) Transcription of recordings. Transcriptions of recorded English- language conversations are often used to help the jury better understand what was said. When, however, the recorded conversation is in a foreign language, it must be translated for the jury. In such cases, translated transcriptions of audio recordings may be admitted into evidence. Some courts both play the recorded foreign-language conversation and admit the translated transcript. Other courts admit the transcript without having the jury listen to the recording, reasoning that the jury will not profit from hearing a foreign-language conversation.

Steven Goode and Olin Guy Wellborn III, Courtroom Handbook on Federal Evidence,

FRE R 604 (2019 Westlaw) (emphasis added) (citations omitted).

Another treatise emphasizes that translated transcripts are independent evidence

based on expert testimony:

Translated transcripts. Where the audible record captures statements or conversations in a language other than English, a transcript in translation

14 is indispensable as a practical matter. . . . The problem of assuring accuracy is compounded, and careful pretrial work by the parties under judicial supervision is essential. Neither the court nor the jury is likely to be qualified to determine the accuracy of the translation by comparing it with the audible record, and both depend heavily on persons fluent in English and the other language. In this instance, the transcript (or transcripts, if competing versions must be offered because of the failure of the parties to agree) must be received as independent evidence, supported by the testimony of the translator, who must qualify as an expert, and if the parties cannot agree on translation issues, competing transcripts should be allowed.

Christopher B. Mueller and Laird C. Kirkpatrick, 5 Federal Evidence § 10:15 (4th ed.

2019, Westlaw) (emphasis added) (footnotes omitted).

And at least two law-review articles explain that the translated transcript is

preferable to original foreign-language recordings. First,

The courts disagree on whether the transcripts can be introduced as substantive evidence. Some courts believe transcripts should be introduced as substantive evidence, but only to aid to [sic] the jury in determining the real issue presented, the content, and the meaning of the tape recordings. . . . Generally tapes are virtually useless to the jury because they either cannot understand the tapes or must adhere to the transcript translation. The tapes may even confuse or mislead the jury, which provides a basis for excluding the tapes under Rule 403 of the Federal Rules of Evidence. Therefore, listening to the tapes is more likely to prejudice the case than to aid the jury in rendering a fair determination.

Beth Gottesman Lindie, Inadequate Interpreting Services in Courts and the Rules of

Admissibility of Testimony on Extrajudicial Interpretations, 48 U. Miami L. Rev. 399,

418–19 (1993) (emphasis added) (footnotes omitted). Second,

Some courts have treated the translations of audiotape recordings of conversations in a language other than English as aids for the jury and the non-English language audiotapes as the primary evidence. The translated transcript is considered a visual aid for the jurors while the audiotape recording is being played. This approach makes little sense when the audiotape recording is in a language other than English. Other courts have held that the purpose of the transcript is to assist the trier of fact in

15 understanding the foreign language spoken on the audio tape. Presumably, the judge and the jurors are not able to understand an audiotape in the non- English language. Furthermore, there is potential danger that bilingual jurors will disagree as to the meaning of the conversation. The better line of cases holds that the translations and not the audiotapes in the non- English language are the best evidence.

Charles M. Grabau & Llewellyn Joseph Gibbons, Protecting the Rights of Linguistic

Minorities: Challenges to Court Interpretation, 30 New Eng. L. Rev. 227, 252 (1996)

(emphasis added) (footnotes omitted).3

Several circuits have recognized the problems posed by recordings in a foreign

language and have consequently endorsed the exclusion of the originals of foreign-

language recordings while admitting translations. The Fifth Circuit, relying on Rule 403,

affirmed a district court’s refusal to play an original tape on the ground that doing so was

more likely to confuse the jury than to help it. See Valencia, 957 F.2d at 1194–96. The

Eighth Circuit likewise affirmed the exclusion of Spanish recordings on the ground that it

was unlikely that the jury could discern relevant information from the recordings. See

United States v. Grajales-Montoya, 117 F.3d 356, 367 (8th Cir. 1997); see also United

States v. Placensia, 352 F.3d 1157, 1165 (8th Cir. 2003) (“We have . . . held translated

3 There is no reason why a translation of a recording must be presented to the jury through a transcript. The translation could be read to the jury, perhaps by having different people read what was said by each voice on the recording. That was done in this case, with the reading being performed by the prosecutor and a law-enforcement officer. As an aid to the jury in following the presentation, a transcript of the translation was displayed on a screen. One advantage of this procedure is that it is unnecessary to provide the jurors with a copy of the transcript during their deliberations. Such a copy may overemphasize that particular evidence, which is one reason why judges do not provide jurors with transcripts of trial testimony (although they sometimes will have the court reporter read excerpts of testimony when deliberating jurors request it). 16 transcripts of foreign language tapes are admissible at the district court’s discretion, even

when introduced without the tape recordings.”). So did the Seventh Circuit. See United

States v. Estrada, 256 F.3d 466, 473 (7th Cir. 2001) (“[T]he district court saw no value in

allowing a presumably English speaking jury to hear tapes that were recorded in

Spanish.”). As the Massachusetts Supreme Judicial Court summarized the matter, “A

foreign language recording generally is not admissible in evidence unless there is relevant

evidence discernible on the recording other than the content of the statements made by

the participants in the conversation.” Commonwealth v. Portillo, 968 N.E.2d 395, 400

(Mass. 2012).

Two points should be emphasized about these decisions excluding foreign-

language recordings. First, these rulings in no way contradict the Best Evidence Rule.

That Rule does not mandate the admission of anything. It is a rule of exclusion not of

admissibility. It excludes copies from evidence; it does not ensure that every original

document or recording must be admitted. The fact that a document or recording is not a

copy does not mean that the jury must see it. If the original is irrelevant, who cares that it

is not a copy? In particular, the above-cited judicial and academic authority supports the

exclusion of foreign-language documents and recordings. Why introduce “evidence” that

will “sound like Greek” to the jury? If the only issue regarding a document or recording

in a foreign language is the meaning of what is said, it is hard to see why the untranslated

original should be presented to the jury. Cf. Romo, 941 N.E.2d at 508 (admission of

foreign-language translation transcripts is not governed by Indiana’s Best Evidence Rule,

17 Ind. R. Evid. 1002, since playing those recordings “would not serve the purpose of the

rule because it could not prove any content to the jury”).

Second, I am not suggesting that Rules 402 and 403 always precludes admission

of an original foreign-language recording. In some circumstances a recording may have

probative value even for a jury that does not understand the meaning of the words

spoken. For example, the recording may assist jurors in determining the identities of the

speakers. And some circuits have expressed the view that it is “relevant to play [a

foreign-language] tape to a jury that does not understand the substance of the

conversation to show the mood or tone of the speakers, or the general context or

ambiance of their conversation.” Valencia, 957 F.2d at 1195 (internal quotation marks

omitted) (although holding that the district court did not abuse its discretion in thinking

otherwise and excluding the recording); see United States v. Cruz, 765 F.2d 1020, 1024

(11th Cir. 1985) (explaining that playing a Spanish recording could allow the jury “to

detect changes in voice modulation and note any hesitancies or other characteristics

which might give meaning to the tape recording”). Some circuits, however, see no

potential relevance of a foreign-language recording. See Grajales-Montoya, 117 F.3d at

367 (the defendant “has suggested no reliable means of enabling people who do not speak

Spanish to interpret inflections and tone, and we cannot think of any, either”); Estrada,

256 F.3d at 473 (quoting Grajales-Montoya).

Thus, the best-evidence question is not whether the original must be admitted, but

whether the translation can be admitted into evidence even though the original has been

excluded. And that question was answered above. As I have already explained, the Best

18 Evidence Rule says that ordinarily a copy is not admissible, but that general rule is

qualified by the language “unless these rules or a federal statute provides otherwise.”

Fed. R. Evid. 1002. And there is such an exception for expert opinion. Under Rule 703

the matters on which an expert relies in forming an opinion “need not be admissible for

the opinion to be admitted,” if experts in the field would reasonably rely on them. Of

course, an expert translator would have to rely on the recording or document being

translated.

The appellate cases relied on by the majority as requiring admission of the original

foreign-language recording are not persuasive. Not one of the cases presented the issue

before us on this appeal—namely, whether the trial court can admit into evidence a

translation of a foreign-language recording when it has refused to admit the recording

itself. In fact, the original recording was admitted in each of the cases. Some include

dicta that it was necessary to admit the original, but none of those cases analyze the issue;

in particular, they fail to recognize any problem with presenting a foreign-language

recording to a jury that cannot understand what is being said, and they ignore the

possibility that the doctrine requiring the introduction of original English-language

documents or recordings may not apply to foreign-language material. Given these lapses,

the dicta are hardly persuasive.

I begin with the three Tenth Circuit cases relied on by the majority opinion:

United States v. Verdin-Garcia, 516 F.3d 884 (2008); United States v. Gomez, 67 F.3d

1515, (1995); and United States v. Rivera, 778 F.2d 591 (1985). All are readily

distinguishable from the present case because the district court had admitted into

19 evidence the original Spanish-language recording and the defendant had not challenged

admission of the recordings. The issue now before us therefore was not considered by

the court. And the only dicta in these cases that might support part of the majority

opinion’s analysis is too ambiguous to be in any way authoritative.

In Verdin-Garcia the argument by the defendant on appeal was that the

translations of the recordings were incorrect and should have been excluded. In the

paragraph of the opinion introducing this issue, the opinion merely stated an historical

fact: “The recordings themselves were admitted as substantive evidence; the translations

were shown to the jury for demonstrative purposes only.” 516 F.3d at 892. I discern

nothing in our opinion that approves or disapproves of what was done in that regard.

(Also, the translator herself testified. See id. I assume her testimony was admitted as

evidence.)

In Gomez a translator had testified “as a witness through whom the transcribed

conversations were admitted into evidence.” See 67 F.3d at 1525. We first rejected an

unpreserved argument that the translator was not qualified. See id. at 1525–26. We also

rejected a best-evidence argument “that the jury was never informed that the primary

evidence was the tape itself not the transcripts; and they were never informed that, should

they detect a discrepancy between the tape and the transcripts, the tape should control.”

Id. at 1526 (internal quotation marks omitted). Because the issue had not been preserved

below, the defendant could prevail only on plain-error review. See id. And his argument

failed under that standard because “[t]he admission of transcripts to assist the trier of fact

lies within the discretion of the trial court,” the defendant’s argument was “wholly

20 without merit,” and the “record clearly reflects that the district court did not abuse its

discretion in admitting the transcripts.” Id. at 1527. We added a footnote, however,

stating that “a cautionary instruction regarding the use of the transcripts as aids only in

understanding the audio tapes would have been preferred, see United States v. Robinson,

707 F.2d 872, 877 (6th Cir. 1983),” although we said that no reversible error resulted

from the absence of the instruction. Id. at 1527 n.15. Perhaps that dictum could be read

to state that testimony translating a foreign-language recording is not admissible

evidence. But I see no need to read it as making that improper statement. (It would be

improper because, as discussed more fully above, if the jury cannot use the translation as

evidence, how can it make any finding based on the meaning of the foreign-language

conversation?) The quoted statement refers to a transcript. The transcribed conversation

was in both English and Spanish, see id. at 1518 n.2, and the English portion of the

transcript was not admissible but merely a jury aid. (The Robinson decision cited in our

footnote involved a transcript of an English conversation, and the concern was

audibility.) Further, the translator testified, see id. at 1525, so the translation may well

have been presented orally, making the entire transcript only a jury aid. I think it would

be misguided to treat an ambiguous dictum, which was largely true, as authority for a

further proposition that is patently unreasonable.

As for the final case, Rivera, the district court had submitted to the jury during its

deliberations some Spanish-language recordings and transcripts of the recordings. See

778 F.2d at 600. The defendant complained that “submission of the tapes to the jurors

placed the one Spanish-speaking juror in the position of an expert on translation of the

21 almost entirely Spanish conversations in the tapes.” Id. We did not say that submission

of the tapes to the jurors was required. We said only that the jurors had the English

transcripts to examine, “the inflections and emphasis in the conversations may have been

significant,” and we would not presume prejudice but rather presume that the jurors were

true to their oaths and observed the court’s instructions. Id. In other words, there may

have been a relevant purpose for listening to the recordings, and we would presume that

there had not been misconduct by the jurors, so there was no reversible error. This is not

a declaration that failure to play the tapes would have been reversible error. To say that

doing X is not reversible error is slim support for the proposition that not doing X is

reversible error. Particularly on questions of admission of evidence, we defer to the trial

court, often being willing to affirm whichever way it decides.

Likewise, none of the out-of-circuit cases relied on in the majority opinion on this

point considered a challenge like the one raised by Defendant here. That is, none

considered whether the district court erred by admitting translations of foreign-language

recordings when the recordings themselves were not admitted into evidence. In each

case, the recordings had been admitted.

In two of the cases, there was not even a foreign-language recording; the court

discussed only how to handle recordings and transcripts of recordings in English. See

United States v. Holton, 116 F.3d 1536 (D.C. Cir. 1997); Government of Virgin Islands v.

Martinez, 847 F.2d 125, 128 (3d Cir. 1988). Some cases are apparently cited by the

majority opinion simply as examples where the foreign-language recordings, and

sometimes the English transcripts, were admitted into evidence (although in at least one

22 case the recordings were never played to the jury). See United States v. Abonce-Barrera,

257 F.3d 959, 962–64 (9th Cir. 2001) (Spanish tapes were played, English translations

were read to the jury, and the jury received transcripts of both. Defendant unsuccessfully

challenged the qualifications of the translator, the quality of the tapes, and the

opportunity to review the translations and transcription.); United States v. Franco, 136

F.3d 622, 625–29 (9th Cir. 1998) (Recordings were placed in evidence but not played for

the jury because they were in a foreign language. The jury was told it could request

recordings be played during deliberations, but no request was made. English transcripts

were also placed in evidence, but only 18 of 110 were read in full to jury. The appellate

court said that it is nonsensical to say that a transcript translation of a foreign-language

tape is only an aid to understanding.); United States v. Garcia, 854 F.2d 1280, 1282–83

(11th Cir. 1988) (While listening to the recording of a conversation that was in both

Spanish and English, the jury was provided an English-language transcript; the jurors

were told that they should determine the reliability of the transcript; the transcript was not

submitted to the jury during deliberations.).

Several other cases cited by the majority opinion state, or quote the trial judge as

stating, that a translated transcript is not admissible in evidence and is merely an aid for

the jury. But they do not say that testimony by a translator is inadmissible, so their

concern may be limited to use of the document; and in any event, they fail to explain how

the jurors are to discern the meaning of the foreign-language conversation if they cannot

treat a translation as evidence. See United States v. Nunez, 532 F.3d 645, 649–52 (7th

Cir. 2008) (The defendant objected that the English-language transcript of a Spanish

23 conversation denoted code words for drugs with quotation marks or footnoted definitions;

the appellate court said that the trial court had erred by telling the jury that it could afford

evidentiary weight to the transcript, but there was no reversible error because the

translators had testified to the meanings of the terms and the instruction to the jury did

not allow it “to consider evidence of the code words’ meaning that was not already in

evidence apart from the transcripts.”); United States v. Ben-Shimon, 249 F.3d 98, 100–01

(2d Cir. 2001) (rejecting challenge to use at trial of transcript of conversation in English

and Hebrew, where judge instructed that transcript was only an aid in listening to

recording); United States v. Martinez, 951 F.2d 887, 888 (8th Cir. 1991) (The defendant

challenged the admission of tape recordings because they were of such poor quality and

challenged the admission of what he termed inaccurate transcripts. The appellate court

says the admission of tapes is within the discretion of the trial court. The transcripts were

not admitted into evidence; the judge said that only the tapes, and not the transcripts,

were to be considered when weighing the evidence.).

The last case cited by the majority opinion (actually the first one it cites), United

States v. Morales-Madera, 352 F.3d 1 (1st Cir. 2003), is unique in that it concerned a trial

in Puerto Rico, where the jury was probably more fluent in the foreign language

(Spanish) than in English. See Andrea Freeman, Linguistic Colonialism: Law,

Independence, and Language Rights in Puerto Rico, 20 Temp. Pol. & Civ. Rts. L. Rev

170, 185 (2011) (2006 survey by United States Census Bureau reported that 95.2% of

Puerto Rico’s population speak Spanish at home). The Jones Act, 48 U.S.C. § 864,

however, requires that federal trials there be conducted in English. The central evidence

24 was 52 recordings of wiretapped conversations in Spanish. See 352 F.3d at 4. The

defendant’s chief complaints on appeal were that a translator did not translate the

recordings into English as they were played and, relevant here, English-language

transcripts were not marked as exhibits or admitted in evidence. See id. The court held

that it was error not to admit the English transcripts in evidence and that “an instruction

that the jury should consider only what is on the tape and not what is in the English

transcript would not be appropriate.” Id. at 9. But the failure to admit into evidence the

transcripts that were presented to the jury was not plain error requiring reversal when the

issue had not been raised below. See id. at 10–12. In the course of its discussion, the

court included the following sentence: “The best evidence rule requires that the tape

recording themselves must be furnished, absent agreement to the contrary, but does not

require that English translations of those tapes be excluded from evidence.” Id. at 9

(emphasis added). The majority opinion quotes the emphasized portion of the sentence to

support its basic thesis—namely, that the Best Evidence Rule requires the admission of

foreign-language recordings. But the focus of the First Circuit opinion was the second

clause of the sentence. In explaining why the transcripts needed to be admitted as

evidence, the court wanted to point out that the Best Evidence Rule “does not require that

English translations of those tapes be excluded from evidence.” The statement that the

rule requires admission of the recordings was dictum, since the recordings had been

admitted and no challenge was raised with respect to that admission. And the only case

authority cited by the court in support of its statement was an opinion that involved an

English-language recording in a Missouri trial. See United States v. Warner, 204 F.3d

25 799 (8th Cir. 2000). Moreover, I would think that in Puerto Rico the admission of an

original recording in Spanish would create little chance of confusion, given the high

likelihood that almost all the jurors could understand it. In a Puerto Rico trial a judge

might reasonably decide that a Spanish-language tape should be treated for best-evidence

purposes as an English-language recording would be treated in the other federal courts.

The First Circuit in Morales-Madera had no occasion to consider the problems that might

arise in presenting a Spanish-language recording to a jury that did not understand the

language. Not so in Wyoming.

To sum up, the admission into evidence of foreign-language recordings is

problematic because the jury is likely to misunderstand the content of anything said.

They should probably be admitted if the party seeking admission can explain how

listening to the recording could help the jury resolve a material issue (such as the identity

of the speakers). Otherwise, their admission should be within the discretion of the trial

judge. As far as I can tell, every circuit court to consider the matter has reached the same

conclusion. There appear to be no cases in which a federal circuit court has reversed a

district court that excluded a foreign-language recording from evidence on the ground

that it would be more likely to confuse than to enlighten the jury. The inadmissibility of

the foreign-language recording does not, however, foreclose the admission into evidence

of a translation of what was said on the recording. That is because the translation is

expert opinion and under Federal Rule of Evidence 702 such opinion can be admissible

even when the material on which it is based is not admissible. This feature of Rule 702

supersedes the Best Evidence Rule, Rule 1002. The expert opinion of the translator can

26 be offered to the jury as testimony or, by stipulation of the parties, through a transcript. If

the opinion is offered as testimony, a transcript may be provided to the jury as an aid in

following the testimony but ordinarily should not be admitted into evidence, just as a

transcript of an English-language recording should ordinarily not be admitted.

With this as background, I now turn to the particulars of Defendant’s trial.

B. Application to This Case

The prosecution’s case against Defendant was strong. The two drug transactions

for which he was convicted were both controlled buys by confidential informant Bryan

Salas. A number of government witnesses testified to their roles in “controlling” the

transactions by searching Salas before and after the transactions and keeping him under

continuous surveillance between the two searches. For the first transaction, (1) Salas and

his car were searched by law-enforcement officers for drugs or other contraband and he

was given funds to buy drugs, (2) officers then continuously observed him before, during,

and after he met the occupants of a white vehicle, and (3) Salas and his car were then

searched again and he delivered methamphetamine to the officers. For the second

transaction, Salas was (1) searched, (2) driven to a Walmart and given funds to buy

drugs, (3) continuously observed by officers stationed in the Walmart while Salas waited

inside for the transaction, (4) continuously observed by officers while he was outside the

Walmart, including when he met the occupants of a white vehicle (not the same vehicle

as on the first occasion), and (5) picked up at Walmart by officers who received

methamphetamine from him and searched him. The white car was then continuously

observed by law-enforcement officers until a local police officer stopped the car shortly

27 after the transaction. At trial the officer identified Defendant as the passenger in the car

and testified that the passenger identified himself by stating Defendant’s name.

Salas testified regarding each transaction and identified Defendant as the person

he dealt with in the car who took his money and gave him the drugs. He identified the

other people in the first car as Carlos Dominguez and McKleen Miranda. Miranda

pleaded guilty to the methamphetamine sale and testified at trial, stating that he was the

driver of the car and identifying Defendant as the passenger who took the money from

Salas and gave him the drugs. Salas identified the people in the second white car as

Defendant and Dominguez, and again testified that the transaction was conducted with

Defendant.

For each transaction, Salas wore a recording device that transmitted to the officers.

There are three tape recordings. Because their content is not important to the analysis, I

leave an abbreviated description of their content to a footnote.4

4 I will identify each recording by the same number as the exhibit that is the translated transcript of that recording. Tape 15 begins shortly before the first transaction. None of it purports to be a recording during the transaction. The only voices recorded are those of Salas and a police officer. Much of it is Salas telling the officer—in English—where he is traveling on the way to and from the transaction. The parts in Spanish are what Salas is saying to someone on the telephone. None of the transcript was read aloud at trial. It is highly doubtful that this recording could have had any significant impact on the jury. Tape 17 is a recording of a telephone conversation between Salas and Defendant to set up the second transaction. It is almost all in Spanish except at the beginning, when the officer supervising the call describes the plan, and at the end, when the officer states that the conversation has ended and gives the time. During the conversation with the seller, everything is in Spanish except a few okays and the seller’s saying “36, at least” and “He’s going to fuck it up, fuck your money too, you know?” Tape 16 is a recording of the second transaction. The first part contains Salas’s side of some telephone conversations. After the first conversation Salas, speaking in 28 Transcripts were prepared for each recording. The transcripts purport to include

all the intelligible Spanish and a translation of that Spanish together with whatever was

said in English. Wyoming Highway Patrol Lieutenant Joseph Scimone testified to the

accuracy of the translations and the transcripts. He testified that he grew up in a bilingual

household speaking Spanish and English, that he had learned Spanish terms used in drug

trafficking through his training and experience, that he had taught Spanish to new

members of the Highway Patrol, and that he had previously testified in federal and state

court about his translations from Spanish. He had listened to the recordings while

reviewing a draft of the transcript and made corrections to the translation as appropriate.

Salas testified to whom he spoke with on the recordings.

The three recordings and the transcripts were provided to defense counsel with

ample time before trial. The transcripts were admitted into evidence. At trial, defense

counsel stated that he had no complaints about the adequacy of the notice he received of

the recordings or translations and said, “I don’t challenge that they are accurate

translations of what Trooper Scimone has heard or read.” R., Vol. 3 at 95. He declined

to ask any questions of Scimone.

During trial the contents of Exhibit 16 and 17 were presented to the jury by having

the prosecutor and Agent Jason Ruby read them. The prosecutor read what was said by

Salas and Dominguez. Ruby read what was said by Defendant. While they read each

English, tells the listening officers what the plan is. Then the recording has three voices, identified as Salas, Defendant, and (for a few words) the driver, Dominguez. Portions of the recorded conversation are in English.

29 transcript, the transcript was displayed on a screen in the courtroom. At no time during

the recitation did defense counsel challenge the accuracy of the translation or complain

that the reading of the transcripts to the jury did not reflect what was said (as translated

into English when necessary) or what was unintelligible.

Because of the care by law-enforcement agents with respect to each controlled

buy, there could be little doubt that on both occasions Salas traded money for

methamphetamine with someone in each white car. The only question was whether

Defendant was involved in the sale. And the only direct evidence of that came from

Salas and Miranda. Thus, their credibility was the chief dispute in the case.

The tape recordings, however, were of little consequence on that issue, because

the identification of the people with whom Salas spoke came solely from Salas and

Miranda. For the tapes to be probative of Defendant’s specific involvement in the

transactions (and thus corroborate the accusations of Salas and Miranda), the jury had to

believe Salas’s identification of the voices. The defense therefore focused on impeaching

the credibility of Salas and Miranda, particularly on the ground that they were receiving

leniency for their own crimes. And by the same token, the specific words spoken in the

transcripts were of minimal concern. It is fair to assume that the defense had little

interest in nitpicking the translation in the transcripts because (1) a review of the

recordings showed that the translations were essentially correct and (2) revisions to a

word here or there would have no impact on the jury.

The majority opinion reverses Defendant’s convictions because the three tape

recordings were not admitted into evidence. The trial judge excluded them because the

30 Wyoming jury would not understand the Spanish. (Defense counsel had represented that

“the vast majority of [the recording during the second transaction] is in Spanish.” R.,

Vol. 3 at 267.) Under Federal Rule of Evidence 402, “Irrelevant evidence is not

admissible.” And under Rule 403, “The court may exclude relevant evidence if its

probative value is substantially outweighed by a danger of one or more of the following:

unfair prejudice, confusing the issues, misleading the jury, undue delay, wasting time, or

needlessly presenting cumulative evidence.” Absent some explanation of why the

recording would assist the jury in assessing the facts of the case, I would have thought

that those two rules of evidence would surely support the district court’s ruling.

In contrast, the English translations of the Spanish-language conversations were

clearly relevant, since they informed the jury of what was said during the arrangements

for and the conduct of the drug sales for which Defendant was convicted. The majority

opinion declares, however, that the translations cannot be admitted into evidence unless

the recordings themselves are admitted. For reasons I have already discussed at some

length, I respectfully disagree. The translations of the recordings were expert testimony

that is admissible regardless of whether the translated recordings themselves were

admitted into evidence.

True, the government never proffered Lt. Scimone as an expert witness, and the

district court never declared him qualified as an expert. But there was no need to do so

since the translation was not challenged. Scimone clearly had the knowledge and

experience to perform the translation. To be sure, it is not uncommon for an attorney

putting on expert testimony to request that the court recognize the witness as an expert

31 and to obtain such recognition. But that procedure is not required. Opposing counsel

may decline to raise the issue of the witness’s qualification as an expert because the

qualifications are clear and a judicial pronouncement that the witness is an expert may

lead the jury to give the witness’s opinion excessive weight. Indeed, having a judge

declare a witness to be qualified as an expert is generally frowned upon. See Fed. R.

Evid. 702, 2000 advisory committee note (“[T]here is much to be said for a practice that

prohibits the use of the term ‘expert’ by both the parties and the court at trial. Such a

practice ‘ensures that trial courts do not inadvertently put their stamp of authority’ on a

witness’s opinion, and protects against the jury’s being ‘overwhelmed by the so-called

‘experts.’”). The American Bar Association’s Civil Trial Practice Standard 14 states:

“‘Qualifying’ Expert Witnesses. The court should not, in the presence of the jury,

declare that a witness is qualified as an expert or to render an expert opinion, and

counsel should not ask the court to do so.” ABA Civil Trial Practice Standard

14 (Aug. 2007) (available at

https://www.americanbar.org/groups/litigation/policy/civil_trial_standards/);5 see

5 It is worth quoting in full the comment to the standard: It is not uncommon for a proponent of expert testimony to tender an expert witness to the court, following a recitation of the witness’s credentials and before eliciting an opinion, in an effort to secure a ruling that the witness is “qualified” as an expert in a particular field. The tactical purpose, from the proponent’s perspective, is to obtain a seeming judicial endorsement of the testimony to follow. It is inappropriate for counsel to place the court in that position. A judicial ruling that a proffered expert is “qualified” is unnecessary unless an objection is made to the expert’s testimony. If an objection is made to an expert’s qualifications, relevancy of expert testimony, reliability or any 32 other aspect of proffered expert testimony, the court need only sustain or overrule the objection. When the court overrules an objection, there is no need for the court to announce to the jury that it has found that a witness is an expert or that expert testimony will be permitted. The use of the term “expert” may appear to a jury to be a kind of judicial imprimatur that favors the witness. There is no more reason for the court to explain why an opinion will be permitted or to use the term “expert” than there is for the court to announce that an out-of-court statement is an excited utterance in response to a hearsay objection. Because expert testimony is not entitled to greater weight than other testimony, the practice of securing what may appear to be a judicial endorsement is undesirable. As United States District Judge Charles R. Richey has observed in a related context, “It may [ ] be an inappropriate judicial comment . . . for the court to label a witness an ‘expert.’” Hon. Charles R. Richey, Proposals to Eliminate the Prejudicial Effect of the Use of the Word “Expert” Under the Federal Rules of Evidence in Civil and Criminal Jury Trials, 154 F.R.D. 537, 554 (1994). The prejudicial effect of this practice is accentuated in cases in which only one side can afford to, or does, proffer expert testimony. When the Advisory Committee on the Federal Rules of Evidence recommended what became the December 1, 2000 amendment to Fed. R. Evid. 702, it cited Judge Richey and ended the Advisory Committee Note accompanying the amendment with the following paragraph: The amendment continues the practice of the original Rule in referring to a qualified witness as an “expert.” This was done to provide continuity and to minimize change. The use of the term “expert” in the Rule does not, however, mean that a jury should actually be informed that a qualified witness is testifying as an “expert.” Indeed, there is much to be said for a practice that prohibits the use of the term “expert” by both the parties and the court at trial. Such a practice “ensures that trial courts do not inadvertently put their stamp of authority” on a witness’ opinion, and protects against the jury’s being “overwhelmed by the so-called ‘experts.’” Hon. Charles Richey, Proposals to Eliminate the Prejudicial Effect of the Use of the Word “Expert” Under the Federal Rules of Evidence in Criminal and Civil Jury Trials, 154 F.R.D. 537, 559 (1994) (setting forth limiting instructions and a standing order employed to prohibit the use of the term “expert” in jury trials) .

33 This Standard suggests that the court should not use the term “expert” and that the proponent of the evidence should not ask the court to do so. The party objecting to evidence also has a role to play in assuring that the court does not appear to be anointing a witness as an “expert.” A party objecting that a witness is not qualified to render an opinion or that a subject matter not the proper subject of expert testimony should avoid using the word “expert” in the presence of the jury. Any objection in the presence of the jury should be “to the admissibility of the witness’ opinion.” If the objecting party objects that testimony is inadmissible “expert” testimony and the court overrules the objection, it may appear that the judge has implicitly found the witness to be an “expert.” When an objection is made, if the proponent wishes to argue the matter, it should be outside the hearing of the jury. See Fed. R. Evid. 103(c) (providing that inadmissible evidence should not be heard by the jury). The utility of the Standard can be undermined if the court is not careful to excise the term “expert” from the instructions it gives to the jury before it deliberates. Juries can be fully instructed on their role in assessing credibility without any mention of the term. The following instruction is illustrative: Some witnesses who testify claim to have special knowledge, skill, training, experience or education that enable them to offer opinions or inferences concerning issues in dispute. The fact that a witness has knowledge, skill, training, experience or education does not require you to believe the witness, to give such a witness’s testimony any more weight than that of any other witness, or to give it any weight at all. It is important for you to keep in mind that the witness is not the trier of fact. You are the trier of fact. It is for you to decide whether the testimony of a witness, including any opinions or inferences of the witness, assists you in finding the facts and deciding the issues that are in dispute. And, it is for you to decide what weight to give the testimony of a witness, including any opinions or inferences of the witness. 6 Stephen A. Saltzburg, Michael M. Martin, & Daniel J. Capra, Federal Rules of Evidence Manual 144 (8th ed. 2002).

34 United States v. Nixon, 694 F.3d 623, 629–31 (6th Cir. 2012) (adopting standard

14); United States v. Bartley, 855 F.2d 547, 852 (8th Cir. 1988); 29 Wright &

Miller § 6264.3 (2d ed.).

I recognize that this court has not prohibited trial judges from declaring in

open court that a witness is an expert (although perhaps we should). But we have

never prohibited the parties and the court from proceeding in accordance with the

better practice recommended by the ABA Standard and the advisory committee on

the Rules of Evidence.

In this case the government called Scimone as a witness, asked him to describe his

background, which clearly established his ability to translate from Spanish into English,

and elicited the witness’s endorsement of the translation in the transcript. Defense

counsel did not object. There was nothing more that the court needed to do or should

have done. See Gomez, 67 F.3d at 1525 (“Mr. Gomez did not object at trial to the

admission of the expert testimony and, therefore, the issue is not properly before this

court.” Only review for plain error would be proper.); 4 Weinstein & Berger § 702[6][a]

(“[O]f course, reliability determinations are necessary only if the opponent objects to the

admissibility of the expert testimony. In the absence of an objection, rulings admitting or

excluding expert testimony without a reliability determination are reviewable only for

plain error.”); United States v. Cristerna-Gonzales, 962 F.3d 1253, 1260 (10th Cir. 2020)

(“To begin with, we question whether there is any error in admitting expert-opinion

testimony without a judicial ruling on the witness’s qualifications when no objection has

been raised to the testimony. A comment to ABA Civil Trial Practice Standard 14 states

35 the Bar’s understanding that ‘[a] judicial ruling that a proffered expert is “qualified” is

unnecessary unless an objection is made to the expert’s testimony.’” ). The Federal Rules

of Evidence Manual states the point succinctly: “The rules set forth [evidence Rules 104,

615, 702, 703, 704, 705, and 706] are triggered when a timely objection is made to expert

testimony. Unless a timely objection is made, the evidence will be admitted.” Federal

Rules of Evidence Manual § 7-2[c].

The majority opinion suggests that Scimone’s testimony was inadmissible because

he did not reasonably rely on the draft of the transcript prepared by unidentified other

people. This suggestion is puzzling. Scimone testified that he listened to the recorded

conversation and that the transcript, as edited by him, was an accurate translation of the

conversation. It is not as if those who prepared the draft translation had access to

anything or performed any test that Scimone did not. He listened to the same recordings

that they did. There is nothing to indicate that he deferred to someone else’s translation.

It is irrelevant what person (or even computer program) prepared the draft, because

Scimone testified that he made corrections so that the transcript presented in court was, in

his view, an accurate translation.6 The only thing he “relied” on other than his own

expertise was the recording itself. Moreover, Defendant never challenged Scimone’s

qualifications to testify to his translation or his reliance on anything he used to perform

the translation. Indeed, he did not challenge the accuracy of the translation.

6 Perhaps his corrections to the translation account for the apparent discrepancies between the translation column and the column that included the original Spanish. He may have deciphered what had been labeled as unintelligible. 36 A final comment on whether the recordings should have been admitted into

evidence. Defendant’s sole argument at trial was that they had to be admitted under the

Best Evidence Rule if the translations were to be admitted. I have already explained why

that argument fails. He could have argued that the recordings should be admitted because

they were relevant. But he did not do so. The trial judge saw no point in playing tapes

that were in a foreign language unfamiliar to the jurors. I recognize that parts of the

recording were in English, but Defendant’s arguments have exclusively focused on the

Spanish-language conversation, and he has not explained how he was prejudiced by the

jurors’ not hearing the English portions. Defense counsel said nothing that would

disabuse the trial court of its conclusion that no purpose would be served by playing the

recording. In any event, Defendant has waived any claim that the tapes would be

relevant by not making them part of the record on appeal.

In short, the district court’s rulings did not violate the Best Evidence Rule.

Defendant’s conviction should be affirmed.