United States v. York

TINDER, Circuit Judge. As Darvell York spoke to Tracy Mitchell about selling him “nine probably hard,” law enforcement was listening. Agents had set up a sting to buy nine ounces of crack cocaine from York, and Mitchell was their informant. At York’s trial, the government played the recorded conversations for the jury. Though York and Mitchell seemed to speak in plain English, without any reference to drugs, the government argued that they were really conversing in the cryptic, coded language of the narcotics trade. The government called two law enforcement witnesses to *418 interpret this drug jargon for the jury. These interpretations made clear that York and Mitchell were negotiating a drug deal, despite their words being facially benign — e.g., “nine probably hard” meant nine ounces of crack cocaine. At this, York cries foul, stating a slew of reasons why he believes this interpretation testimony should have been thrown out. After reviewing the admissibility of each agent’s testimony, we find that the district court was correct in admitting the vast majority of the agents’ translations; only a small portion of one agent’s testimony was erroneously admitted. Consequently, the crux of this case is whether those few errors were harmless. We conclude that they were and therefore affirm. I On April 80, 2003, law enforcement executed the sting to buy crack from York. Agents equipped Mitchell with an audio transmitter and an audio recording device to monitor in real time and record any of Mitchell’s conversations. They gave him $6,500 cash to make the buy. Mitchell waited for York at a residential jobsite where Mitchell was supervising a landscaping crew, while a couple of blocks away agents looked on with binoculars and a video camera. York first showed up in the early afternoon. As he talked to Mitchell, the covert audio devices picked up what he had to say: MITCHELL: ____ So I was (U/I) going to get a half. You know what I’m saying? And then I was gonna boost the half up like, cause you can, how much can you make off of a half of one? YORK: I don’t know. It depends on the work you know? You still want the shit to be sellable, man. You know? MITCHELL: .... Just get you know what I’m saying? Nine, nine probably hard right? And then I’ll flip that mother fucker and then I’ll come back, cause she own [sic] me $20,000 for this. That’s all I got left to do right here. YORK: So what you trying to do now? MITCHELL: Just got, just bring me nine. YORK: (U/I) you want me to cook it? MITCHELL: What, it gonna be soft? YORK: Huh? MITCHELL: It’s gonna be soft? YORK: (U/I).... I ain’t did nothin’ in a while, I got to get another mother fucker to get it though. But it gonna be right though. MITCHELL: Okay. Well, that’s fine. But then I’m be here till like eight.... Mitchell and York then got into Mitchell’s van (which the police had previously searched for drugs and found none), and York told Mitchell, “get your money straight,” and said, “That shit costs, uh, six nigger.” They talked for a little while longer in the van, making a couple more references to numbers and money (“five dollar” and “fifty-five”). Then they got out and York took off in his car. Mitchell then rendezvoused with the agents. They debriefed him and searched his person and his car. They did not find any drugs, but they did find that Mitchell had only $500 of the initial $6,500 they gave him. The agents instructed Mitchell *419 to head back to the jobsite and wait for York to return. While he waited, a woman arrived who officers assumed (after checking her car’s plates) to be Mitchell’s wife. Though the audio transmitter’s battery had died and police could not hear their conversation, police observed Mitchell hand her what appeared to be a set of keys. They did not see the woman give anything to Mitchell. York pulled up around six o’clock that evening and this time he had somebody else in the car with him. Mitchell walked over to the passenger side of York’s car and started talking with York and the other man. Then York said, “That’s nine right, that’s nine like that, you wait on it.” The police watched with binoculars and the video camera, but Mitchell’s position blocked their view. So the officers could not see whether York handed something to Mitchell. After Mitchell and York chatted a bit more about Mitchell’s landscaping job, York left. Mitchell then circled back with the agents. They searched him and again found the $500 in cash. This time, however, they found something else — a clear plastic bag that contained nine ounces of crack cocaine. York was indicted on one count of knowingly and intentionally distributing cocaine base (specifically crack) in violation of 21 U.S.C. § 841(a)(1). His first trial ended without a unanimous verdict, resulting in a mistrial. York was retried a few months later. The government’s case was nearly identical at both trials — the same witnesses testified at each and their testimony was substantially the same. In neither trial did the informant Mitchell testify. At the second trial, the jury heard from a number of witnesses, including a chemist, a fingerprint expert, and a set of law enforcement officers. Two of those officers are of concern in this case. The first was FBI Agent Mike Brown, who was one of the .primary agents handling the investigation of York and who helped execute the sting. Brown explained the sting operation to the jury, described what the jurors saw as they watched the video of York meeting with Mitchell, and identified the voices in the audio recordings as Mitchell’s and York’s. He also said that he heard the sound of money being counted while eavesdropping when York and Mitchell were in the van. In addition, Brown described his meeting with Mitchell in between York’s visits, where Mitchell had only $500 of the original $6,500, and his rendezvous with Mitchell after York’s second visit, where Brown found Mitchell with what looked like (and was later determined to be) crack cocaine. But the government did not use Brown solely as a fact witness. Brown had extensive experience in prior drug cases. So the government, without first formally offering Brown as an expert, asked Brown to give his opinion about the meaning of certain words and phrases that Mitchell and York used in their conversations. Brown obliged: “half’ meant half a kilo of cocaine, “nine” meant nine ounces, “hard” meant crack cocaine, “soft” meant powder cocaine, “work” meant the drug business, “cook” meant converting powder cocaine into crack, and “boost up” meant diluting a given quantity of cocaine into a larger volume to have more to sell. Brown also interpreted several words as references to money: York’s reference to “six” meant $6,000, which Brown said was the cost of the drugs; “five dollar” meant $500; and “fifty-five” meant $5,500. Brown wasn’t the only witness to interpret the drug lingo in the recorded conversations. The government called (and formally offered) Officer Robert Coleman as an expert witness. Coleman had extensive experience in narcotics investigations but he was not involved with York’s investiga *420 tion. He only reviewed the transcripts of the recorded conversations so he could give his opinion on their meaning. Most of Coleman’s translations of the code words were identical to Brown’s — the words referred to drugs and drug paraphernalia. Plus, Coleman testified to a few phrases that Brown did not, such as York’s telling Mitchell, “get your money straight,” which Coleman interpreted as York telling Mitchell to “get his money together for the nine-ounce purchase.” Coleman also testified to his knowledge of narcotics transactions in the local area. He testified that wholesale amounts of cocaine are sold in half, quarter, or eighth of a kilogram quantities and that a quarter kilogram (or nine ounces) of crack costs between $5,000 and $6,000 and can be as much as $9,000. The jury convicted York of delivering 50 grams or more of cocaine base. The district court sentenced York to 360 months’ imprisonment and 10 years of supervised release. York appeals both his conviction and his sentence. York seeks a new trial by arguing that the district court should have excluded both Brown’s and Coleman’s interpretation testimonies. We review each agent’s testimony in turn. York seeks re-sentencing based on the retroactive amendment to the crack cocaine guidelines. II. A. Agent Brown’s Interpretation Testimony York levels an array of attacks against Brown’s interpretations of the drug jargon that laced York’s and Mitchell’s recorded conversations. Through various interrelated arguments, York contends that Brown’s interpretation testimony was inadmissible under Federal Rules of Evidence 702 and 403, Federal Rule of Criminal Procedure 16(l)(g), and the Sixth Amendment’s Confrontation Clause. After untangling York’s claims, we find a portion of Brown’s testimony problematic. The district court should have excluded Brown’s interpretations of the words “six,” “five dollar,” and “fifty-five.” The rest of Brown’s interpretations, however, was admissible. To begin, we must determine whether Brown’s interpretations were admissible only as expert opinion testimony under Fed.R.Evid. 702. The government concedes that they were, and we agree. Opinions or inferences based on “scientific, technical, or other specialized knowledge within the scope of Rule 702” are not admissible as lay testimony under Fed. R.Evid. 701. Such opinions or inferences, drawn from facts outside the witness’s first-hand knowledge of the case, are admissible only as expert testimony. United States v. Conn, 297 F.3d 548, 553-54 (7th Cir.2002). Brown’s interpretation testimony fit the “expert” mold. Though Brown had first-hand knowledge of York’s investigation, the government asked Brown to rely on his experience in prior crack cocaine investigations to explain the hidden meaning of certain words in York’s and Mitchell’s conversations. See United States v. Oriedo, 498 F.3d 593, 603 & n. 10 (7th Cir.2007). For the most part, Brown did not claim that he learned the meaning of these words during the course of his investigation of York. (We say “for the most part” because on a few occasions it was unclear what body of knowledge Brown relied on to inform his interpretation of certain words, which as you will later see, causes us some concern.) Cfi United States v. Rollins, 544 F.3d 820, 833 (7th Cir.2008) (finding law enforcement witness’s interpretations of code words as admissible lay testimony where witness based interpretation only on listening first *421 hand to numerous recorded telephone calls in that particular investigation). So by generally relying on his specialized knowledge, Brown testified as an expert. Expert testimony has its benefits for the party who offers it, but it also has its burdens. Rule 702 requires that an expert be qualified “by knowledge, skill, experience, training, or education” to render his opinion, and that the opinion “assist the trier of fact to understand the evidence or to determine a fact in issue.” Rule 702 also requires that “(1) the testimony is based upon sufficient facts or data, (2) the testimony is the product of reliable principles and methods, and (3) the witness has applied the principles and methods reliably to the facts of the case.” In addition, prior to trial, the government must disclose to the defendant a written summary of the expert’s testimony, which “describe[s] the witness’s opinions, the bases and reasons for those opinions, and the witness’s qualifications.” Fed. R.Crim.P. 16(a)(1)(G). Finally, if the witness testifies as both a fact witness and an expert witness in the same trip to the witness stand (like Brown did here), the government and the court must take some special precautions to make clear for the jury when the witness is relying on his expertise and when he is relying only on his personal knowledge of the case. See, e.g., United States v. Upton, 512 F.3d 394, 401 (7th Cir.2008). 1. Qualifying Agent Brown as an “Expert” York argues that the government failed to fulfill a number of these burdens when it asked Brown to opine on the meaning of York’s and Mitchell’s conversations. First, York contends that Brown’s interpretation testimony should have been excluded because the district court failed to formally “qualify” Brown as an expert and did not permit York to conduct a voir dire regarding Brown’s qualifications. District courts have a “gatekeeping” duty to ensure that witnesses do not offer expert testimony before the court is satisfied that Rule 702’s requirements are met. See Kumho Tire Co., Ltd. v. Carmichael, 526 U.S. 137, 147, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999); United States v. Moore, 521 F.3d 681, 684 (7th Cir.2008); Conn, 297 F.3d at 555-56. However, “[a] judge is not obliged to look into the questions posed by Rule 702 when neither side either requests or assists.” Moore, 521 F.3d at 685. York’s trial counsel failed to object to Brown’s qualifications or any other aspect of Brown’s testimony until well after Brown began defining words like “half’ and “hard.” Only near the end of Brown’s interpretation testimony did York’s counsel object to Brown’s “basis of knowledge,” to which the court responded that Brown’s “extensive involvement in prior drug narcotics investigations” made him “competent to testify to his understanding” of York’s and Mitchell’s drug jargon. We are hard-pressed to say that York’s general foundation objection preserved any challenge to Brown’s qualifications, the lack of voir dire, or whether Brown’s testimony was based on “sufficient facts or data” and “reliable principles and methods.” See Fed.R.Evid. 103(a)(1) (requiring a “timely objection ... stating the specific ground of objection, if the specific ground was not apparent from the context” (emphasis added)); Rollins, 544 F.3d at 834. Even if it did, any error flowing from the district court’s failure to formally anoint Brown an expert was harmless. We have routinely held that “narcotics code words are an appropriate subject for expert testimony.” United States v. Farmer, 543 F.3d 363, 370 (7th Cir.2008); see also United States v. Foster, 939 F.2d 445, 451 (7th Cir.1991). A law *422 enforcement officer’s understanding of the drug trade, which comes from that officer’s prior experience policing illicit narcotics transactions, is “specialized knowledge” within Rule 702. See Oriedo, 498 F.3d at 603 & n. 10. So we allow officers whose testimony is based on some aspect of that understanding (such as the meaning of drug code words), rather than on firsthand knowledge of the particular investigation in the case, to testify as experts. Id. at 603-04; United States v. Hughes, 970 F.2d 227, 236 (7th Cir.1992). Notably, York does not challenge Brown’s qualifications, his methods, or the accuracy of his testimony here on appeal. Brown testified that he had served 17 years as an FBI agent and been involved in approximately 200 narcotics investigations prior to testifying. Brown also testified that during his experience in drug investigations he learned some of the language of the drug trade and he relied on that knowledge to define most of the drug lingo. Therefore, given Brown’s qualifications, and no attempt to disparage them here on appeal, we find that Brown would have easily qualified as an expert had the court conducted the formal Rule 702 analysis. Along these lines, we find that York suffered no prejudice by being unable to conduct a formal voir dire of Brown before he testified. York was able to and did cross-examine Brown on his qualifications and methods at trial. On appeal, York gives us no reason to question those qualifications or methods. So we won’t. See Oriedo, 498 F.3d at 604. (“Although Mr. Oriedo makes the blanket assertion that he was deprived of an opportunity to cross-examine adequately Agent Gourley or prepare a defense, before this court he questions neither the accuracy of the statements offered nor the qualifications of Agent Gourley to make them.”). 2. Expert Disclosure Requirements Next, York hints that the government ignored the expert disclosure requirements of Fed.R.Crim.P. 16(a)(1)(G). Indeed, it appears the government did. Since the government planned to elicit expert opinion testimony from Brown, it should have provided a summary of Brown’s testimony to York’s counsel prior to trial. See Oriedo, 498 F.3d at 604. But York never raised this issue below, which limits our review to plain error. United States v. Navarro, 90 F.3d 1245, 1259 (7th Cir.1996). On appeal, York does not identify any prejudice that he suffered as a result of this non-disclosure, which would justify outright exclusion of Brown’s testimony (or any other sanction under Fed. R.Crim.P. 16(d)(2) for that matter). See United States v. Duvall, 272 F.3d 825, 829 (7th Cir.2001). We don’t see how there could be any. Brown’s testimony at York’s second trial mirrored his testimony from the first. At the first trial, Brown explained his extensive experience with narcotics investigations and then relied on that expertise to interpret many of the same code words he defined at the second trial. Moreover, York’s lawyer was the same for both trials. So defense counsel could not claim that he suffered some unfair surprise and was caught unaware of Brown’s qualifications and opinions, and the bases and reasons for those opinions, going into the second trial. This is not to say that every time a defendant is retried the government need not disclose the experts from the prior trial whom the government intends to call again. But here, where Brown’s testimony was nearly identical in both trials, and where York does not allege any disadvantage in preparing for the second trial or cross-examining Brown, we cannot see any prejudice that would justify exclusion nor any plain error that would justify reversal. *423 3. Helpfulness of Agent Brown’s Testimony York next argues that Brown interpreted some words and sounds that were not drug code and therefore needed no interpretation. Specifically, York challenges Brown’s interpretation of the numbers “six,” “nine,” “five dollar,” and “fifty-five,” as well as Brown’s comment that he heard the sound of money being counted over the audio transmitter. This testimony, according to York, exceeded the proper scope of expert testimony under Rule 702. We see two different claims growing out of this single challenge. First, York contends that, by interpreting words already within the jury’s understanding, Brown’s testimony did not meet Rule 702’s requirement that expert testimony “assist” the jury. Second, because these words and sounds were not “code,” in York’s view, Brown’s testimony was not based on sufficient facts and reliable methods as Rule 702 requires. Because York’s trial objection arguably encompasses these claims, we review them for an abuse of discretion. Farmer, 543 F.3d at 370. Turning to that first claim, we have discussed that the Rules of Evidence allow expert law enforcement witnesses to translate drug jargon and code words that might seem entirely innocuous to an untrained jury. United States v. Ceballos, 302 F.3d 679, 687-88 (7th Cir.2002); Foster, 939 F.2d at 451-52 (“Despite our country’s “war on drugs’ and its accompanying media coverage, it is still a reasonable assumption that jurors are not well versed in the behavior of drug dealers.”). But this presupposes that the terms to be interpreted are in fact drug code and not words “that the jury can evaluate for itself.” United States v. de Soto, 885 F.2d 354, 361 (7th Cir.1989). “Interpretations” of unambiguous words or phrases that are plainly within the jury’s understanding are unlikely to be admissible under Rule 702; they would not “assist the trier of fact to understand the evidence or to determine a fact in issue.” See United States v. Rollins, 862 F.2d 1282, 1292 (7th Cir.1988). Instead, they would merely put an expert gloss on a conclusion the jury should draw. In Ceballos, 302 F.3d at 687-88, we upheld agents’ interpretations of simple pronouns such as “it,” “them,” and “both” as referring to methamphetamine shipments, in part, because defendants used those pronouns ambiguously in their conversation (they mentioned no other nouns to which the pronouns could refer). Given this ambiguity, we concluded that the agents’ experience interpreting drug code language would be helpful to the jury. Id. at 688; see also Rollins, 862 F.2d at 1292 (upholding agent’s interpretation of “t-shirts,” “stuff,” and “it” as code words referring to cocaine). Here, the terms “six,” “nine,” “five dollar,” and “fifty-five” were facially ambiguous. York told Mitchell, “That shit costs ... six ... ”, and then later said, “That’s nine right, that’s nine like that, you wait on it.” The question is, six and nine of what? Though York’s reference to “cost” gives the jury some indication that “six” refers to payment, the term “six” was still sufficiently ambiguous (does it mean money or something else and how much?). The references to “five dollar” and “fifty-five” were equally unclear. Given our decisions in Ceballos and Rollins, we have no qualms concluding that Brown’s interpretation of these vague terms would assist the jury. 4. Foundation for Agent Brown’s Interpretations But that alone doesn’t mean Brown’s testimony was admissible — Rule 702 requires more than a qualified expert and helpful testimony. Brown must have *424 had a reliable basis for opining that words like “six” and “fifty-five” are in fact drug lingo. Fed.R.Evid. 702; see also Conn, 297 F.3d at 555 (“The second prong of the test set forth in Rule 702 requires that the testimony be the product of reliable principles and methods.”). York argues that Brown had no basis for such an opinion because these numbers do not refer to anything. To illustrate, York contrasts his case to United States v. Moon, 512 F.3d 359, 363 (7th Cir.2008), where we upheld an agent’s interpretation of “four, five, or six meals” as referring to between four and six kilograms of cocaine. Though the agent interpreted the numbers as quantities of drugs, the code word in Moon was the word “meals.” York’s case is different, he argues, because the numbers in his and Mitchell’s conversations lacked any reference to another noun that could serve as the coded term. York supports his view with the Second Circuit’s opinion in United States v. Dukagjini, 326 F.3d 45, 55 (2d Cir.2003). In Dukagjini, a law enforcement expert witness interpreted the words “six” and “ten” in the expressions, “tell 'em to bring ... the six or whatever,” and “tell him to come with the ten,” to refer to quantities of heroin. Id. (emphasis added). The Second Circuit found that admitting this testimony was erroneous because the words “six” and “ten,” though ambiguous, were not drug code: “There was no evidence that these phrases were drug code with fixed meaning either within the narcotics world or within this particular conspiracy.” Id. Drawing on Dukagjini, York argues that words like “six” and “fifty-five” do not have “fixed meanings” and therefore should not be treated as drug code that an expert can interpret. We find York’s view too narrow. Experts need not establish that certain words have fixed meanings only in the narcotics world or in the particular conspiracy before they can interpret those words. Experts can determine, based on their expertise, that certain words have drug-related meanings within the context of a single conversation. In Ceballos, for example, agents interpreted the words “it,” “them,” and “both” as referring to shipments of methamphetamine. 302 F.3d at 687. Those words certainly lack any “fixed meaning” in the narcotics world or elsewhere — “it” does not always mean meth. But in the context of that conversation, where the pronoun “it” had no antecedent, the agents, drawing on their expertise, had a reliable basis to conclude that those words referred to drugs. Id. at 687-88. The same is true here. Mitchell’s and York’s conversations were laced with words that Brown testified were common drug parlance (and did have fixed meanings in the drug trade) — i.e., “soft” meant powder cocaine, “hard” meant crack cocaine, and “cook” meant processing powder into crack. Brown knew Mitchell and York were talking about drugs. So when Mitchell asked for “nine, nine probably hard,” the word “nine” was not just dangling in the conversation, unlinked to any drug code word, as York suggests. “Nine” and “hard” went together, just like “four, five, and six” went together with “meals” in Moon, 512 F.3d at 363. So Brown had a reliable basis to opine that York and Mitchell were negotiating a crack deal (York also told Mitchell what the “shit costs”), and Brown’s experience in narcotics transactions could inform his opinion about common quantities and prices for the drugs. Between his expertise and the context of the conversation, Brown could interpret “nine” as the agreed-upon quantity and “six,” “five dollar,” and “fifty-five” as efforts to negotiate the price. Brown also had a sufficient basis, though not an expert basis, to testify that he *425 heard the sound of money being counted on the audio transmitter. A witness’s testimony about the sounds he heard is one of the “ ‘prototypical example[s] of the type of evidence contemplated by the adoption of Rule 701.’ ” Fed.R.Evid. 701 advisory committee’s note (insertion in original) (quoting Asplundh Mfg. Div. v. Benton Harbor Eng’g, 57 F.3d 1190, 1196 (3d Cir. 1995)); see also Conn, 297 F.3d at 554 (“Lay opinion testimony most often takes the form of a summary of first-hand sensory observations.”). So this part of Brown’s testimony was admissible as lay opinion testimony. 5. Dual Testimony But did Brown actually rely on his expertise when he interpreted the code words? Or did he rely on some other basis, such as a conversation with the non-testifying Mitchell, which might pose problems under the Rules of Evidence and the Sixth Amendment? That’s the crux of York’s final two challenges to Brown’s testimony — the dual nature of Brown’s testimony and the Confrontation Clause concerns that such dual testimony raises. Because Brown’s general “basis of knowledge” objection did not preserve these two claims, we review them only for plain error. United States v. Pree, 408 F.3d 855, 868-69 (7th Cir.2005). York argues that Brown impermissibly testified as both an expert and a fact witness in the same trip to the witness stand. Though such a practice is routinely upheld, particularly where experienced law enforcement officers were involved in the particular investigation at issue, e.g., United States v. Mansoori, 304 F.3d 635, 654 (7th Cir.2002), there are some inherent dangers with this kind of dual testimony, see Upton, 512 F.3d at 401; de Soto, 885 F.2d at 360; Dukagjini, 326 F.3d at 53-54. For example, the witness’s dual role might confuse the jury. United States v. Goodwin, 496 F.3d 636, 641 (7th Cir.2007); Fed. R.Evid. 403. Or, the jury might be smitten by an expert’s “aura of special reliability” and therefore give his factual testimony undue weight. United States v. Brown, 7 F.3d 648, 655 (7th Cir.1993). Or, “the jury may unduly credit the opinion testimony of an investigating officer based on a perception that the expert was privy to facts about the defendant not presented at trial.” Upton, 512 F.3d at 401. Given these dangers, district courts must take some precautions to ensure the jury understands its function in evaluating this evidence. Id. The jury needs to know when an agent is testifying as an expert and when he is testifying as a fact witness. “The potential for prejudice in this circumstance can be addressed by means of appropriate cautionary instructions and by examination of the witness that is structured in such a way as to make clear when the witness is testifying to facts and when he is offering his opinion as an expert.” Mansoori, 304 F.3d at 654. We have recognized other precautions as well, such as the government’s establishing the proper foundation for the witness’s expert opinions, Farmer, 543 F.3d at 370-71, and the district court allowing the defense to rigorously cross-examine the expert about his interpretation of the drug lingo, id. at 371; United States v. Parra, 402 F.3d 752, 759-60 (7th Cir.2005). The protective steps taken in this case were not the model of how to handle a witness who testifies in a dual capacity. We recognize that the government established an adequate foundation for Brown’s testimony, as we have discussed, outlining his years of expertise with drug investigations. And we certainly credit the fact that the court put no limits on York’s opportunity to cross-examine Brown; York’s counsel took full advantage of that *426 opportunity to delve into both Brown’s qualifications and his interpretations of certain words in the recorded conversations. Counsel repeatedly questioned Brown’s basis for his understanding of the words, highlighting the fact that neither Mitchell nor York ever specifically mentioned cocaine or crack in their conversations. These measures went part of the way toward minimizing the dangers of dual testimony. But the court and the government were less vigilant in instructing the jury and structuring Brown’s testimony. For instance, though the court did instruct the jury on how it should evaluate opinion testimony from witnesses with special knowledge or skill, this instruction came at the end of the trial. It would have been far more effective for the court to have explained Brown’s dual role to the jury before Brown testified and then flag for the jury when Brown testified as a fact witness and when he testified as an expert. See Upton, 512 F.3d at 401 (“Before [Detective] Eversman’s testimony, the district court gave a cautionary instruction explaining that Eversman would be serving both functions as a witness.” (emphasis added)). What gives us the greatest cause for concern, though, is the structure of Brown’s testimony. The government started off well. It appropriately signaled to the jury that Brown was relying on his expertise and not his knowledge of York’s investigation when it asked Brown whether, during his involvement in over 200 investigations, he learned some terms of the drug trade. In its follow-up questions, the government took a similar tack, prefacing its questions with phrases like, “based on your experience in crack cocaine investigations .... ” This structure helped minimize jury confusion. See Farmer, 543 F.3d at 371 (approving questions with similar introductory remarks). But then things got murky. The government switched back to questioning Brown about the investigation, which of itself might not have been problematic, had the government not decided, several moments into Brown’s factual testimony, to go back and question Brown about a few more code words — “six,” “fifty-five,” and “five dollar.” Seamlessly switching back- and-forth between expert and fact testimony does little to stem the risks associated with dual-role witnesses. Even more problematic was the way in which the government prefaced these questions: “Based on your experience of [sic] crack cocaine investigations and this investigation in particular .... ” (emphasis added). This phrasing explicitly mixed Brown’s dual bases of knowledge, leaving the jury to wonder who was testifying, Brown-the-expert or Brown-the-case-agent. Given this heightened possibility for juror confusion, coupled with the lack of a timely cautionary instruction and the fact that we cannot discern whether Brown’s interpretations were actually based on his expertise or a conversation with Mitchell, we conclude that the court erred in admitting Brown’s responses to the government’s questions about “six,” “fifty-five,” and “five dollar.” (We will address whether those were plain errors momentarily.) Other than these three interpretations, though, Brown’s testimony did not offend Rules 403 and 702 to such an extent that we can say the district court erred in admitting it. 6. Confrontation Lastly, York argues that Brown’s interpretation testimony violated his Sixth Amendment right “to be confronted with the witnesses against him.” York’s argument grows out of Crawford v. Washington, in which the Supreme Court held that *427 the Sixth Amendment’s Confrontation Clause prohibits admitting testimonial hearsay evidence unless the declarant is unavailable and the defendant had a prior opportunity to cross-examine him. 541 U.S. 36, 68, 124 S.Ct. 1354, 158 L.Ed.2d 177 (2004); see also United States v. Tolliver, 454 F.3d 660, 664-65 (7th Cir.2006). Our focus here is on the fact that Crawford applies only to hearsay, which must be a statement offered for the truth of the matter asserted. Crawford, 541 U.S. at 59 n. 9, 124 S.Ct. 1354 (“[The Confrontation Clause] does not bar the use of testimonial statements for purposes other than establishing the truth of the matter asserted.”); see Fed.R.Evid. 801(c). We are no stranger to Crawford-based Confrontation Clause challenges to the admission of previously recorded conversations between informants and defendants. E.g., United States v. Van Sach, 458 F.3d 694, 700-02 (7th Cir.2006); Tolliver, 454 F.3d at 666. We have held that playing the tapes of those conversations for the jury does not violate the Confrontation Clause so long as those tapes are offered to provide context for the defendant’s own admissions. United States v. Nettles, 476 F.3d 508, 517-18 (7th Cir.2007); Van Sach, 458 F.3d at 701; Tolliver, 454 F.3d at 666. When offered for context and not for the truth, the declarant’s statements are not hearsay, United States v. Gajo, 290 F.3d 922, 930 (7th Cir.2002); United States v. Davis, 890 F.2d 1373, 1380 (7th Cir.1989), and thus not subject to Crawford. In this case, the government offered Mitchell’s recorded statements to help the jury to understand York’s statements during his dealings with Mitchell. Mitchell’s statements were not offered for their truth — i.e., it was irrelevant whether Mitchell actually desired nine ounces of cocaine or intended to “boost up” the drugs. Mitchell’s statements were offered to show how York reacted to them. For example, after Mitchell said, “Just bring me nine,” York responded, ‘You want me to cook it?” Mitchell’s statement allowed the jury to understand that York was not offering to fix supper. In other words, Mitchell’s statements put York’s reactions into “context” and hence were not hearsay. Moreover, we see no indication that Mitchell tried to “put words into [York’s] mouth or ... persuade [York] to commit more crimes in addition to those that [York] had already decided to commit.” Nettles, 476 F.3d at 518. So admitting Mitchell’s statements, by itself, did not offend York’s confrontation rights. York concedes as much here on appeal. But York contends that Brown’s interpretations of Mitchell’s statements make this case different. In essence, York argues that Brown’s interpretations transformed Mitchell’s recorded statements from merely providing context to being offered for their truth. But interpreting an informant’s statements would not alter the government’s use of those statements. Brown translated Mitchell’s code words into terms that jurors could understand. If Mitchell had not used code and instead said plainly to York, “I need nine ounces of crack cocaine,” that statement would still be admissible as context for York’s responses. Whether Mitchell spoke to York in drug lingo or plain English would not affect the purpose for which those statements were used. Because Mitchell’s out-of-court statements, translated or untranslated, were offered as context for York’s side of these inculpatory discussions, they do not implicate the Confrontation Clause. We might have a confrontation problem, however, if Brown based his interpretations on his own conversations with Mitchell and not on his knowledge of the drug trade and review of the transcripts. See *428 United States v. Silva, 380 F.3d 1018, 1020 (7th Cir.2004); United States v. Mejia, 545 F.3d 179, 198-99 (2d Cir.2008). For instance, if Mitchell told Brown that “hard” meant crack and Brown relied on that information when he interpreted the word “hard” at trial, Mitchell’s statements to Brown (as opposed to Mitchell’s conversations with York) would have been offered for their truth. No matter that Mitchell’s recorded conversations were offered for context, Mitchell’s statements to the agents would have been hearsay and the agents’ reliance upon them might implicate Crawford. But we see little or no evidence of that here. Brown did speak to Mitchell during the investigation: he “debriefed” Mitchell after both encounters with York. Brown testified that, after the first encounter, he told Mitchell to “go back and wait for Mr. York to deliver the quantity of drugs that he just ordered.” York argues that Brown’s knowledge of Mitchell’s “order,” including the details of that order (e.g., the thing ordered, quantity, and cost), could only have come from a conversation with Mitchell. Not so. As we have discussed, Brown had a working knowledge of the meanings of drug jargon and code words. Brown listened to Mitchell’s and York’s conversations in real time and, by the time he debriefed Mitchell, was well aware that they were negotiating a drug transaction. Moreover, before testifying at trial, Brown reviewed the transcripts of the recordings and testified about the meaning of certain words based on his experience. We have no indication that those interpretations were based on conversations with Mitchell. All we have is a mere possibility of an impermissible basis for three of those interpretations — “six,” “fifty-five,” and “five dollar” — which Brown defined relying on both his expertise and involvement in York’s investigation. Because we have already excluded those interpretations as improperly safeguarded dual testimony, however, we need not decide whether their admission also violated Crawford. In sum, we find that most of Brown’s interpretation testimony was admissible. However, given the lack of precautions taken to minimize the dangers of dual testimony, Brown’s interpretations of “six,” “fifty-five,” and “five dollar” as referring to certain dollar amounts should have been excluded. B. Officer Coleman’s Interpretation Testimony York also challenges Coleman’s testimony. Like Brown, Coleman interpreted the drug jargon and code words in York’s and Mitchell’s conversations. Unlike Brown, Coleman was formally qualified as an expert and had no experience with York’s investigation beyond reviewing the audio recording transcripts. Still, York argues that the district court should have excluded Coleman’s testimony for two of the same reasons that he believes Brown’s testimony was inadmissible. First, York contends that Coleman’s testimony violated the Confrontation Clause. But we know from our previous discussion of this issue that York’s argument here must fail. Simply because Coleman interpreted Mitchell’s words based on his expertise did not change the government’s use for playing the audio tapes — to provide context for York’s admissions. Using the tapes and Coleman’s interpretation of them did not implicate Crawford. Moreover, Coleman was not involved in the investigation and never spoke with Mitchell. Coleman interpreted Mitchell’s statements based solely on his expertise; York does not contend otherwise. So Coleman’s testimony did not *429 draw on any hearsay from Mitchell and therefore did not infringe York’s confrontation rights. Second, York argues that Coleman interpreted words that needed no interpretation, thereby exceeding the proper scope of expert testimony. York challenges all of Coleman’s interpretation testimony and specifically targets Coleman’s translation of numbers like “six” and “nine” as well as his interpretation of the phrase “get your money straight” (a phrase on which Brown did not comment). We review for an abuse of discretion. Farmer, 543 F.3d at 370. As we have discussed, York’s and Mitchell’s vague or coded references to drugs and money were ambiguous and not readily understood by lay jurors. So Coleman’s interpretations of words like “hard,” “soft,” “six,” and “nine” assisted the jury in understanding those words. Fed.R.Evid. 702; see Ceballos, 302 F.3d at 688. In addition, Coleman’s expertise gave him a reliable basis to opine on the meanings of those words. Coleman did not rely on his personal knowledge of the investigation— he didn’t have any beyond the transcripts. Instead, Coleman testified that, in his experience, nine ounces of cocaine (equal to a quarter kilogram) was a wholesale quantity and that the price of that quantity was “anywhere between five or $6,000 up to maybe 9,000.” Coleman’s knowledge of common quantities and prices gave him a reliable basis to interpret the otherwise undefined terms “six” and “nine” as $6,000 and nine ounces of cocaine. Therefore those interpretations were admissible. This reasoning also extends to Coleman’s interpretation of “get your money straight.” York contends that Coleman’s testimony was unhelpful because this phrase had no other reasonable interpretation than the one Coleman gave it: ‘York is telling Mitchell to get his money together for the nine-ounce purchase.” We disagree. The phrase might have meant a variety of things, such as “get your money from a clean source,” or it might have referred to a desire for bills of certain denominations, or York might have been telling Mitchell to physically straighten up the cash he brought with him. In other words, we think the phrase “get your money straight” was just another form of drug slang and, without Coleman’s interpretation, would have remained ambiguous to jurors. The court did not err in admitting that interpretation. C. Harmless Error To review where we are at this point, we think that the district court should have excluded Brown’s interpretations of “six,” “fifty-five,” and “five dollar” as improper dual testimony. But that doesn’t mean we must reverse. Under either a plain error standard or an abuse-of-discretion standard, if those errors were harmless, York’s conviction will stand. United States v. Ortiz, 474 F.3d 976, 982 (7th Cir.2007). Harmlessness means that the jury would have convicted even absent the errors. Id.; see also United States v. Owens, 424 F.3d 649, 656 (7th Cir.2005) (“The test for harmless error is whether, in the mind of the average juror, the prosecution’s case would have been ‘significantly less persuasive’ had the improper evidence been excluded.”). We are convinced that failing to exclude these three interpretations was harmless; the evidence against York was overwhelming. The jury knew that after Mitchell’s first meeting with York, Mitchell had $6,000 less than when he started. And after their second meeting, Mitchell had nine ounces of crack cocaine on him. The admissible portions of Brown’s testimony and all of Coleman’s bolstered the inference that Mitchell and York negotiated *430 and executed a drug deal and diminished any possibility that Mitchell received the drugs from anywhere else, such as from his wife. Their interpretations of “half,” “nine,” “hard,” “soft,” “work,” “cook,” “sellable,” and “boost up” made York’s intentions clear. Even without Brown’s interpretation of certain numbers, Coleman told the jury that York and Mitchell were discussing prices, that $6,000 was within the range of going rates for cocaine, and that in his experience “six” meant $6,000. We think it impossible that Brown’s testimony reinforced Coleman’s in such a way that the exclusion of a small portion of Brown’s testimony would have caused the jury to reach a different verdict. Because any error below was harmless beyond a reasonable doubt, we affirm York’s conviction. III. York wants to petition the district court to reduce his sentence in light of the retroactive application of the revised crack guidelines. See Supplement to the 2007 United States Sentencing Guidelines Manual at 1-4 (Mar. 3, 2008) (U.S.S.G. § lB1.10(c)); United States Sentencing Commission Guidelines Manual, Supplement to Appendix C, 226-31 (2008) (Amendment 706). So he asks us to remand his sentence. But remand is not required to pursue that avenue of relief. United States v. Tatum, 548 F.3d 584, 588 (7th Cir.2008). Instead, York should file a motion in the district court pursuant to 18 U.S.C. § 3582(c)(2). Id. IV. In light of the foregoing, we affirm the judgment of the district court.