"The Sixth Amendment ... prohibits the introduction of testimonial statements by a nontestifying witness, unless the witness is 'unavailable to testify, and the defendant had a prior opportunity for cross-examination.' " Ohio v. Clark , --- U.S. ----, 135 S. Ct. 2173, 2179, 192 L.Ed.2d 306 (2015) (quoting Crawford v. Washington , 541 U.S. 36, 54, 124 S.Ct. 1354, 158 L.Ed.2d 177 (2004) ). Because the court's decision contravenes this core tenet of the Sixth Amendment by permitting the State to introduce testimonial statements made by an unidentified interpreter working from an unidentified location without calling that interpreter as a witness, I respectfully dissent.
To secure defendants' rights to confront their accusers, the Sixth Amendment generally *424prohibits the State from admitting "testimonial statements."1 Id. The first task in a Confrontation Clause analysis, therefore, must be to identify the statement in question. As an example, the court states that "Lopez-Ramos admitted to having sexual intercourse with the victim," presumably relying on the statement "We [sic] had intercourse with her." But Lopez-Ramos never said "We had intercourse with her." He said something in Spanish, and then an interpreter, appointed and paid by the police, said, "Lopez-Ramos said, 'We had intercourse with her.' "2
I agree with the court that "a defendant cannot complain that he was denied the opportunity to confront himself." Thus, the State was fully entitled (subject to the Minnesota Rules of Evidence) to admit the Spanish versions of Lopez-Ramos's statements. But Lopez-Ramos's statements in Spanish are not the statements at issue in this case. The statements at issue are the statements-made by the interpreter-purporting to translate what Lopez-Ramos said. But I know of no instance where a court has held that a third-party declarant's testimonial statements to the police-even statements alleging a confession by the defendant-are admissible without affording the defendant an opportunity to cross-examine the declarant. And the reason why is simple: "Testimonial statements of witnesses absent from trial [are admissible] only where the declarant is unavailable, and only where the defendant has had a prior opportunity to cross-examine." Crawford , 541 U.S. at 59, 124 S.Ct. 1354.3 "To be sure, the [Confrontation] Clause's ultimate goal is to ensure reliability of evidence, but it is a procedural rather than a substantive guarantee. It commands, not that evidence be reliable, but that reliability be assessed in a particular manner: by testing in the crucible of cross-examination."4 Id. at 61, 124 S.Ct. 1354.
*425The court sidesteps Crawford 's mandate by concluding that the Sixth Amendment does not apply to the interpreter's statements because he was acting as a "language conduit." But this language-conduit theory has no support in our precedent, and is undermined by our analysis in State v. Caulfield , 722 N.W.2d 304 (Minn. 2006), and the Supreme Court's subsequent decisions in Melendez-Diaz v. Massachusetts , 557 U.S. 305, 129 S.Ct. 2527, 174 L.Ed.2d 314 (2009), and Bullcoming v. New Mexico , 564 U.S. 647, 131 S.Ct. 2705, 180 L.Ed.2d 610 (2011).
Each of these cases involves a similar scenario: The primary evidence against the defendant was a forensic laboratory report. Without calling the analyst who prepared the report, the State offered the report into evidence, the district court received the report, and the defendant was convicted. Bullcoming , 564 U.S. at 651, 131 S.Ct. 2705 ; Melendez-Diaz , 557 U.S. at 308-09, 129 S.Ct. 2527 ; Caulfield , 722 N.W.2d at 306. In each case, the reviewing court concluded that the report was testimonial evidence against the defendant, and if the State wished to offer the report into evidence, the Confrontation Clause required the State to call the authoring analyst.5 Bullcoming , 564 U.S. at 652, 131 S.Ct. 2705 ; Melendez-Diaz , 557 U.S. at 311, 129 S.Ct. 2527 ; Caulfield , 722 N.W.2d at 306-07.
This case closely parallels Bullcoming , Melendez-Diaz , and Caulfield . Here, the primary evidence against Lopez-Ramos was the interpreter's report of what Lopez-Ramos said during his interview. Without calling the interpreter to testify-and without even proffering the interpreter's full name and location-the State offered the interpreter's translations into evidence, the district court admitted the translations, and Lopez-Ramos was convicted. Following Bullcoming , Melendez-Diaz , and Caulfield , Lopez-Ramos's conviction should be reversed because the State did not call the interpreter so that he could be cross-examined about his translations.
We are not the first court to consider the implications of Bullcoming and Melendez-Diaz on the admissibility of translator statements. As the court notes, both the Eleventh Circuit and Maryland's intermediate appellate court have applied those cases to reject the language-conduit theory. See United States v. Charles , 722 F.3d 1319, 1329-30 (11th Cir. 2013) ; Taylor v. State , 226 Md.App. 317, 130 A.3d 509, 539-41 (Md. Ct. Spec. App. 2016). Both cases strongly critique the view-wrongfully endorsed by this court's majority-that judges can "make a threshold determination of the interpreter's honesty, proficiency, and methodology without testimony from the one witness whose testimony could best prove the accuracy of the interpretations -the interpreter himself or herself." Taylor , 130 A.3d at 539 (emphasis added); see also Charles , 722 F.3d at 1327-28.
The court attempts to distinguish these cases, arguing that "[b]ecause we believe the facts of Crawford are materially different *426from cases involving an interpreter, the underlying logic of decisions in cases like Charles and Taylor is distinguishable." Assuming arguendo that the facts of Crawford can be distinguished from translator cases (perhaps because the Crawford declarant was a lay witness who saw the substantive events of the crime, whereas translators perform an expert analysis-translation-after the fact), Charles and Taylor were not based solely on Crawford ; they rely just as much, if not more, on Bullcoming and Melendez-Diaz . As I discuss below, the court errs when it concludes that the laboratory analysts of those cases are distinguishable from the translators in Charles , Taylor , and this case. The logic of Charles and Taylor is sound, and the court errs in rejecting it.
Turning to Bullcoming and Melendez-Diaz , the court attempts to distinguish those cases, as well as implicitly Caulfield , by arguing that "unlike a forensic laboratory analyst, a foreign language interpreter simply converts information from one language to another without adding content," but in contrast, "[a] laboratory analyst must input knowledge and content in order to take a biological sample and generate a report on the sample, including a definitive test result." But this is an illusory distinction, created by conflating what interpreters aspire to do with what they actually do. In the ideal world imagined by the court, interpreters "simply convert[ ] information from one language to another without adding content." In actuality, however, interpreters, like laboratory analysts, often add content and nuance. And also like analysts, interpreters make mistakes.6
Indeed, one can draw a parallel between the translation process and between each step of the chemical-analysis process that the court lists. Interpreters must "take a sample" by listening to what the native speaker is saying. In doing so, interpreters, like analysts, may make "errors" in taking the sample if they mishear a word. See Roseann Dueñas González et al., Fundamentals of Court Interpretation: Theory, Policy, and Practice 576-79 (2d ed. 2012). Next, interpreters must apply their knowledge to the content by determining what English word or phrase most accurately conveys the native speaker's statement. Again, interpreters, like analysts, may make errors in making this determination. Id. at 779 ("[I]nterpreter error is inevitable."). Finally, interpreters must report their final result by stating (in English) what their determination is. Like analysts, interpreters may make errors in this report by saying (either intentionally or inadvertently) something other than what they believe to be the most accurate English translation. Id. at 637.7
The flaws in the language-conduit theory also become apparent if one considers what the outcome of this case would be with two minor factual changes. Suppose that, instead of speaking Spanish, Lopez-Ramos *427spoke English. Assume further that Lopez-Ramos and the officer were in different rooms, with a mutually trusted "conduit" walking back and forth from one room to another, conveying each other's messages. At trial, the State seeks to have the police officer testify that he was told by the conduit that Lopez-Ramos confessed to the crime. Without a doubt, the State would be required to call the conduit in order to admit that testimony. There would be no question that the statement "Lopez-Ramos said, 'We had intercourse with her' " was the conduit's statement, not a statement of Lopez-Ramos, even if the conduit was just conveying what was said without adding content. And Lopez-Ramos would be entitled to challenge the veracity of that statement by cross-examining the conduit.
The court appears to accept as much, not contesting that in such a scenario, the conduit would be a witness against Lopez-Ramos that the State would be required to call if it wished to admit the conveyed statements. But then the court dismisses the hypothetical as unhelpful because "it does not recognize the difference between the function of an interpreter and the function of a witness who is offering testimony against the accused." I am baffled that such a scenario can present a Confrontation Clause issue if Lopez-Ramos and the conduit are speaking English, but that the constitutional problem somehow goes away if Lopez-Ramos speaks Spanish and an interpreter translates into English. Surely, if Lopez-Ramos has the right to confront a translator who only relays statements that are already in English, the need for confrontation increases-not decreases-if the interpreter is required to not only convey the statements, but also undertake the task of translating them from Spanish to English. If the interpreter is, as the court puts it, a "witness who is offering testimony against [Lopez-Ramos]" when conveying English statements, most certainly he retains that role when conveying Spanish statements.
Of course, the court points out that Lopez-Ramos never challenged the adequacy or accuracy of his translated statements prior to trial. But this is of no moment because the accuracy of the translation is irrelevant under the Confrontation Clause. Indeed, the dissent in Melendez-Diaz raised essentially the same argument-that "[w]here ... the defendant does not even dispute the accuracy of the analyst's work, confrontation adds nothing." 557 U.S. at 340, 129 S.Ct. 2527 (Kennedy, J., dissenting). But the Court rejected this argument, reiterating Crawford 's holding that regardless of whether there are specific challenges to the veracity of an expert's analysis, "the Constitution guarantees " the defendant the right to test the analysis " 'in the crucible of cross-examination.' " Id. at 317-18, 129 S.Ct. 2527 (majority opinion) (emphasis added) (quoting Crawford , 541 U.S. at 61, 124 S.Ct. 1354 ). The same holds true here; regardless of whether Lopez-Ramos challenged the accuracy of the interpreter's translations, the Constitution guarantees him the right to test them via cross-examination.8
Moreover, it is the State's obligation to ensure a fair trial and not Lopez-Ramos's obligation to affirmatively challenge the translations. See *428State v. Kindem , 338 N.W.2d 9, 15 (Minn. 1983) ("[T]he state's obligation was to prove its case in a fair way ...."). Indeed, in Melendez-Diaz , the state raised precisely the same argument, that the Court "should find no Confrontation Clause violation ... because [the defendant] had the ability to subpoena the analysts." 557 U.S. at 324, 129 S.Ct. 2527. The Court rejected this argument, reasoning that "the Confrontation Clause imposes a burden on the prosecution to present its witnesses, not on the defendant to bring those adverse witnesses into court." Id. The Supreme Court concluded that the value of the Confrontation Clause is not replaced by a system that permits the state to present its evidence via out-of-court accusations and then wait for the defendant to subpoena the declarants if he so chooses. Id. at 324-25, 129 S.Ct. 2527.
The court also relies on Minn. Stat. §§ 611.30 -.34 (2018) and our Code of Professional Responsibility for Interpreters in the Minnesota State Court System as evidence of the protections native speakers receive against incorrect interpretation. Again, those protections are irrelevant to a Confrontation Clause analysis. But even if they were relevant, Lopez-Ramos did not receive those protections. First, as the court notes, qualified interpreters must "take an oath[ ] to make to the best of the interpreter's skill and judgment a true interpretation." Minn. Stat. § 611.33, subd. 2. The interpreter in this case took no such oath. After being connected with the interpreter, the interrogating officer asked the interpreter to introduce himself to Lopez-Ramos, read Lopez-Ramos his Miranda rights in Spanish (even though his first language was Mam), after which the interpreter immediately began translating interrogation questions and answers.
The court's reliance on our Code of Professional Responsibility for Interpreters is similarly misplaced. Our Code only applies to "persons, agencies and organizations who administer, supervise, use, or deliver interpreting services within the Minnesota state court system." Code of Prof'l Responsibility for Interpreters in the Minn. State Court Sys., Applicability. But the interpreter here was not providing interpreting services within the Minnesota state court system. He was providing them to a police department. Nothing in the record suggests that the department had a code of conduct applicable to its interpreters, and even if there was such a code, there is no indication in the record that the interpreter in this case was ever informed of it.9
Finally, the court suggests that holding that Lopez-Ramos had the right to confront the interpreter in court would imply that, whenever a transcript of a past proceeding was used, the defendant would have the right to confront the court reporter who prepared the transcript. The court is wrong. Only testimonial hearsay is implicated by the Confrontation Clause. See Crawford , 541 U.S. at 68, 124 S.Ct. 1354 ("Where nontestimonial hearsay is at issue, it is wholly consistent with the Framers' design to afford the States flexibility in their development of hearsay law ...."). A necessary requirement for a statement to be testimonial is that "the 'primary purpose' of the [statement] was to 'creat[e] an out-of-court substitute for trial testimony.' "
*429Clark , --- U.S. ----, 135 S. Ct. at 2180 (quoting Michigan v. Bryant , 562 U.S. 344, 358, 131 S.Ct. 1143, 179 L.Ed.2d 93 (2011) ). But the primary purpose of court reporters is not to create a substitute for trial testimony; it is to create an official record of the proceedings to aid in the administration of justice. Therefore, the holding I would reach does not implicate court reporters.
Because: (1) the interpreter is the declarant of the statement, "Lopez-Ramos said, 'We had intercourse with her' "; (2) that statement was testimonial; and (3) the State did not call the interpreter at trial, the district court erred when it denied Lopez-Ramos's motion to suppress the statement. Obviously, the district court's decision to admit the statement was not harmless beyond a reasonable doubt. Accordingly, I would reverse Lopez-Ramos's conviction and remand for a new trial. At such a trial, the State could either offer the live testimony of the AT&T interpreter, or have a different interpreter in the courtroom translate Lopez-Ramos's recorded statement.
The State argues that there has been no determination that the statements in question were testimonial. But "[s]tatements taken by police officers in the course of interrogations are ... testimonial under even a narrow standard." Crawford , 541 U.S. at 52, 124 S.Ct. 1354.
Of course, the interpreter did not literally say "Lopez-Ramos said" before rendering each translation. But it was unquestionably implicit in the interpreter's statements. As the State points out in its brief, the interpreter was certainly not claiming to have had intercourse with the victim; the interpreter was saying that Lopez-Ramos said he had intercourse with her.
Crawford identifies one of "[t]he most notorious instances" of evidence the clause was intended to guard against as the 1603 trial of Sir Walter Raleigh. Crawford , 541 U.S. at 44, 124 S.Ct. 1354. In that trial, the attorney general read onto the record a declaration by Lord Cobham claiming that "Raleigh and he [were] to meet to confer about the distribution of ... money" obtained from the King of Spain for the furtherance of sedition. See Trial of Sir Walter Raleigh , 1 Jardine's Crim. Tr. 400, 411, 415 (1832). Raleigh demanded "to have [his] accuser brought here face to face to speak," citing statutes that required that, in treason cases, "accusers must be brought in person before the party accused at his arraignment, if they be living." Id. at 418. The court refused Raleigh's request, noting the statutes in question had been repealed because "they were found to be inconvenient." Id. at 420. Raleigh was subsequently found guilty and sentenced to death. See id. at 449, 451.
Of course, Lopez-Ramos was not and cannot be sentenced to death, but the exact same circumstances are present here: a third-party relayed an alleged confession to investigators, the State presented the alleged confession to the jury at Lopez-Ramos's trial, Lopez-Ramos demanded that the State call the interpreter, the judge refused, and Lopez-Ramos was convicted based primarily on the alleged confession.
To this end, the court is incorrect in stating that my disagreement is "premised primarily on the assumption that 'interpreters make mistakes.' " My discussion of interpreter mistakes throughout this opinion is merely intended to rebut the presumption-upon which the court bases much of its analysis-that an interpreter is some sort of infallible translating machine that does not make judgment calls when rendering its translation. But even if the interpreter's reliability was unquestionable, Crawford makes clear that Lopez-Ramos would still have the procedural right to cross-examine the interpreter. See Crawford , 541 U.S. at 61, 124 S.Ct. 1354 ; see also infra at 426-27.
Admitting the report also would not have violated the Confrontation Clause if the authoring analyst was unavailable to testify and the defendant had a pretrial opportunity to cross-examine the analyst. Bullcoming , 564 U.S. at 652, 131 S.Ct. 2705.
One such example can be found in this very case. At trial, Lopez-Ramos was asked if he remembered telling the investigating officer "that [he] had some incident with [the victim]." The court-appointed interpreter initially translated his response as "Ah, I don't remember it because I was drunk. I was not within my five senses. I was ask-asked-being asked questions that I was not understanding." However, the interpreter subsequently corrected his translation to "I was under fear."
Even this description of the interpretation process is a gross over-simplification. Although no one model of the interpretation process has gained universal acceptance, all reflect a complicated, multi-step process that is considerably more nuanced than the court's facile description of "simply convert[ing] information from one language to another without adding content." See, e.g. , González et al., supra , at 817-19 (laying out proposed models of the interpretation process).
The court also states that the State did not verify the interpreter's identification and physical location "because Lopez-Ramos never challenged the adequacy or accuracy of the translation." In addition to the accuracy of the translation being irrelevant, there is also no evidence to support this causal connection. The State admitted it did not verify the identity or location of the interpreter, but it never offered any explanation to the district court for its failure to do so.
Indeed, the paucity of information about this interpreter is startling. The State offered no evidence of who he was beyond his first name and interpreter-identification number, no evidence of where he was located, and most importantly, no evidence of his training or experience. Accordingly, even if the State had called the interpreter, it is likely that the interpreter would not have been allowed to testify under Minn. R. Evid. 702 unless the State laid more foundation as to his qualifications. See Minn. R. Evid. 604 ("An interpreter is subject to the provisions of these rules relating to qualification as an expert ....").