Before SMITH, WIENER, and ELROD, Circuit Judges. JENNIFER WALKER ELROD, Circuit Judge: Alexis Aguilar-Alonzo pleaded guilty to aiding and abetting the possession with intent to distribute marijuana. The district court imposed a two-level enhancement to Aguilar-Alonzo’s base offense level under U.S.S.G. § 2D1.1(b)(15)(A) (2016) for “us[ing] fear, impulse, friendship, affection, or some combination thereof to involve another individual” in the offense. Aguilar-Alonzo appeals, contending that the evidence does not support the two- level enhancement and that the district court clearly erred in applying the enhancement. We agree and VACATE and REMAND. I. Aguilar-Alonzo, his girlfriend Yudilet Chavez-Hernandez, and eight others were charged with aiding and abetting the possession with intent to distribute more than 100 kilograms of marijuana but less than 1000 kilograms in violation of 21 U.S.C. § 841(a)(1) and 18 U.S.C. § 2. Aguilar-Alonzo pleaded guilty without a plea agreement. After her arrest, Chavez-Hernandez told investigators that “she had been dating Aguilar-Alonzo for approximately one year and . . . knew [he] was involved in drug trafficking,” but that she had never participated in “picking up a load of narcotics” herself prior to this offense. This time, Aguilar-Alonzo requested that she accompany him “to pick up the [marijuana].” She told the investigators that she agreed to participate “out of fear he would break up with her.” The presentence investigation report (PSR) calculated a base offense level of 24 under the 2016 Sentencing Guidelines §§ 2D1.1(a)(5) and (c)(8). It recommended a two-level enhancement under § 3B 1.1(c) of the Guidelines because Aguilar-Alonzo “was an organizer, leader, manager, or supervisor.” Because of this two-level enhancement for leadership, the PSR recommended an additional two-level enhancement for two independent reasons. First, Aguilar-Alonzo “us[ed] friendship or affection to involve [Chavez-Hernandez] in the illegal transport of controlled substance[s], [Chavez-Hernandez] received little or no compensation [from] the transport of the controlled substances, and [Chavez-Hernandez] had minimal knowledge of the scope and structure of the enterprise.” See U.S.S.G. § 2D1.1(b)(15)(A) (2016).[1] Second, Aguilar-Alonzo, “knowing that [Chavez-Hernandez] was pregnant, involved [her] in the offense.” See § 2D1.1(b)(15)(B)(iii). With a three-level reduction for acceptance of responsibility under § 3E1.1(a) and (b), Aguilar-Alonzo’s total offense level was 25. Based on the offense level of 25 and a criminal history category of II, the Guidelines range was 63 to 78 months of imprisonment. Aguilar-Alonzo objected to the two-level enhancement for use of affection, asserting that “Chavez-Hernandez had more [than minimal] knowledge” of the scope and structure of the criminal enterprise and that merely being “in a dating relationship does not trigger” that enhancement. The probation officer asserted that the requirements for the enhancement were satisfied. Regarding use of affection, the probation officer noted that (1) “Chavez-[Hernandez] was engaged in a romantic relationship with [Aguilar-Alonzo],” (2) Aguilar-Alonzo “was aware Chavez-[Hernandez] was pregnant with his child at the time of the instant offense,” and (3) “Chavez- [Hernandez] agreed to help [Aguilar-Alonzo] as she feared [he] would terminate his relationship with her.” At sentencing, Aguilar-Alonzo contended that nothing in the record indicated that he “made any kind of action or said any words” suggesting that he would end the relationship if she refused to participate in the offense. Aguilar-Alonzo asserted that, even though she may have felt that way, he “ha[d] to make some sort of actual action or words” to be eligible for the enhancement. In addition, Aguilar-Alonzo contended that Chavez-Hernandez had more than minimal knowledge of the scope and structure of the enterprise: “[S]he knew [that] he was involved in drug trafficking,” she saw him pick up and unload drugs in the past, and she then agreed to participate in the offense following his request. Aguilar-Alonzo also objected to the second independent basis for the two- level enhancement, claiming it was unwarranted under § 2D1.1(b)(15)(B) because “it [was] unclear whether [Aguilar-Alonzo] knew at the time of the offense that [Chavez-Hernandez] was pregnant.” Aguilar-Alonzo asserted that “he didn’t find out until she was arrested,” at which point she called from jail to tell him “I think I just found out I’m pregnant,” and that he had a letter from her to him explaining “[she] wish[ed she] would have told [him] that [she] was pregnant.” The prosecutor conceded that the government did not have evidence proving that Aguilar-Alonzo knew of Chavez-Hernandez’s pregnancy at the time of the offense. The district court explicitly found that the facts did not establish whether Aguilar-Alonzo knew, at the relevant time of the offense, that Chavez- Hernandez was pregnant and declined to sustain the enhancement under § 2D1.1(b)(15)(B). Nevertheless, the district court determined that the two- level enhancement was warranted under § 2D1.1(b)(15)(A) because it was “apparent from the facts . . . that [Aguilar-Alonzo] used fear, impulse, friendship, affection, or some combination thereof to involve [Chavez- Hernandez]” in the offense. The court overruled Aguilar-Alonzo’s objection to the enhancement for a leadership role. With this two-level enhancement for the use of affection, the total offense level was 25, resulting in a Guidelines range of 63 to 78 months. The district court declined to depart from the Guidelines recommendation and sentenced Aguilar-Alonzo to 70 months of imprisonment and 5 years of supervised release. On appeal, Aguilar-Alonzo challenges only the two-level enhancement under § 2D1.1(b)(15)(A), asserting that the district court clearly erred in concluding that he “used fear, impulse, friendship, affection, or some combination thereof to involve” Chavez-Hernandez in the offense because “[t]he [G]uideline’s plain language requires active employment of affection to induce another to participate in the offense,” but “[n]o evidence supported a finding that Aguilar[-Alonzo] actively induced Chavez[-Hernandez's] participation by playing on affection.” Aguilar-Alonzo further argues that this error prejudiced him because, without the erroneous two-level enhancement, his Guidelines range would have been 60 to 63[2] months, and there is no indication that the district court would have imposed the same 70-month sentence if it had considered the lower range. II. “To preserve error, an objection must be sufficiently specific to alert the district court to the nature of the alleged error and to provide an opportunity for correction.” United States v. Neal, 578 F.3d 270, 272 (5th Cir. 2009). At the sentencing hearing, Aguilar-Alonzo argued that he did not take any action or make any statement that would have led Chavez-Hernandez to believe that he would end the relationship if she refused to participate in the offense. His argument was sufficiently specific to alert the district court to the nature of the issue raised on appeal. See Neal, 578 F.3d at 272. “We review the district court’s interpretation and application of the Guidelines de novo, and its factual findings for clear error.” United States v. Zuniga, 720 F.3d 587, 590 (5th Cir. 2013). “A finding is ‘clearly erroneous’ when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.” United States v. U.S. Gypsum Co., 333 U.S. 364, 395 (1948). The proponent of an adjustment to the defendant’s base offense level bears the burden of establishing the factual predicate “by a preponderance of the relevant and sufficiently reliable evidence.” United States v. Richardson, 781 F.3d 237, 249 (5th Cir. 2015) (quoting United States v. Alfaro, 919 F.2d 962, 965 (5th Cir. 1990)). “Where . . . the district court commits a significant procedural error such as miscalculating the Guidelines range, we must vacate the sentence unless the error did not affect the selection of the sentence imposed.” Neal, 578 F.3d at 274. “The proponent of the sentence has the burden of establishing that the error was harmless.” Id. III. On appeal, Aguilar-Alonzo only challenges the district court’s two-level enhancement under U.S.S.G. § 2D1.1(b)(15)(A) (2016), which reads: If the defendant receives an adjustment under § 3B1.1 . . . and[:] (i) The defendant used fear, impulse, friendship, affection, or some combination thereof to involve another individual in the illegal purchase, sale, transport, or storage of controlled substances, (ii) the individual received little or no compensation from the illegal purchase, sale, transport, or storage of controlled substances, and (iii) the individual had minimal knowledge of the scope and structure of the enterprise . . . increase by 2 levels. The district court concluded that all three parts of this Guideline were met and applied the two-level increase. Aguilar-Alonzo specifically challenges the district court’s finding that he “used affection to involve . . . [Chavez- Hernandez] in the” offense under part (i) of the Guideline quoted above. Aguilar-Alonzo contends that it was clear error for the district court to apply this enhancement and that the error prejudiced him by resulting in a greater sentence than the court would have otherwise imposed. In order to resolve this case, we must determine the proper interpretation of the term “used” in the Guideline, ascertain whether it was clear error for the district court to apply the enhancement in light of our interpretation of the word, and determine whether any error was harmless. A. Aguilar-Alonzo and the government dispute both the proper interpretation of the verb “used” in the context of the Guidelines and the application of that interpretation to these facts. Aguilar-Alonzo contends that “[t]he plain language of [§ 2D1.1(b)(15)(A) requires] that a defendant must actively employ or play upon affection to induce another to commit [an] offense.” Alternatively, the government argues that “used” encompasses a broader definition, including “passively exploiting another’s affection.” We agree with Aguilar-Alonzo’s interpretation of “used” under the Guidelines. We interpret the Sentencing Guidelines according to “the typical rules of statutory interpretation,” “[s]tarting with the text of the relevant guideline[.]” United States v. Stanford, 883 F.3d 500, 511 (5th Cir. 2018). “If the language is unambiguous, and does not lead to an ‘absurd result,’ the court’s inquiry begins and ends with the plain meaning of that language.” Id. (quoting United States v. Koss, 812 F.3d 460, 473 (5th Cir. 2016)). “When a word is not defined by statute, we normally construe it in accord with its ordinary or natural meaning.” Smith v. United States, 508 U.S. 223, 228 (1993). In a variety of criminal statutory contexts, we have consistently interpreted the ordinary and natural meaning of the verb “use” to require active employment of something, as has the Supreme Court. For example, the Supreme Court explained that proof of “use” in “[18 U.S.C.] § 924(c)(1) requires evidence sufficient to show an active employment of the firearm by the defendant, a use that makes the firearm an operative factor in relation to the predicate offense.” Bailey v. United States, 516 U.S. 137, 143 (1995), superseded by statute on other grounds as stated in Welch v. United States, 136 S. Ct. 1257 (2016). In Jones v. United States, the Supreme Court determined that 18 U.S.C. § 844′s requirement “that [a] building be ‘used’ in an activity affecting commerce” was “most sensibly read to mean active employment for commercial purposes, and not merely a passive, passing, or past connection to commerce.” 529 U.S. 848, 855 (2000). This interpretation of “use” has been reaffirmed many times.[3] The proper interpretation of “used” under the Guideline requires active employment of affection in return for involvement in the offense, and that interpretation is consistent with the previous interpretations discussed above. The only other circuit court of appeals decision to address the “used . . . affection” provision of § 2D1.1(b)(15)(A) is consistent with our interpretation of the word. The Third Circuit, in United States v. Best, affirmed the district court’s two-level enhancement for Best’s use of affection to involve his girlfriend in a drug trafficking offense. 639 F. App’x 848, 853-54 (3rd Cir. 2016). At trial, “Best admitted . . . that he convinced [his girlfriend] to steal pills and blank prescriptions by telling her that he loved her[,]” “ that she was never in trouble before meeting him[,]” and “ that if not for him, [his girlfriend] ‘wouldn’t be here,’ facing criminal charges.” Id. at 853. In that case, Best was held subject to the two-level enhancement after he explicitly told his girlfriend that he loved her for the purpose of getting her to steal pills for him. Best supports our interpretation of “use” under § 2D1.1(b)(15)(A) (2016) as requiring active employment of affection on the part of the defendant. B. We next address whether it was clear error for the district court to apply the two-level enhancement under the Guidelines. “A finding is ‘clearly erroneous’ when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.” U.S. Gypsum Co., 333 U.S. at 395. As the proponent of the adjustment to Aguilar-Alonzo’s base offense level, the government bears the burden of establishing the factual predicate of that adjustment “by a preponderance of the relevant and sufficiently reliable evidence.” Richardson, 781 F.3d at 249 (citation omitted). The government fails to meet this burden. After reviewing the evidence, we are left with the definite and firm conviction that the district court committed error by applying the two-level enhancement for use of affection. Under our interpretation of the verb “used” in § 2D1.1(b)(15)(A), the evidence in the record does not support a two-level enhancement to Aguilar- Alonzo’s offense level because he did not actively employ or play upon affection to induce involvement by Chavez-Hernandez in the offense. The evidence consists of two primary facts, neither of which support a finding that Aguilar- Alonzo actively used affection to induce Chavez-Hernandez’s involvement in the offense. First, Aguilar-Alonzo asked Chavez-Hernandez to participate in the offense. Second, Chavez-Hernandez asserted that she agreed to participate “out of fear [Aguilar-Alonzo] would break up with her” if she did not participate. The government essentially argues that, just because Aguilar-Alonzo had the opportunity to use affection to recruit his girlfriend, we should assume that he did, despite the lack of evidentiary support. The government attempts to support this result by emphasizing that Chavez-Hernandez had no previous involvement in criminal activity, had little to gain financially from the crime, and had a limited role in the crime. However, the government offered no support for the proposition that Aguilar-Alonzo did something or said something to cajole or force her to help him. Her fear that Aguilar-Alonzo would break up with her was simply her own subjective assessment of the situation. We see no indication that Aguilar-Alonzo actually did or said anything about their relationship when he asked her to help him. There is nothing in the record whatsoever that indicates Aguilar-Alonzo even knew of Chavez-Hernandez’s subjective fear that he might break up with her but for her involvement in the offense. “No evidence suggests that [Aguilar-Alonzo] cajoled, wheedled, threatened a breakup, or otherwise actively played on [Chavez-Hernandez's] affection or emotions about the relationship.” Perhaps, if Aguilar-Alonzo knew of this fear prior to his request for Chavez-Hernandez’s involvement, a finding that he used affection may be appropriate. That is not the case here. Chavez-Hernandez’s own subjective assessment cannot be solely relied on as evidence that Aguilar-Alonzo used affection to induce her involvement. It cannot be that the mere existence of a romantic, familial, or any other kind of pre-existing relationship would support a two-level enhancement under § 2D1.1(b)(15)(A) when one person, who fits the other criteria under the Guidelines, recruits another to assist in the offense. Aside from the existence of the relationship and Chavez-Hernandez’s subjective fear of a breakup, there is nothing in the record that could even possibly relate to use of affection. Neither does Chavez-Hernandez’s pregnancy support the conclusion that Aguilar-Alonzo actively used affection to involve her in the offense. The district court explicitly found that the facts did not establish whether Aguilar-Alonzo knew, at the relevant time of the offense, that Chavez-Hernandez was pregnant. The fact that Aguilar-Alonzo and Chavez-Hernandez at some point had a sexual relationship, without more, is not enough to establish that Aguilar-Alonzo “actively employed” affection to involve her in the offense. The Third Circuit’s opinion in Best well illustrates why the facts of the instant case do not support the same two-level use of affection enhancement. See 639 F. App’x at 853-84. Unlike in Best, no record evidence demonstrates that Aguilar-Alonzo actively employed affection to induce involvement by Chavez-Hernandez. In Best, the defendant told his girlfriend that he loved her in order to convince her to steal pills and blank prescriptions. Id. at 853. In the present case, Aguilar-Alonzo simply asked Chavez-Hernandez for her help. We see nothing in the record that indicates Aguilar-Alonzo actively employed the continuation of their relationship in return for Chavez-Hernandez’s involvement in the offense. Aside from adopting the PSR, which stated that Chavez-Hernandez agreed to participate in the offense out of fear that Aguilar- Alonzo would break up with her—an uncommunicated, subjective feeling—the district court did not identify how Aguilar-Alonzo actively employed affection to induce involvement by Chavez-Hernandez in the offense. We therefore have a firm and definite conviction that an error has been made by the district court in applying the two-level enhancement for use of affection. C. “Where . . . the district court commits a significant procedural error such as miscalculating the Guidelines range, we must vacate the sentence unless the error did not affect the selection of the sentence imposed.” Neal, 578 F.3d at 274. “The proponent of the sentence has the burden of establishing that the error was harmless.” Id. As the proponent of the sentence, the government also bears the burden of establishing that the error of applying the two-level enhancement was harmless to Aguilar-Alonzo. The government fails to address harmlessness of the error in its brief. Accordingly, any argument as to harmlessness is forfeited. However, even if we were to consider the issue, the district court’s error was not harmless. The district court determined that Aguilar-Alonzo’s total offense level was 25. With a criminal history category of II, this offense level yielded a Guidelines range of 63 to 78 months of imprisonment. Without the two-level enhancement under § 2D1.1(b)(15)(A), Aguilar-Alonzo’s total offense level would have been 23, yielding a Guidelines range of 60-63 months of imprisonment.[4] The district court’s sentence of 70 months of imprisonment is entirely outside the appropriate Guidelines range. The record lacks evidence that the court intended to increase Aguilar-Alonzo’s sentence beyond the Guidelines range. In fact, the district judge, at the sentencing hearing, stated that he intended to remain within the recommended Guidelines range. Consequently, the district court’s error in imposing the § 2D1.1(b)(15)(A) enhancement affected the court’s selection of the sentence it imposed. Accordingly, the error was not harmless, as it affected Aguilar-Alonzo’s “substantial rights.” See Fed. R. Crim. P. 52(a). IV. For the foregoing reasons, we VACATE Aguilar-Alonzo’s sentence and REMAND for further proceedings consistent with this opinion. JERRY E. SMITH, Circuit Judge, dissenting: The majority’s rejection of Judge Counts’s factual finding is error under the correct standard of review. It’s a puzzlement that the majority doesn’t even mention that proper standard, given that both of the judges in the majority are well aware of it. Because Aguilar-Alonzo can’t prevail under that standard or under the equally valid standard that the majority employs, I respectfully dissent. At sentencing, “[a] factual finding is not clearly erroneous as long as it is plausible in light of the record as a whole.” United States v. Dinh, 920 F.3d 307, 310 (5th Cir. 2019) (Elrod, J.).[5] Although the government carefully and accurately recites this standard, Aguilar-Alonzo’s counsel, an experienced former federal public defender, does not, presumably because it’s impossible for him to prevail in light of it. The long and short of it is that, at least at sentencing, a defendant can’t establish clear error without showing, as a necessary condition, that the district judge’s factual finding is “implausible.” Black’s Law Dictionary defines “plausible” as “[c]onceivably true,” “possibly correct or even likely,” “reasonable.”[6] That standard should be read in tandem with the standard that the majority accurately describes: “A finding is ‘clearly erroneous’ when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.”[7]Aguilar-Alonzo’s challenge fails under either standard, applied separately, and even more so under both. The facts are that the defendant was twenty-one; his girlfriend was nineteen. He lived in a travel trailer behind her parents’ house. They had been dating at least a year and had had sexual intercourse. She knew she was pregnant; the district judge declined to find, one way or the other, whether the defendant knew that.[8] But it is undisputed that he remembered the carnal knowledge, and no evidence suggests that the relationship was still anything but close. As the majority explains, the defendant “requested that she accompany him ‘to pick up the [marijuana].’ She told investigators that she agreed to participate ‘out of fear he would break up with her.’” All agree that the sole question is whether “the defendant used fear . . . [or] affection[9] . . . to involve” his girlfriend in the illegal activity. To answer that, we don’t consider whether, on the best view of the evidence, the defendant did or didn’t use affection. That is hardly our role on appeal. To the contrary, we ask only whether it is plausible—that is, “conceivably true,” “possibly correct,” or “reasonable”[10]—that the defendant did so. If it is any of those three, we must affirm. I won’t waste much space to recount that at federal sentencing, the evidentiary rules are considerably relaxed. For example, the sentencing judge can consider hearsay and a wide range of circumstantial evidence, and the defendant has no right to confront witnesses or conduct cross-examination. The judge can make inferences from the evidence he hears and can apply reason and experience. And he can rely on the findings in the presentence report. At a minimum, the judge knew that the couple was in a one-year romantic relationship that wasn’t on the wane. They had had sexual relations, and the girlfriend knew she was pregnant. She was only nineteen. She told the probation officer that she had “agreed to [defendant's] request [to assist him] out of fear he would break up with her.” From these bare facts, do we know, for a certainty, that asking his girlfriend to help with the crime was a threat to break up with her, leaving her with an unborn child and without a companion, if she refused? We do not, and neither did the district judge. But what we do know (for a certainty) is that she felt that way. And nothing even remotely suggests that that feeling was unfounded. Her fear that the defendant would end the relationship is based on all the circumstances of which she was aware. That is her reading of their communications. It is unrefutable circumstantial evidence of the direct evidence that is absent from this record: that the defendant made a threat to end the romance if his girlfriend didn’t assist.[11] To prevail on appeal, the defendant must convince this court that the district judge’s inference from the circumstantial evidence is an hallucination that isn’t “conceivably true,” “possibly correct,” or “reasonable.” To the contrary, there is enough reason to believe that the defendant, knowing this young woman was enamored, chose her, rather than someone else, to supply the backup help he needed to commit the crime. By deciding as it has, the panel majority necessarily implies that Judge David Counts engaged in a flight of fancy by entertaining an explanation that is outside the bounds of ordinary reason and human experience and that invents a story that defies logic and has no basis in the record. Although I’m confident that the majority doesn’t mean to plaster Judge Counts with the accusation that he has lost his senses, that is the necessary implication of the majority’s analysis, which I had hoped it would reconsider. As I’ve explained, the majority announces another, albeit accurate, standard of review: that clear error requires us to reach “the definite and firm conviction that a mistake has been committed.” Importantly, that articulation of the standard approaches the facts in a decidedly different way. For the majority to decide that “a mistake has been committed,” it necessarily declares that the opposite of the district court’s finding is true: that the defendant did not ask his girlfriend to help by employing fear or affection. We can acknowledge, arguendo, that either finding is believable from these facts. In my view, it’s more likely that he played on their relationship than it is that he picked her out randomly, from among the earth’s inhabitants, to request assistance. But what I think isn’t the point. By saying “a mistake has been committed,” the majority imposes a necessary finding of its own: that the defendant didn’t employ affection or fear. That’s the way the “definite and firm conviction” standard works. The district judge found that the facts are “X,” but we say the facts instead are “Y.” The judge found that the traffic light was green, but we, in our appellate wisdom, read the facts differently and find, by definite and firm conviction, that it was red. The situation, however, is that we don’t and can’t know for sure, and that’s why the rules and caselaw defer to the district judge to make that decision, protected by steep deference on appeal after the judge has heard witnesses and examined the case in open court. This defendant loses under either articulation of the standard of review. Under the standard that the majority ignores, it is error to suggest that Judge Counts’s finding of fear or affection is pure fancy in the sense that it is not “conceivably true,” that it is not “possibly correct,” and that it is not “reasonable.” And under the standard that the majority acknowledges, it is error to say not only that Judge Counts’s finding is unsupported by the record but also that the opposite (i.e., no use of fear or affection) is affirmatively true. The majority confronts none of this, content only to emphasize what it sees as a dearth of facts, ignoring the relaxed evidentiary procedures and the broad use of circumstantial evidence and inferences at sentencing.[12] ***** At bottom, the decision to reject Judge Counts’s factual finding does grave disservice to the sentencing regime. It undermines the carefully structured assignment of factfinding to an independent Article III judge. The dynamics of federal sentencing are unique. Especially in venues such as Pecos, Texas, where this case is sited, scores of criminal defendants queue up for sentencing hearings on a routine day. The district judge carefully studies the presentence report, hears from the attorneys, and takes allocution. He or she then makes factual findings and announces a sentence. The evidentiary standards are relaxed for good reason, given the caseload. That is not to say that the proceedings are inadequate. Much to the contrary. They are the optimum way to make fair and accurate factual decisions and rulings of law in a relatively short period of time. Here, the probation office prepared a comprehensive presentence report, with addenda, totaling more than twenty pages. Judge Counts conducted a twenty-minute sentencing hearing, with a transcript of seventeen pages, and made detailed findings, giving ample time for participation. The crush of the sentencing docket can be accommodated by dedicated judges and staff who read and apply the law to the facts. The procedures used to present and incorporate those facts get the job done well. Nitpicking by circuit judges who (with the best of intentions) misunderstand or misapply the standards of review do not. I respectfully dissent.
