Under the Immigration and Nationality Act (“INA”), “[ajriy alien who is convicted of an aggravated felony at any time after admission” is eligible for removal. 8 U.S.C. § 1227(a)(2)(A)(iii). One type of aggravated felony under the INA is “a theft- offense (including receipt of stolen property) ... for which the; term of imprisonment [is] at least one year.” 1⅛ § 1101(a)(43)(G). In finding petitioner Evandro De Lima eligible for removal, the Board of Immigration Appeals (“BIA”) concluded that third-degree larceny under Connecticut law, Conn. Gen. Stat. § 53a-124, is one such offense. For the following reasons, we uphold that finding.
I.
De Lima is a native and citizen of Brazil. He became a lawful permanent resident of the United States in 2011, three years before he was convicted of third-degree larceny under section 53a-124 of the Connecticut General Statutes.1 In March 2015, removal proceedings commenced against De Lima on the basis that his conviction was for a “theft offense” within the meaning of 8 U.S.C. § 1101(a)(43)(G) and was therefore an “aggravated felony” that rendered him eligible for removal. Id. § 1227(a)(2) (A) (iii).
*262Section 53a-119 of the Connecticut General Statutes provides that a person commits larceny “when, with intent to deprive another of property or to appropriate the same to himself or a third person, he wrongfully takes, obtains or withholds such property from an owner.” Conn. Gen. Stat. § 53a-119. Larceny “includes, but is not limited to,” things like “embezzlement,” id. § 53a-119(l); “[o]btaining property by false pretenses,” id. § 53a-119(2); “[o]btaining property by false promise,” id. § 53a-119(3); “defrauding a public community,” id. § 53a-119(6); “theft of services,” id. § 53a-119(7); “library theft,” including “mutilat[ing] a book or other archival library materials ..." so as to render it unusable or reduce its value,” id. § 53a-119(12); “theft of utility service,” including “wireless radio communications,” id. § 53a-119(15); and “air bag fraud,” whereby a person fraudulently “obtains property from such other person or a third person by knowingly selling, installing or reinstalling any object, including any counterfeit air bag or nonfunctional air bag ... in lieu of an air bag that was designed in accordance with federal safety requirements,” id. § 53a-119(16).
Larceny comes in several degrees under Connecticut law. To establish the degree relevant here (third-degree larceny), the state must prove one of the following additional factors: (a) the offender stole a motor vehicle worth ten thousand dollars or less; (b) “the value of the property or service exceeds two thousand dollars”; (c) “the property consists of a public record, writing or instrument kept, held or deposited according to law with or in the keeping of any public office or public servant”; or (d) “the property consists of a sample, culture, microorganism, specimen, record, recording, document, drawing or any other article, material, device or substance which constitutes, represents, evidences, reflects or records a secret scientific or technical process, invention or formula or any phase or part thereof,” as “secret” is defined therein. Id § 53a-124.
In an oral decision on April 10, 2015, an immigration judge found De Lima removable and ordered him removed. De Lima timely appealed to the BIA. Before the Board, he argued that section 53a-124 is broader than the definition of a “theft offense” under the INA, and therefore cannot categorically count as an aggravated felony. Specifically, he claimed that the federal definition of a generic “theft offense” requires permanent intent to deprive another of property, and the Connecticut statute does not, both because it criminalizes theft of property without the intent to permanently deprive the owner of the property, and because it criminalizes theft of services. Therefore, reasoned De Lima, it is possible for a person to be convicted under section 53a-124 for something that would not be considered a “theft offense”, under the federal definition.
The BIA rejected De Lima’s claims and dismissed his appeal. De Lima then timely petitioned our court for review. We review purely legal challenges like those raised here de novo, though we accord deference to the BIA’s “reasonable interpretation of statutes and regulations falling within its bailiwick.” Segran v. Mukasey, 511 F.3d 1, 5 (1st Cir. 2007).
II.
Because the INA’s list of aggravated felonies, see 8 U.S.C. § 1101(a)(43), does not perfectly correspond to state criminal codes, “the BIA and courts of appeal must often ascertain whether a particular state law fits within the enumerated aggravated felonies.” Lecky v. Holder, 723 F.3d 1, 4 (1st Cir. 2013). To do so, we apply the so-called “categorical approach,” which “looks to the statutory definition of *263the offense of conviction, not to the particulars of the alien’s behavior.” Mellouli v. Lynch, — U.S. -, 135 S.Ct. 1980, 1986, 192 L.Ed.2d 60 (2015); see Moncrieffe v. Holder, 569 U.S. 184, 133 S.Ct. 1678, 1684, 185 L.Ed.2d 727 (2013). In substance, we identify the elements of the state offense for which the person was previously convicted; we identify, to a reasonable possibility, the minimum conduct that the state would have deemed to have satisfied those elements; and then we ask whether that conduct would also satisfy one of the INA’s listed “generic” aggravated felonies. Moncrieffe, 133 S.Ct. at 1684-85; see Esquivel-Quintana v. Sessions, — U.S. -, 137 S.Ct. 1562, 1567, 198 L.Ed.2d 22 (2017).
Consistent with that approach, De Lima advances three arguments for finding that his Connecticut conviction is not a conviction for a “theft offense” because the range of conduct sufficient to sustain a conviction for third-degree larceny under Connecticut law is broader than that which constitutes a “theft offense” under the INA. We address each argument in turn.
A.
De Lima argues, first, that section 53a-124 is overbroad because it imposes liability for takings of property even by one who does not intend to deprive another permanently of the property, as evidenced by the statute’s imposition of criminal liability for mutilating a library book, replacing a car’s airbags with something else, or intercepting wireless radio communications.
This argument runs into our holding in Lecky. There, the petitioner challenged whether his conviction under Connecticut’s second-degree larceny statute, Conn. Gen. Stat. § 53a-123, could be cited as a conviction for a “theft offense” subjecting him to removal under the INA. Lecky, 723 F.3d at 4. Like its- third-degree larceny statute, Connecticut’s second-degree larceny statute incorporates the definition of larceny contained in section 53a-119. See Conn. Gen. Stat. § 53a-123. Like De Lima, the petitioner in Lecky argued that the generic “theft offense” under the INA does not reach temporary deprivations of property, yet the Connecticut statute does, as evidenced by its inclusion of library theft, airbag fraud, and obtaining wireless radio communications. Lecky, 723 F.3d at 5. We observed that where a person has been convicted for theft of property, “(t]he BIA has made it clear that a theft offense requires the intent to deprive an owner of property rights, but such deprivation need not be permanent nor total.” Id. at 6 (citing Matter of V-Z-S-, 22 I. & N. Dec, 1338, 1345-46 (B.I.A. 2000)). Finding the BIA’s interpretation of the INA reasonable, and noting that the Second Circuit had done so as well in its decisions in Abimbola v. Ashcroft, 378 F.3d 173 (2d Cir. 2004), and Almeida v. Holder, 588 F.3d 778 (2d Cir. 2009), we deferred to the BIA and rejected the petitioner’s argument. See Lecky, 723 F.3d at 5-6 (citing Chevron, U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 843, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984)).
De Lima urges us to find Lecky no longer controlling in light of the Supreme Court’s recent decisions in Moncrieffe and Mellouli. See Holder v. Sessions, 848 F.3d 500, 502 (1st Cir. 2017) (citing, inter alia, United States v. Carter, 752 F.3d 8, 18 n.211 (1st Cir. 2014)) (recounting the exception to stare decisis whereby intervening pronouncements from the Supreme Court undermine an existing panel decision). He argues that these recent cases indicate that the court in Lecky erred by deferring to the BIA’s interpretation of “theft offense” under the INA. Instead, argues De Lima, Moncrieffe and Mellouli show either that the BIA’s decision in V-Z-S- deserves less deference, or that the BIA unreason*264ably interprets the INA anytime it finds that a generic offense is broader than the common-law version of that offense,
De Lima’s argument concerning Lecky and the degree of deference we accord the BIA takes two forms. The first, most clearly articulated in De Lima’s opening brief, is- that post-Lecky Supreme Court decisions suggest that the BIA should default to the common law unless Congress expressly indicates otherwise. But in Taylor v. United States, 495 U.S. 575, 110 S.Ct. 2143, 109 L.Ed.2d 607 (1990), the Supreme Court expressly rejected the argument “that Congress meant to include only a special subclass of [generic offenses]” like “those that would have been [the generic offenses] at common law.” Id. at 598, 110 S.Ct. 2143. There is nothing in Moncrieffe, Mellouli, dr any other intervening Supreme Court decision that suggests that Taylor is no longer good law or does not apply in this context; in fact, Moncrieffe itself relies on Taylor in explicating the categorical approach that applies-to cases like this one, and Mellouli, in turn, relies on Moncrieffe to do the same: See Mellouli, 135 S.Ct. at 1986 (citing Moncrieffe, 133 S.Ct. at 1684-85); Moncrieffe, 133 S.Ct. at 1684-85 (citing Gonzales v. Duenas-Alvarez, 549 U.S. 183, 186, 127 S.Ct. 815, 166 L.Ed.2d 683 (2007) (citing Taylor, 495 U.S. at 599-600, 110 S.Ct. 2143)).
The second form of this argument,, more fully fleshed- out in De Lima’s reply brief and at oral argument, is that Lecky should not control the outcome of this case be-' cause Moncrieffe and Mellouli require that the BIA construe narrowly the ambiguous generic offenses in the INA. In substance* this is an argument that these two cases demonstrate that the rule of lenity must always trump deference in defining precisely what a “theft offense” is under the INA, For three reasons, we disagree.
First, neither Moncrieffe nor Mellouli addresses the subject of the interplay between deference and lenity in construing a provision of the INA. The Court in. Mon-crieffe confronted the BIA’s interpretation of a state statute, not the INA itself. Chevron was therefore not implicated, so no deference was afforded to the BIA’s interpretation .when the Court stated that “ambiguity in criminal statutes referenced by the INA must be. construed in the nonciti-zen’s favor.” Moncrieffe, 133 S.Ct. at 1693 (citing Carachuri-Rosendo v. Holder, 560 U.S. 563, 581, 130 S.Ct. 2577, 177 L.Ed.2d 68 (2010), and Leocal v. Ashcroft, 543 U.S. 1, 11 n.8, 125 S.Ct. 377, 160 L.Ed.2d 271 (2004)). Similarly, lenity did not replace Chevron deference in Mellouli. Rather, deference simply proved to be unwarranted in Mellouli because the BIA’s interpretation of the interplay between the INA, the Comprehensive Drug Abuse Prevention and Control Act of 1970, 21 U.S.C, §§ 801-971, and Kansas state law did not make any sense. See Mellouli, 135 S.Ct. at 1989.
Second, to the extent that De Lima’s argument is that lenity (or some form of it) plays a role in construing provisions of the INA that trigger deportation, or removal, that role is well established and long predates Lecky. See, e.g., Kawashima v. Holder, 565 U.S. 478, 489, 132 S.Ct. 1166, 182 L.Ed.2d 1 (2012); Leocal, 543 U.S. at 11 n.8, 125 S.Ct. 377; INS v. St. Cyr, 533 U.S. 289, 320, 121, S.Ct. 2271, 150 L.Ed.2d 347 (2001); Fong Haw Tan v. Phelan, 333 U.S. 6, 10, 68 S.Ct. 374, 92 L.Ed. 433 (1948) (“We resolve the doubts in favor of [the alien] because deportation is a drastic measure....”). Nothing has changed in that regard post-Lecky, and neither Mel-louli nor Moncrieffe suggests otherwise.
Third, under our case law, even when lenity is potentially applicable, it plays no role “unless there is a grievous ambiguity *265or uncertainty in the language and structure of [a statute], such that even after, a court has seize[d] every thing from which aid can be derived, it is still left with an ambiguous statute.” Soto-Hernandez v. Holder, 729 F.3d 1, 6 (1st Cir. 2013) (quoting Chapman v. United States, 500 U.S. 453, 463, 111 S.Ct. 1919, 114 L.Ed.2d 524 (1991)). The rule therefore “ ‘cannot apply to contravene the BIA’s reasonable interpretation’ of an immigration statute where the agency makes use of ‘ordinary principles of statutory construction.’ ” Garcia v. Sessions, 856 F.3d 27, 41 (1st Cir. 2017) (quoting Soto-Hernandez, 729 F.3d at 6). And this is precisely what the BIA did in V-Z-S- when it decided that “a taking of property constitutes a ‘theft’ [under the INA] whenever there is criminal intent to deprive the owner of the rights and benefits of ownership, even if such deprivation is less than total or permanent.” V-Z-S-, 22 I. & N. Dec. at 1346.
Lecky thus remains good law in this circuit. As such, it forecloses De Lima’s claim that “theft offense” must be construed narrowly to exclude theft committed without intent to permanently deprive. See Lecky, 723 F.3d at 6 (citing V-Z-S-, 22 1. & N. Dec. at 1345-46).
b;
De Lima next argues that Connecticut’s inclusion of theft of services as larceny renders the crime broader than the generic “theft offense” definition in the INA. The government argues that Lecky also disposes of this argument. We are not so sure. On the one hand, our-decision in Lecky expressly approved of the reasoning and holding in Abimbola, in which the Second Circuit found that Connecticut’s third-degree larceny offense .is categorically a “theft offense” under the INA despite its imposition of criminal liability for theft of services. Lecky, 723 F.3d at 6; Abimbola, 378 F.3d at 178-80, In Lecky, we found that Abimbola and Almeida, 588 F.3d at 789 (holding that second-degree larceny under Connecticut law is categorically a “theft offense” aggravated felony under the INA), were “well-reasoned opinions” that were “both on point and persuasive,” and we stated that the petitioner did pot “convince[] us to part ways with those opinions.” Lecky, 723 F.3d at 6;
On the other hand, as mentioned earlier, Lecky involved a challenge to a conviction under a different statute than the statute under which De Lima was convicted. And while that difference was immaterial to our disposition of De Lima’s first argument, it might be material to our consideration of his second. The petitioner in Lecky was convicted under a portion of Connecticut’s second-degree larceny statute that imposes liability for larceny where “the property, regardless of its nature or value, is taken from the person of another.” Conn. Gen. Stat. § 53a-123(a)(3). We noted that because the relevant portion of the statute expressly provided for the taking of property “from the person of another,” some of the examples provided in section 53a-119 could not possibly constitute second-degree larceny under section 53a-123(a)(3). The definition of third-degree larceny under section 53a-124 lacks the precise language upon which we relied in Lecky and contains a subsection that expressly provides for liability for theft of a “service” whose value “exceeds two thousand dollars,” id. § 53a-124(a)(2).2
*266In the end, we need not concern ourselves with either the scope of Lecky’s preclusive effect or with precisely defining the metes and bounds of our holding in that case, because we are unpersuaded by De Lima’s argument that section 53a-124’s imposition of criminal liability for theft of services renders the statute too broad to categorically constitute a “theft offense” aggravated felony under the INA. It is true, as De Lima says, that the traditional common-law definition of theft was limited to property, and that services were not considered property in many common-law jurisdictions. But this is the type of argument rejected in Taylor, where the Supreme Court declined to give the generic term “burglary” its common-law meaning because “that meaning [wa]s obsolete or inconsistent with the statute’s purpose.” 495 U.S. at 594, 110 S.Ct. 2143. The Court determined that “burglary” under the Career Criminals Amendment Act of 1986, 18 U.S.C. § 924(e), was not confined to the term’s traditional common-law meaning, but rather included the broader array of conduct captured within the definition in the Model Penal Code and prohibited by burglary statutes adopted by numerous states at the time the federal statute was passed. Taylor, 495 U.S. at 597-98, 598 n.8, 110 S.Ct. 2143.
So, too, did Congress’s use of the term “theft offense” rather than merely the term “theft” imply an intent to reach more broadly than the singular common-law notion of theft. See Ilchuk v. Att’y Gen. of the U.S., 434 F.3d 618, 622 (3d Cir. 2006) (“[I]t was Congress’s intent for a ‘theft offense’ to be more broadly defined than the common-law definition of larceny, and ... by using that phrase, rather than ‘theft,’ Congress signaled that it was not presenting an exhaustive list of offenses, but rather, a definition with broad meaning.”). At the time of the enactment of § 1101(a)(43)(G), the Model Penal Code had for several years provided for criminal liability for theft of services, and over half the states had criminalized theft of services under their respective criminal codes. See United States v. Corona-Sanchez, 291 F.3d 1201, 1216 n.6 (9th Cir. 2002) (en banc) (Rymer, J., dissenting in part) (citing American Law Institute, Modern Penal Code and Commentaries II § 223.7, cmt. 1 (1980)), abrogated on other grounds by statute as éxplained in United States v. Gomez-Mendez, 486 F.3d 599, 604-05 (9th Cir. 2007). Taylor teaches that in using generic terms, like “burglary” in 18 U.S.C. § 924(e), Congress intended to adopt “the generic sense in which the term is now used in the criminal codes of most [sjtates.” 495 U.S. at 598, 110 S.Ct. 2143. It fairly follows that Congress “intended to incorporate a modern understanding of theft,” Abimbola, 378 F.3d at 178-79, in which case it' likely intended theft of services to fall within the ambit of the term “a theft offense.”
In concluding otherwise, the Ninth Circuit en banc court observed that services are not property. See Corona-Sanchez, 291 F.3d at 1208; accord United States v. Juarez-Gonzalez, 451 Fed.Appx. 387, 392 (5th Cir. 2011) (unpublished opinion). But why this should make any difference is unclear. The Ninth Circuit observed that theft of services was “generally [not] included within the scope of ordinary theft statutes because one’s labor is not one’s ‘property.’” Corona-Sanchez, 291 F.3d at 1208. But the more relevant point is that the thieving of services was already a theft offense in most states at the time Congress enacted the relevant provision of the INA.
The Ninth Circuit also observed that “the Supreme Court has carefully maintained the distinction between ‘property’ and other rights when construing criminal statutes.” Id. (citing McNally v. United *267States, 483 U.S. 350, 356, 107 S.Ct. 2875, 97 L.Ed.2d 292 (1987), superseded on other grounds by statute as recognized in Cleveland v. United States, 531 U.S. 12, 19-20, 121 S.Ct. 365, 148 L.Ed.2d, 221 (2000)). Be that as it may, the word “property” does not appear in the INÁ’s provision denominating “theft offense[s]” as aggravated felonies. See 8 U.S.C. § 1101(a)(43)(G). And the fact that certain states criminalized as larceny only the theft of property simply does not mean that the laws of many other states criminalizing the theft of services are not also theft offenses. Otherwise, the Ninth Circuit offered no reason for its holding besides a desire to maintain a national uniformity that does not exist.
We have considered, too, the language of V-Z-S- in which the BIA itself described a taking of property as a theft.3 De Lima would read this statement as saying that only a taking of property can constitute a theft. We see no reason to read the BIA’s statement in such a restrictive manner. The question of services Versus property was not before the BIA in VZ-S-. Rather, at issue was the question whether an alien could b.e said to have been convicted of a “theft offense” by sustaining a conviction- under a state law that criminalized the taking .of property even absent intent to deprive permanently, traditionally an element of common-law theft. The BIA considered and decided which thefts of property eqnstitute a “theft offense”; it did not, in so doing, construe “theft offense” as including only tangible-property theft crimes.
We therefore hold that Connecticut’s third-degree larceny statute’s imposition of criminal liability for theft of services does not broaden the offense beyond the limits of a “theft offense” under the INA. See Abimbola, 378 F.3d at 178; cf. Ilchuk, 434 F.3d at 622-23 (finding that Pennsylvania’s theft-of-services statute is categorically a “theft offense” aggravated felony under the INA).
C.
De Lima argues, finally, that his conviction cannot categorically qualify as a conviction for a “theft offense” under 8 U.S.C. § 1101(a)(43)(G) because Connecticut’s third-degree larceny statute criminalizes theft by fraud, which the BIA itself does not treat as a “theft, offense.” See Matter of Garcia-Madruga, 24 I. & N. Dec. 436, 440-41 (BIA 2008) (finding fraud crimes do not pass the categorical test because theft requires lack of consent, while fraud necessarily involves the acquisition of property by consent involuntarily given). De Lima, though, did not advance this argument before the BIA. The law is clear that “theories not advanced before the BIA may not be surfaced for the first time in a petition for judicial review of the BIA’s final order.” Pérez Batres v. Lynch, 796 F.3d 157, 160 (1st Cir. 2015) (quoting Makhoul v. Ashcroft, 387 F.3d 75, 80 (1st Cir. 2004)); see Mazariegos-Paiz v. Holder, 734 F.3d 57, 62 (1st Cir. 2013); Sousa v. INS, 226 F.3d 28, 31-32 (1st Cir. 2000). This limitation on the scope of our review is jurisdictional. Mazariegos-Paiz, 734 F.3d at 62-63.
At oral argument, counsel suggested that we could nevertheless consider De Lima’s argument that a state statute that prohibits theft by fraud is not categorically a theft offense under the INA, because the argument is simply a different approach to the overbreadth challenge that he mounted *268both before the BIA and in his petition for review. This suggestion is untenable in light of our holding in Ravindran v. INS, 976 F.2d 754 (1st Cir. 1992). There, an immigration judge denied a petitioner’s application for asylum because the petitioner did not have a well-founded fear of persecution. Id. at 757. The petitioner appealed to the BIA asserting that the immigration judge was wrong because the petitioner had demonstrated his fear of, and the likelihood he would suffer, persecution on account of his political opinions. Id. at 760. When the BIA affirmed, the petitioner sought relief in our court, asserting that “the BIA did not consider his claim of persecution on the basis of membership in a particular social group.” Id. We noted that it indeed did not, but that its failure was “attributable to the fact that petitioner failed to raise this claim before the BIA,” and we accordingly found that we lacked jurisdiction to consider his challenge. Id. at 760-61.
Here too, we are confronted with a challenge that was not presented to the BIA: De Lima contended before the BIA only that the criminal activities proscribed in subsections 7,12,15, and 16 of section 53a-119 fell outside the scope of the generic federal definition of a theft offense based on' the' arguments we have addressed above. Now he wants to argue that one or more different subsections of the 'Connecticut statute fail to qualify as a theft offense for a different reason.
■ Telling an agency that subsection A does not qualify as a theft offense for reason X simply does not raise, much less exhaust, the argument that subsection B does not qualify as a theft offense for reason Y. True, both arguments feed into the common ultimate conclusion that a conviction under the broad Connecticut statute is not categorically a theft offense. But if we were to deem the assertion of this ultimate conclusion to be sufficient to exhaust all independent routes to reaching such a conclusion, we would present both the BIA and the opposing party with an unfair and daunting task. Quite literally, each would have to generate, sua sponte, a list of all of the possible reasons why third degree larceny might not be a theft offense, and then perform a categorical analysis of all eighteen subsections of the Connecticut, statute, mapping each reason, against each subsection. Even a single such categorical analysis is an arduous task, requiring a close analysis of the specific statutory language put at issue, see Swaby v. Yates, 847 F.3d 62, 65-66 (1st Cir. 2017); of the ease law interpreting that language, see United States v. Fish, 758 F.3d 1, 4-5. (1st Cir. 2014); and of the extent to which the elements made relevant by that language match or fall within what the BIA has reasonably interpreted the INA’s “theft offense” provision to include, see Descamps v. United States, — U.S.-, 133 S.Ct. 2276, 2283, 186 L.Ed.2d 438 (2013) (citing Taylor, 495 U.S. at 600, 110 S.Ct. 2143). The ensuing categorical evaluation is often difficult and time consuming. See United States v. Tavares, 843 F.3d 1, 19 (1st Cir. 2016) (“The result is a Rube Goldberg jurisprudence of abstractions piled on top of one another in a manner that renders doubtful anyone’s confidence in predicting what will pop out at the end”). In this very case, for example, considering just the property and services provisions of selected subsections within section 53a-119 occupies over two-thousand words of analysis, even with the benefit of Becky’s spade work addressing one aspect of that consideration. The reasons why one subsection may broaden the elements' of the offense beyond the federal definition, while another subsection does not, can be as varied as the number of subsections to be examined. The analysis can be even more complicated, too, in cases (unlike this *269one) where the parties do not stipulate to the indivisibility of the statute in question. See Moncrieffe, 133 S.Ct. at 1684 (observing. that a modified categorical approach applies to “state statutes that contain several different crimes, each described separately”); Mathis v. United States, — U.S. -, 136 S.Ct. 2243, 2256-57, 195 L.Ed.2d 604 (2016) (instructing that, to determine whether a statute contains multiple different crimes with different elements or one crime that can be committed by multiple different means, a court should look to the statute, to the state’s case law, and, if those sources cannot answer the question, to the relevant charging instrument in the case before the court).
There is no precedent in our circuit that even remotely supports defining exhaustion so loosely as to encompass De Lima’s newly - minted challenge. Rather, Ravin-dr an and the subsequent similar cases we have cited, above, dictate the contrary: a reasonably elaborate argument that. requires substantial analysis and development to support a general proposition is not exhausted merely because a party raises an entirely independent argument, to support that same general proposition.
In the absence of First Circuit precedent providing grounds for us to review this belated challenge, our dissenting colleague relies on the Second Circuit’s decision in Gill v. I.N.S., 420 F.3d 82 (2d Cir. 2005). Gill rested on two grounds: a loose reading of the statutory exhaustion requirement, and a broad reading of the court’s power to override such requirements to avoid “manifest-injustice.” Id. at 86-88. The latter ground has since been implicitly rejected by the Supreme Court. See Grullon v. Mukasey, 509 F.3d 107, 115 (2d Cir. 2007) (citing Bowles v. Russell, 551 U.S. 205, 127 S.Ct. 2360, 168 L.Ed.2d 96 (2007)). And no other circuit has since applied Gill’s loose construction of the statutory exhaustion requirement of 8 U.S.C. § 1252(d)(1). Moreover, Gill involved an argument that was, in the court’s words, logically “subsidiary” to the argument made before the BIA, 420 F.3d at 87, not an alternative argument that stood on its own legs (as De Lima seeks to advance).
In deciding what arguments have been suitably developed before our own court, we regularly decline to assume for ourselves the burdens that De Lima would have us impose on the BIA. For example, in United States v. Whindleton, 797 F.3d 105 (1st Cir. 2015), we considered whether Assault with a Dangerous Weapon (“ADW”) under Massachusetts General Laws eh. 265, § 15B(b), categorically qualified as a “violent felon/’ under the Armed Career Criminal Act (“ACCA”), 18 U.S.C. § 924(e). Whindleton, 797 F.3d at 107. We found that it did, id at 116, but expressly declined to consider one possible argument to the contrary—that Massachusetts’s ADW might not be a “violent felony” because the least culpable mental state necessary to sustain a conviction might be less culpable than that of a violent felony— because “the defendant [had] not developed, or even expressly asserted, any argument” to that effect, id. at 116- n,12. Similar examples abound in other types of cases. See, e.g., Eldridge v. Gordon Bros. Grp., 863 F.3d 66, 83-84 (1st Cir. 2017) (finding waiver in a civil case where a mismanagement-based breach-of-implied-warranty claim was raised and argued, but only arising out of an unjustified liquidation, not also as to the mismanagement of stores); Coningford v. Rhode Island, 640 F.3d 478, 482-83 (1st Cir. 2011) (finding a habeas corpus petitioner failed to exhaust a claim in state- court, and rejecting the argument “that by arguing generally that the wayward introduction of prior bad acts evidence rendered his trial unfair, [petitioner] presented the state court with a *270claim based on a ‘particular right’ guaranteed by the Constitution”); United States v. Slade, 980 F.2d 27, 31 (1st Cir. 1992) (applying the raise-or-waive rule in a criminal case, noting that “a party is not at liberty to articulate specific arguments for the first time on appeal simply because the general issue was before the district court”).
We observe, too, that the dissent’s attempt to mitigate the unfortunate harm caused by a procedural defalcation leads it to intrude on the BIA’s expertise even as it professes not to do so. The dissent projects onto the BIA’s decision in Garda-Madruga a finding that no fraudulent taking can be a “theft offense” under the INA. But the BIA in Garda-Madruga determined only that fraudulent takings are “ordinarily” not theft offenses. Garcia-Madruga, 24 I. & N. Dec. at 440. The decision in that matter sheds insufficient light on whether the BIA, were it presented with the argument De Lima now raises before us, might interpret “theft offense” under the INA to include the fraudulent takings proscribed in section 53a-119 on the basis that the INA’s definition of consent differs from Connecticut’s. See id. at 440 n.5 (declining to “discount the possibility that the theft and fraud aggravated felony compartments are not watertight such that certain crimes ... may constitute both a theft, offense and one ‘involving] fraud’ ” (alteration in original)). Indeed, the BIA’s holding in Matter of Ibarra, 26 I. & N. Dec. 809, 811-13 (BIA 2016), shows that the INA has come to such a conclusion before. And the mere fact that section 53a-119 criminalizes a number of types of fraudulent takings likewise has little to do with whether one or more of those types of fraudulent takings would not constitute a “theft offense.”
Finally, the dissent offers a series of policy reasons why exhaustion of theories of relief should not be required. In a perfect world, Congress might be persuaded by some of these reasons. It might, for example, limit exhaustion requirements to arguments that turn on matters with regard to which the INA has special expertise. Or it might make an exception for really strong arguments that would change a case’s result. But we do not have the authority to adopt these changes to the law on our own accord.
De Lima also advances quite a different twist on his theft-by-fraud argument. Returning to his preserved argument that theft of services is not a theft offense, he argues that if theft of services has a home in the INA’s category of aggravated felonies, it is more like theft by fraud than it is like theft generally, and theft by fraud is an aggravated felony under the INA only when the victim’s loss is more than $10,000, see 8 U.S.C. § 1101(a)(43)(M)(i). Whether De Lima has preserved his ability to marshal this point in support of his clearly preserved argument that theft of services is not a theft offense, we need not decide. What distinguishes theft by fraud from theft of property is not the object of the theft. Rather, it is the means by which the theft is accomplished. Moreover, when Congress creates a general category of “theft offense” and a special category for one type of theft (theft by fraud), the logical inference is that other types of theft not specially dealt with remain in the general category.
III.
For the foregoing reasons, we uphold the BIA’s decision. De Lima’s petition for review is denied.
. De Lima was also convicted of fourth-degree larceny under Connecticut law, Conn. Gen. Stat. § 53a-125, for a separate larceny. That conviction, however, was subsequently vacated and played no part in De Lima’s removal proceedings.
. The government has not provided Shepard documents that would shed light on whether De Lima was convicted under a particular subsection of section 53a-124. See Conteh v. Gonzales, 461 F.3d 45, 59 (1st Cir. 2006) (citing Shepard v. United States, 544 U.S. 13, 20-23, 125 S.Ct. 1254, 161 L.Ed.2d 205 (2005)).
. "[A] taking of property constitutes a ‘theft’ whenever there is criminal intent to deprive the owner of the rights and benefits of ownership, even if such deprivation is less than total or permanent.” V-Z-S-, 22 I. & N. Dec. at 1346.