Appeal from the United States District Court for the Southern District of Indiana, Indianapolis Division. No. 1:24CR00153-001 — James P. Hanlon , Judge .

ARGUED MAY 18, 2026 — DECIDED AUGUST 12, 2026

Before SCUDDER, KIRSCH, and TAIBLESON, Circuit Judges .

TAIBLESON, Circuit Judge . This appeal requires us to interpret the word “conspiring” in the United States Sentencing Guidelines, where it appears in the definition of 'crime of violence.” At Christopher Lloyd’s sentencing for unlawful possession of a firearm, the district court found that Lloyd had previously committed a “crime of violence” under the Guidelines— specifically, conspiring to commit armed robbery under Indiana law. That finding increased Lloyd's Guidelines Page 2range. Lloyd appeals, arguing that Indiana defines “conspiring” more broadly than the Guidelines do, so no Indiana conspiracy conviction can be a 'crime of violence' under the Guidelines.

The distinction Lloyd points to is that under Indiana law, a defendant may be guilty of conspiracy if his only coconspirator is an undercover agent— i.e. , not a genuine participant. This is called a “unilateral” conspiracy. By contrast, in many (but not all) other jurisdictions, a conspiracy must be “bilateral,” meaning that at least two coconspirators are genuine participants. The Guidelines, for their part, define “crimes of violence” to include certain conspiracies without specifying anything about their bilateral or unilateral nature.

We apply the familiar categorical approach in interpreting this Guidelines provision. That approach requires us to divine the elements of the “generic” crime of conspiracy as of the time the relevant Guideline was enacted, and then compare generic conspiracy to Indiana conspiracy. Following that method, we conclude that generic conspiracies around the time of the relevant Guideline’s enactment were only bilateral, and that Indiana conspiracy is therefore not a categorical match to conspiracy under the Guidelines. This means that Lloyd’s Indiana conspiracy conviction should not have counted as a “crime of violence” under the Guidelines. We vacate and remand to the district court for resentencing.

In 2024, Lloyd fell asleep with a gun on him at a White Castle restaurant in Indiana. He was charged with, and ultimately pleaded guilty to, possessing a firearm as a felon in violation of 18 U.S.C. § 922(g)(1).

Page 3At sentencing, the parties disputed Lloyd’s proper Guidelines offense level , which turned on his previous felony convictions. For a § 922(g) offense like Lloyd's, the Sentencing Guidelines assign a higher base offense level to a defendant who has “at least two felony convictions” for “a crime of violence or a controlled substance offense .” U.S. SENT’G GUIDELINES MANUAL § 2K2.1(a)(2) (2024). The Presentence Investigation Report stated that Lloyd had two qualifying felony convictions, both under Indiana law: dealing in cocaine, a 'controlled substance offense,' and conspiracy to commit armed robbery, a “crime of violence.” Lloyd objected, arguing that his Indiana conspiracy conviction did not constitute a “crime of violence,” so he only had one qualifying prior felony (which would produce a lower Guidelines offense level) .

The term 'crime of violence' is defined in U.S.S.G. § 4B1.2. As relevant here, it includes 'any offense under federal or state law, punishable by imprisonment for a term exceeding one year, that … has as an element the use, attempted use, or threatened use of physical force against the person of another; or … is … robbery,' as well as the '[i]nchoate [o]ff ense s” of “conspiring to commit any such offense. ” On its face, that definition includes conspiracies to commit armed robbery. But Lloyd argued that Indiana’s conspiracy statute criminalizes both unilateral and bilateral conspiracies and is therefore broader than the generic form of conspiracy reflected in the Guidelines. As a result, Lloyd argued, not all Indiana conspiracies categorically mee t the Guidelines' definition of conspiracy. Under the logic of the categorical approach, that would mean no Indiana conspiracy conviction could be a 'crime of violence” under Guidelines § 4B1.2—including Lloyd’s own armed robbery conspiracy, even though his conspiracy was, in fact, a bilateral one.

Page 4The district court overruled Lloyd's objection , finding that in 2023—the time the district judge deemed relevant—the generic definition of conspiracy included the unilateral theory, so there was not a categorical mismatch, and Lloyd’s Indiana conspiracy conviction constituted a “crime of violence.” Applying the enhanced base offense level reflected in the PSR, Lloyd’s Guidelines range was 77 to 96 months’ imprisonment. The judge sentenced Lloyd to 96 months’ imprisonment, and Lloyd now appeals.

Whether a defendant’s prior conviction counts as a “crime of violence” under the Guidelines “is a question of law we review de novo ” when, as here, it was preserved below. United States v. Taylor , 630 F.3d 629, 632 (7th Cir. 2010).

The parties agree on the analytical framework for determining whether Indiana conspiracy convictions count as conspiracies under § 4B1.2: We apply “the familiar categorical approach, under which this court must compare the least-culpable conduct of the state offense with the generic federal offense. ” United States v. Pemberton , 85 F.4th 862, 866 (7th Cir. 2023) (citing Mathis v. United States , 579 U.S. 500, 509 (2016)). 'If the state offense sweeps more broadly than the generic federal offense, then the defendant does not qualify for the enhancement.” Id. Under the categorical approach, “the facts of each defendant’s conduct” are irrelevant; instead, we look only to “the elements of the statute of conviction” to determine the scope of the state offense. Pugin v. Garland , 599 U.S. 600, 603– 04 (2023) (quoting Taylor v. United States , 495 U.S. 575, 601 (1990)).

Page 5The parties also agree about Indiana law: Indiana’s conspiracy statute, IND. CODE § 35-41-5-2, criminalizes both bilateral and unilateral conspiracies. See Garcia v. State , 271 Ind. 510, 394 N.E.2d 106, 109–10 (1979); Tidwell v. State , 644 N.E.2d 557, 559 (Ind. 1994). This means that in Indiana, “proof that the co-conspirator intended to carry out the conspiracy is not required,” Tidwell , 664 N.E.2d at 559, so the statute covers “a person who agrees to commit a crime with, for example, an undercover agent who does not sincerely conspire to break the law,” Pemberton , 85 F.4th at 865.

This leaves us with one principal task: D efining the scope of 'generic' conspiracy, which will tell us the meaning of “conspiring” in § 4B1.2 of the Guidelines. If the word “conspiring” in the Guidelines includes both bilateral and unilateral conspiracies, then Lloyd’s Indiana conspiracy conviction can be a “crime of violence.” But if the word “conspiring” in the Guidelines is limited to bilateral conspiracies, then Lloyd’s Indiana conspiracy conviction does not count. We encountered a nearly identical question only a few years ago in United States v. Pemberton , 85 F.4th 862, and outlined the path to answering it. There, our review was for plain error, so we ultimately did not resolve the question. We do so today.

Identifying the elements of a generic crime requires us to ascertain 'the offense as commonly understood.' Descamps v. United States , 570 U.S. 254, 257 (2013); see also Pugin , 599 U.S. at 609 (defining a generic crime by identifying 'the widespread and contemporary understanding of' the offense ). We conduct that analysis as of the time of the relevant Guideline’s “enact[ment].” Pemberton , 85 F.4th at 867; see also, e.g. , United States v. Hill , 963 F.3d 528, 532 (6th Cir. 2020). Here, that meansPage 6 1989, which is the year that the Sentencing Commission added conspiracies to the definition of 'crime of violence.' The Commission did so in an “Application Note[]” to § 4B1.2 specifying that the term “crime of violence” includes the inchoate crimes of 'aiding and abetting, conspiring, and attempting to commit' the described offenses. U.S.S.G. § 4B1.2 cmt. n.1 (1989).

The Government resists this conclusion and instead urges us to define generic conspiracy as of 2023. That year, the Sentencing Commission relocated the language about inchoate offenses from its original location in § 4B1.2's application notes to the body of § 4B1.2, in subsection (d). See U.S.S.G. § 4B1.2(d) (2023); U.S.S.G. supplement to app. C, amend. 822 at 245–47 (2023). In the Government’s view, we should treat this reorganization as “substantive” and “meaningful,” and therefore consider 2023 to be the time of § 4B1.2(d)'s 'passage.” The district court agreed with this approach.

We view the 2023 amendment differently. The amendment made no substantive change to the pertinent text; instead, it moved it. The Commission explained that the move responded to a circuit split about the authoritative weight of § 4B1.2’s application notes. Some courts (but not ours) had declined to follow certain Guidelines application notes in the wake of Kisor v. Wilkie , 588 U.S. 558 (2019). See United States v. White , 97 F.4th 532, 537–38 (7th Cir. 2024) (describing circuit split). The 2023 amendment confirmed that, whatever one's views of the application notes in general, inchoate offense s like conspiracy are included within the meaning of “crime of violence” under § 4B1.2. See U.S.S.G. supplement to app. C, amend. 822 at 245–47 (2023) (“The amendment addresse[d] this circuit conflict by moving, without change, the Page 7commentary including certain inchoate and accessory offenses in the definitions of 'crime of violence’ and ‘controlled substance offense' to the text of the guideline .”); United States v. Pugh , 147 F.4th 801, 806 (7th Cir. 2025) (noting this court’s treatment of the application note was unchanged by the 2023 amendment). There is no reason to read the amendment as affecting the substantive meaning of the word ' conspiring.” Cf. Antonin Scalia & Bryan A. Garner, Reading Law 257 (2012) (“legislative restyling exercises,” such as “the nonsubstantive redrafting” of Federal Rules, do “not amend prior enactments unless [they] do[] so clearly”). Thus, in ascertaining the generic elements of conspiracy, we look to the law as it existed in 1989.

We must determine whether, in 1989, the “generally accepted contemporary meaning' of 'conspiring' included participating in unilateral conspiracies. See Taylor , 495 U.S. at 596. “In doing so, we consider contemporaneous federal and state laws, dictionary definitions, and the Model Penal Code.' Pemberton , 85 F.4th at 867.

Traditionally, conspiracies at common law were bilateral. See, e.g. , Garcia , 394 N.E.2d at 108 (“The bilateral concept is the traditional view of conspiracy as derived from common law. It is formulated in terms of two or more persons agreeing to commit a crime, each with intent to do so.”) (collecting cases); State v. Marian , 62 Ohio St. 2d 250, 405 N.E.2d 267, 269 (1980) (“Traditionally a criminal conspiracy has existed only if there has been a meeting of the minds of two or more people to act together to achieve an unlawful purpose.”). Through the midtwentieth century, federal and state jurisdictions all followed suit, defining conspiracy in accordance with that term's Page 8longstanding historical meaning. See, e.g. , Dierdre A. Burgman, Unilateral Conspiracy: Three Critical Perspectives , 29 DEPAUL L. REV. 75, 78-83 (1979); Peter Buscemi , Conspiracy: Statutory Reform Since the Model Penal Code , 75 COLUM. L. REV. 1122, 1122–26 (1975).

That started to change about sixty years ago. In 1962, after a close vote, the drafters of the Model Penal Code (MPC) broke new ground and adopted a “unilateral” concept of conspiracy. The MPC did so by defining “conspiracy” in terms of an individual defendant’s agreement to commit a crime, rather than focusing on an agreement between two or more people. MODEL PENAL CODE § 5.03(1) (A.L.I. 1962). As the comments explained, this language 'depart[ed] from the traditional view of conspiracy as an entirely bilateral or multilateral relationship.” MODEL PENAL CODE § 5.03 cmt. 2(b) (A.L.I. 1985).¹ And so the unilateral approach was born. See Buscemi , supra , at 1136 (“Under a unilateral formulation, the crime of conspiracy is committed when a person agrees to proceed in a prohibited manner; under a bilateral formulation, the crime of conspiracy is committed when two or more persons agree to proceed in such manner.”). The MPC’s adoption of the unilateral approach marked the beginning of a movement, as some jurisdictions began embracing this broader understanding of conspiracy. See generally id . at 1122–28 (surveying changes to state statutes).

Page 9Federal law, however, did not. See Pemberton , 85 F.4th at 867. For well over a century now, the general federal conspiracy statute has consistently required that “two or more persons” participate in a conspiracy, which is classic bilateral language. See, e.g. , LXX REV. STAT. § 5440 (1875); 18 U.S.C. § 88 (1926); 18 U.S.C. § 371 (1988); 18 U.S.C. § 371 (2024). During the late 1970s, Congress considered conspiracy legislation in the proposed Federal Criminal Code Reform Act, which employed the new unilateral language focusing on the defendant’s 'agree[ment] with' another to engage in criminal conduct. Criminal Code Reform Act of 1977, S. 1437, 95th Cong. § 1002(a) (1977). But that legislation did not pass. See 37 CONG. Q. 1899, 1900 (1979); H.R. REP. No. 95-29, at III, 3, 9–12 (1979).

For its part, the Supreme Court has long described the offense of 'conspiracy' as 'a partnership in crime' or 'the agreement of two persons,' while not addressing the bilateral-vs.-unilateral question directly. Pinkerton v. United States , 328 U.S. 640, 643-44 (1946); see also, e.g. , Ziglar v. Abbasi , 582 U.S. 120, 153 (2017) (“Conspiracy requires an agreement— and in particular an agreement to do an unlawful act— between or among two or more separate persons.”). The very reason conspiracies are dangerous, the Supreme Court has explained, is that “collective criminal agreement—partnership in crime—presents a greater potential threat to the public than individual delicts.” Callanan v. United States , 364 U.S. 587, 593 (1961). And for our own part, we have repeatedly reiterated that “an agreement must exist among coconspirators , that is, those who actually intend to carry out the agreed-upon criminal plan,” meaning “[a] defendant is not liable for conspiring solely with an undercover government agent or a government informant.” United States v. Corson , 579 F.3d 804, 811 (7th Cir.

Page 102009) (citing United States v. Mahkimetas , 991 F.2d 379, 383 (7th Cir. 1993)). We are aware of-and the parties have pointed to—no disagreement among the circuits on this question, now or in 1989.

The MPC did, however, inspire some states to recognize unilateral conspiracies. See, e.g., Marian , 405 N.E.2d at 269, 271 (interpreting OHIO REV. CODE ANN. § 2923.01). But as of 1989, at least, a clear majority of states still required bilateral agreement. Though the parties have some quibbles over the precise count, they agree that at least 28 states plus D.C., Puerto Rico, and the Virgin Islands used a bilateral-only approach in 1989.² And even in states that amended their conspiracy statutes and implemented language similar to the MPC, some courts refused to abandon the bilateral approach, finding the revisions insuffi cient to support such a dramatic departure from the long-established understanding of conspiracy. See, e.g. , People v. Foster , 99 Ill. 2d 48, 457 N.E.2d 405, 406–09 (1983); State v. Grullon , 212 Conn. 195, 562 A.2d 481, 484-86 (1989); State v. Pacheco , 125 Wash. 2d 150, 882 P.2d 183, 185–88 (1994).

Secondary sources also point to the conclusion that, as of 1989, a “conspiracy” generally required bilateral agreement. Every prominent legal dictionary that we know of from around that time defined 'conspiracy' in terms of an agreement between two people, rather than adopting the MPC’s focus on the individual defendant. See, e.g. , Black's Law Page 11Dictionary (5th ed. 1979) ( defining ' [c]riminal conspiracy” as “[a]n agreement or confederacy of two or more persons to do a criminal or unlawful act or to do a lawful act in an unlawful or criminal manner”); Black’s Law Dictionary (6th ed. 1990) (same); Bryan A. Garner, A Dictionary of Modern Legal Usage 124, 145 (1st ed. 1987) (defining ' [c]onspiracy” to require “an agreement between two or more people to behave in a manner that will automatically constitute an offense by at least one of them”); Merriam-Webster’s Dictionary of Law 98 (1st ed. 1996) (defining “conspiracy” as “an agreement between two or more people to commit an act prohibited by law or to commit a lawful act by means prohibited by law”). The Supreme Court has relied on the same sources to ascertain the generic definitions of statutory terms. See Esquivel-Quintana v. Sessions , 581 U.S. 385, 392 (2017); Pugin , 599 U.S. at 604.

Together, federal law, state law, and secondary sources all indicate that generic conspiracy in 1989 required bilateral agreement. This leaves the Government with the MPC’s outlier approach , which represented a conscious effort to change the law rather than to distill it. The Government points to the state law “trend,” inspired by the MPC, towards expanding the meaning of conspiracy to embrace unilateral conspiracy. But trends alone cannot carry the day under the Supreme Court's categorical-approach cases, at least not until the trends garner majority support. And if anything, the Government’s point only reinforces that the default understanding of conspiracy at the time was still limited to bilateral conspiracy. Thus the word “conspiring” in Guidelines § 4B1.2 refers only to bilateral conspiracies.

To be fair, we acknowledge this conclusion is not exactly intuitive. The Sentencing Commission wrote the wordPage 12 'conspiring' without specifying how many coconspirators must genuinely agree, at a time when the law on this point was in flux nationwide. The word's plain meaning in 1989 might therefore have been broad enough to encompass both types of conspiracies—just like the unadorned word 'office,' in 2026, might encompass both a home office and an office in the workplace. But that line of reasoning would be inconsistent with the Supreme Court’s categorical-approach precedents, and so we pursue it no further. Instead, defining a generic offense requires us to choose a side and specify the elements that represented a consensus view of conspiracy at the time of enactment. See, e.g. , Esquivel-Quintana , 581 U.S. at 395–97 (holding that generic meaning of “sexual abuse of a minor” requires that “the victim must be younger than 16,” which was the age of consent in the majority of states, even though 16 states had higher ages of consent). The Government points to no case in which a majority of the pertinent jurisdictions and sources defined an offense one way while this court or the Supreme Court defined the generic offense more broadly.

A s a practical matter, we also recognize the significance of our conclusion. The parties agree that nowadays, at least twenty states have conspiracy statutes adopting the unilateral approach, like Indiana does. If we are right, then none of those states’ conspiracy crimes can be “crimes of violence” under § 4B1.2. That result produces odd disparities: For example, an Illinois conspiracy may still qualify as a “crime of violence,” while identical conduct committed just across the border in Indiana may not. But this is the result demanded by the categorical approach, and so it is the one we adopt here.³

Page 13* * *

Because the generic definition of conspiracy in 1989 required bilateral agreement, Lloyd’s conspiracy conviction under Indiana Code § 35-41-5-2 does not satisfy § 4B1.2 and should not have subjected him to an enhanced base offense level under § 2K2.1(a)(2). Lloyd’s sentence is VACATED and his case is REMANDED for resentencing consistent with this opinion.