Esteras v. United States

606 U.S. 185, 145 S. Ct. 2031

Opinions

Syllabus

NOTE: Where it is feasible, a syllabus (headnote) will be released, as is being done in connection with this case, at the time the opinion is issued. The syllabus constitutes no part of the opinion of the Court but has been prepared by the Reporter of Decisions for the convenience of the reader. See United States v. Detroit Timber & Lumber Co., 200 U. S. 321, 337.

SUPREME COURT OF THE UNITED STATES

Syllabus

ESTERAS v. UNITED STATES

CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT

No. 23–7483. Argued February 25, 2025—Decided June 20, 2025*

Edgardo Esteras pleaded guilty to conspiring to distribute heroin, and the District Court sentenced him to 12 months in prison followed by a 6-year term of supervised release. While on supervised release, Esteras was arrested and charged with domestic violence and other crimes. The District Court revoked Esteras’s supervised release and ordered 24 months of reimprisonment, explaining that Esteras’s earlier sentence had been “rather lenient” and that his revocation sentence must “promote respect for the law,” a consideration enumerated in 18 U. S. C. §3553(a)(2)(A). The Sixth Circuit affirmed, holding that a district court may consider §3553(a)(2)(A) when revoking supervised release. Held: A district court considering whether to revoke a defendant’s term of supervised release may not consider §3553(a)(2)(A), which covers retribution vis-à-vis the defendant’s underlying criminal offense. Pp. 4–17.

(a) In determining the appropriate sentence for a federal defendant, a district court must consider 10 factors set forth in §3553(a). Among those factors is §3553(a)(2)(A), which references “the need for the sentence imposed” “to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense.” This provision speaks to the retributive purpose of punishment.

A judge who imposes a term of imprisonment may also impose a post-imprisonment term of supervised release. Section 3583(c) enumerates the factors that a court must consider when deciding to impose

2

supervised release. Eight of the ten §3553(a) factors are listed. Significant here, §3583(c) excludes §3553(a)(2)(A), which covers retribution vis-à-vis the defendant’s underlying criminal offense. When a court decides whether to revoke a defendant’s term of supervised release , §3583(e) provides that the same eight factors apply. So the question is this: In determining whether to revoke supervised release, may a district court account for these omitted factors—and specifically §3553(a)(2)(A)? Esteras says no; the Government says yes. Pp. 4–8.

(b) The Court agrees with Esteras. District courts cannot consider §3553(a)(2)(A) when revoking supervised release. That conclusion follows from the well-established canon of statutory interpretation—“expressio unius est exclusio alterius”—which means that expressing one item of an associated group excludes another item not mentioned. See Chevron U. S. A. Inc. v. Echazabal, 536 U. S. 73, 80. While Congress elsewhere set forth ten factors that must generally inform a district court’s sentencing decisions, it provided in §3583(e)—the provision governing the revocation of supervised release—that courts must consider only eight of those ten factors. The natural implication is that Congress did not intend courts to consider the other two factors, including §3553(a)(2)(A).

The statutory structure confirms this negative inference. Neighboring provisions governing the imposition and revocation of other kinds of sentences instruct courts to consider all the §3553(a) factors. But for supervised release—and supervised release only—Congress omitted §3553(a)(2)(A). This is a distinction with a difference. And Congress ’s decision to exclude retribution from the calculus also comports with supervised release’s role in the criminal justice scheme. Supervised release “is not a punishment in lieu of incarceration.” United States v. Granderson, 511 U. S. 39, 50. Rather, it “fulfills rehabilitative ends” and “provides individuals with postconfinement assistance.” United States v. Johnson, 529 U. S. 53, 59–60. So when a defendant violates a condition of supervised release, courts must consider the forward -looking sentencing ends, but may not consider the backward- looking purpose of retribution.

The Court has twice interpreted the omission of §3553(a)(2)(A) from the provision governing the imposition of supervised release to mean that district courts may not consider that factor. See Tapia v. United States, 564 U. S. 319; Concepcion v. United States, 597 U. S. 481. Although Tapia and Concepcion both deal with the imposition of supervised release under §3583(c), the same reasoning applies to revocation under §3583(e): the omission of §3553(a)(2)(A) means that courts may not consider it. Pp. 8–10.

(c) The Government’s counterarguments are unpersuasive. The Government reads the exclusion of §3553(a)(2)(A) from §3583(e) to

3

mean that district courts must consider enumerated factors and need not (but may) consider unenumerated ones. Given sentencing judges’ discretion, the Government argues, Congress would have been more explicit had it intended to foreclose the consideration of retributive aims when revoking supervised release. But what Congress said is clear: The itemized list in §3583(e) is exhaustive and supplies the entire universe of factors courts may consider. And the Government’s reading trivializes the omission of §3553(a)(2)(A): As the Government recognizes, a court may “consider” an enumerated factor but give it no weight. Yet there is negligible difference between saying that a court must consider a given factor (but may give it no weight) and saying that a court may consider a given factor (if the court so chooses).

Next, the Government argues that Esteras’s reading is unworkable because considering other enumerated factors—such as “the nature and circumstances of the offense” under §3553(a)(1)—will necessarily imply consideration of the retributive principles captured by §3553(a)(2)(A). While the Government asks how a court may consider the nature and circumstances of the offense without also considering the retributive principles captured by §3553(a)(2)(A), the answer is straightforward. Courts may consider the offense’s nature and circumstances to inform the considerations set forth in §§3553(a)(2)(B), (C), and (D)—deterrence, incapacitation, and rehabilitation. But courts cannot consider them as relevant to §3553(a)(2)(A)’s retributive focus.

Finally, the Government interprets §3583(g), which mandates revocation in certain circumstances, as subject to the full list of §3553(a) factors. And it would be anomalous, the Government says, if courts could consider retribution in mandatory revocations under §3583(g) but not in discretionary revocations under §3583(e)(3). The correct reading of §3583(g) does not affect the Court’s analysis. That §3583(e) might operate differently from §3583(g) is no reason to disregard §3583(e)’s plain meaning. Pp. 11–16.

(d) When appellate courts review a claim that the district court has impermissibly relied on §3553(a)(2)(A), much will turn on whether the defendant has properly objected to the district court’s impermissible reliance on §3553(a)(2)(A). If not, the defendant’s appeal will be governed by plain-error review. But if a court considers retribution for the original offense over the defendant’s objection, and the error was not harmless, then the court of appeals should vacate the order and remand for the court to apply the correct standard. None of this turns the Court’s reading of §3583(e) into a “substance-free reverse magic- words requirement,” as the Government suggests. Brief for United States 37. The Government conflates the proper interpretation of §3583(e) with an appellate court’s ability to enforce that interpretation . Pp. 16–17. 88 F. 4th 1163, 95 F. 4th 1004, vacated and remanded.

4

BARRETT, J., delivered the opinion of the Court, in which ROBERTS, C. J., and THOMAS, KAGAN, and KAVANAUGH, JJ., joined, and in which SOTOMAYOR and JACKSON, JJ., joined as to all but Part II–B. SOTOMAYOR, J., filed an opinion concurring in part and concurring in the judgment, in which JACKSON, J., joined. JACKSON, J., filed an opinion concurring in part and concurring in the judgment. ALITO, J., filed a dissenting opinion , in which GORSUCH, J., joined.

Footnotes

1. *Together with Jaimez, fka Watters v. United States and Leaks v.

Opinion of the Court

NOTICE: This opinion is subject to formal revision before publication in the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, pio@supremecourt.gov, of any typographical or other formal errors.

SUPREME COURT OF THE UNITED STATES

_________________

No. 23–7483

_________________

EDGARDO ESTERAS, PETITIONER v.

UNITED STATES

TIMOTHY MICHAEL JAIMEZ, FKA TIMOTHY M. WATTERS, PETITIONER v. UNITED STATES TORIANO A. LEAKS, JR., PETITIONER v.

UNITED STATES

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT [June 20, 2025]

JUSTICE BARRETT delivered the opinion of the Court. A criminal sentence may include both time in prison and a term of supervised release. 18 U. S. C. §3583(a). Supervised release comes with conditions—for instance, the defendant must refrain from committing another crime. §3583(d). If the defendant violates one of these conditions, then the district court may revoke the term of supervised release and require reimprisonment. But a court may do so only “after considering” an enumerated list of sentencing factors: those “set forth in section 3553(a)(1), (a)(2)(B), (a)(2)(C), (a)(2)(D), (a)(4), (a)(5), (a)(6), and (a)(7).” §3583(e). Conspicuously missing from this list is §3553(a)(2)(A), which directs a district court to consider “the need for the sentence imposed” “to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense.” The Sixth Circuit held that a district court may consider that factor nonetheless.

2

We disagree. Congress’s decision to enumerate most of the sentencing factors while omitting §3553(a)(2)(A) raises a strong inference that courts may not consider that factor when deciding whether to revoke a term of supervised release . This inference is consistent with both the statutory structure and the role that supervised release plays in the sentencing process. Accordingly, we vacate the judgments of the Sixth Circuit and remand for further proceedings.

I

In 2018, Edgardo Esteras pleaded guilty to conspiring to distribute heroin. Varying downward from the 15-to-21- month sentencing range, the District Court sentenced him to 12 months in prison, to be followed by a 6-year term of supervised release.

When Esteras completed his term of imprisonment, his period of supervised release began. Three years in, the probation office notified the District Court that Esteras had been arrested. According to the mother of his children, Esteras had threatened to kill her and had fired three rounds into her vehicle. Esteras was charged in municipal court with domestic violence, aggravated menacing, and criminal damaging, although the charges were ultimately dismissed at the request of the victim.

The District Court held a revocation hearing, at which it found that Esteras had violated the conditions of his supervised release. The District Court remarked that Esteras was “no stranger to law violations and no stranger to federal court” and that his previous drug sentences had been “rather lenient.” App. 96a. The District Court revoked his supervised release and ordered 24 months of reimprisonment , an upward variation from the advisory 6-to-12-month range, with three more years of supervised release to follow.

At this point, Esteras’s counsel objected, arguing that the District Court had impermissibly considered “the factor in

3

Section 3553(a)(2)(A).” Id., at 105a. While Esteras’s counsel lodged his objection for the record, he recognized that Sixth Circuit precedent allows district courts to consider §3553(a)(2)(A) when revoking a term of supervised release. See United States v. Lewis, 498 F. 3d 393, 399–400 (2007). The District Court in turn acknowledged that part of its decision “certainly [was] the need for the sentence imposed, to promote respect for the law.” App. 105a. In a subsequent written order, the District Court explained that it had “considered the factors and conditions for sentencing listed in 18 U. S. C. §3553(a)” and had imposed the term of reimprisonment in part to “promote respect for the law.” Id., at 115a–116a.

The Sixth Circuit affirmed, providing two justifications for its view that district courts may consider §3553(a)(2)(A) when revoking a term of supervised release. See 88 F. 4th 1163 (2023). First, §3583(e) does not say “that the court may consider ‘only’ those factors” specifically enumerated, so there was no indication that Congress meant to displace the “considerable discretion” that district courts typically enjoy “over supervised-release decisions.” Id., at 1167. Second , a rule prohibiting district courts from considering §3553(a)(2)(A) would be unworkable. As the panel saw it, “the purportedly forbidden considerations mentioned in §3553(a)(2)(A) tend to be ‘essentially redundant’ with the permitted ones,” so “[t]o think about the one requires the judge to think about the other.” Ibid. (quoting Lewis, 498 F. 3d, at 400).

The circuits are divided as to whether district courts may consider §3553(a)(2)(A) when revoking a term of supervised release.¹ We granted certiorari to resolve this split. 604

4

U. S. ___ (2024).2

II

A

When a district court sentences a federal defendant, the judge may impose a term of imprisonment, a term of probation , or a fine. See §3551(b). In determining the appropriate sentence, the court must consider certain factors set forth in §3553(a). Some are obvious: The court must consider “the nature and circumstances of the offense,” “the history and characteristics of the defendant,” and “the kinds of sentences available.” §§3553(a)(1), (3). Others require courts to consider guidelines or policies issued by the Sentencing Commission. §§3553(a)(4)–(5). And still others reflect discrete policy aims, such as “the need to avoid unwarranted sentence disparities” among similarly situated defendants and “the need to provide restitution to any victims of the offense.” §§3553(a)(6)–(7).

While all these factors are important, §3553(a)(2) captures the traditional heartland of criminal sentencing. It requires courts to consider:

5

“the need for the sentence imposed—

“(A) to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense;

“(B) to afford adequate deterrence to criminal conduct ;

“(C) to protect the public from further crimes of the defendant; and “(D) to provide the defendant with needed educational or vocational training, medical care, or other correctional treatment in the most effective manner.”

“These four considerations—retribution, deterrence, incapacitation , and rehabilitation—are the four purposes of sentencing generally.” Tapia v. United States, 564 U. S. 319, 325 (2011). They speak to the questions at the core of any system of criminal justice: What sentence does the defendant deserve? What sentence will deter criminal conduct in the future? What sentence will protect the public? And what sentence is most likely to help the defendant rehabilitate for transition back into society?

A judge who imposes a term of imprisonment may—and sometimes must—impose a postimprisonment term of supervised release. §3583(a). “Supervised release is ‘a form of postconfinement monitoring’ that permits a defendant a kind of conditional liberty by allowing him to serve part of his sentence outside of prison,” subject to conditions on his behavior. Mont v. United States, 587 U. S. 514, 523 (2019) (quoting Johnson v. United States, 529 U. S. 694, 697 (2000)). Some conditions are mandatory, such as requiring that the defendant not commit additional crimes and, in most cases, that the defendant submit to drug testing. §3583(d). Others are up to the discretion of the court. Ibid.

When a court “determin[es] whether to include a term of supervised release,” as well as the length and conditions of such a term, it must “consider the factors set forth in section

6

3553(a)(1), (a)(2)(B), (a)(2)(C), (a)(2)(D), (a)(4), (a)(5), (a)(6), and (a)(7).” §3583(c). Only two of the §3553(a) factors are absent from this list. The first is §3553(a)(2)(A), which requires the court to consider the need for the sentence imposed to “reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense,” i.e., the retributive purposes of sentencing. The second is §3553(a)(3), which requires the court to consider “the kinds of sentences available.” When a court decides whether to revoke a defendant’s supervised release, the very same subset of §3553(a) factors—that is, all but two—applies . §3583(e)(3).3 The reason for excluding §3553(a)(3) from the list is selfevident : When a district court decides whether to revoke a term of supervised release, there is no need to consider “the kinds of sentences available,” because supervised release is the only matter at issue.4 The exclusion of §3553(a)(2)(A) is

7

more significant—and here, controversial. Esteras argues that district courts cannot consider the retributive purpose articulated in §3553(a)(2)(A) when revoking supervised release . The Government insists that courts can consider it, although they are not required to do so.

B

To see what is at stake, keep in mind that §3553(a)(2)(A) does not speak of retribution generally—it references the need for the sentence “to reflect the seriousness of the offense ” and to “provide just punishment for the offense.” (Emphasis added.) In the context of a revocation hearing, the “offense” is the underlying crime of conviction, not the violation of the supervised-release conditions. The opening provision of Title 18’s sentencing chapter clearly uses “offense ” to refer to a criminal conviction. See §3551(a) (stating that the sentencing provisions apply to “a defendant who has been found guilty of an offense described in any Federal statute”). Yet conduct that constitutes a violation of supervised release “need not be criminal.” Johnson, 529 U. S., at 700; see, e.g., §3583(g)(3) (mandating revocation when a defendant fails to comply with drug testing imposed as a condition of release). Furthermore, neighboring provisions in §3583 consistently use the word “offense” to refer to the defendant’s original crime of conviction, while using the word “violation” to refer to the conduct that triggers revocation . Compare §3583(e)(3) (referencing the “offense that resulted in the term of supervised release”); §3583(h) (“the offense that resulted in the original term of supervised re-

8

lease”), with §3583(d)(3) (referencing a “violation of a condition of supervised release”); §3583(i) (same).5 So the question is this: In determining whether to revoke a defendant’s term of supervised release (per §3583(e)), may a district court account for the need to exact retribution for the defendant’s underlying crime (per §3553(a)(2)(A))? Esteras says no; the Government says yes.

III

A

We agree with Esteras: District courts cannot consider §3553(a)(2)(A) when revoking supervised release. This conclusion follows directly from the application of a well- established canon of statutory interpretation: “expressio unius est exclusio alterius”—in plain English, “ ‘expressing one item of [an] associated group or series excludes another left unmentioned.’ ” Chevron U. S. A. Inc. v. Echazabal, 536 U. S. 73, 80 (2002) (quoting United States v. Vonn, 535 U. S. 55, 65 (2002); alteration in original). Here, §3553(a) lays out 10 factors that inform a district court’s sentencing decision .6 Section 3583(e) provides that a district court may revoke a term of supervised release “after considering” 8 of

9

these 10 factors. The natural implication is that Congress did not intend for courts to consider the other two factors, §3553(a)(2)(A) and §3553(a)(3). Indeed, the expressio unius canon has particular force here because the §3553(a) sentencing factors constitute an “established series,” such that any “omission” from that series necessarily “bespeaks a negative implication.” Chevron U. S. A. Inc., 536 U. S., at 81. And the fact that Congress included almost the entire list makes the exclusion of §3553(a)(2)(A) and §3553(a)(3) all the more glaring.

The statutory structure confirms this negative inference. Neighboring provisions that govern the imposition and revocation of sentences other than supervised release instruct the court to consider all the factors in §3553(a). For instance , when imposing a term of probation, the court “shall consider the factors set forth in §3553(a) to the extent that they are applicable.” §3562(a). And the court may revoke such a term “after considering the factors set forth in section 3553(a).” §3565(a); see also §3582(a) (same for imposition of term of imprisonment); §3572(a) (same for imposition of a fine). So for supervised release—and for supervised release only—Congress omitted §3553(a)(2)(A). This, we think, is a distinction with a difference. After all, our task is to “give effect to, not nullify Congress’ choice to include” that factor “in some provisions but not others.” Gallardo v. Marstiller, 596 U. S. 420, 431 (2022).

Congress’s decision to exclude retribution from the calculus also comports with the role of supervised release in our current criminal justice scheme. Fines, probation, and imprisonment are a court’s primary tools for ensuring that a criminal defendant receives just deserts for the original offense . Supervised release, by contrast, “is not a punishment in lieu of incarceration.” United States v. Granderson,

10

511 U. S. 39, 50 (1994). Rather, it “fulfills rehabilitative ends” and “provides individuals with postconfinement assistance .” United States v. Johnson, 529 U. S. 53, 59–60 (2000). So when a defendant violates the conditions of his supervised release, it makes sense that a court must consider the forward-looking ends of sentencing (deterrence, incapacitation, and rehabilitation), but may not consider the backward-looking purpose of retribution.

Finally, our conclusion is consistent with this Court’s precedent. We have twice interpreted the omission of §3553(a)(2)(A) from the provision governing the imposition of supervised release (§3583(c)) to mean that district courts may not consider that factor when imposing supervised release . In Tapia, we stated that, “when imposing a term of supervised release,” “a court may not take account of retribution (the first purpose listed in §3553(a)(2)).” 564 U. S., at 326.7 So too, in Concepcion v. United States, we remarked that “in determining whether to include a term of supervised release, and the length of any such term, Congress has expressly precluded district courts from considering the need for retribution.” 597 U. S. 481, 494 (2022). Although Tapia and Concepcion both deal with the imposition of supervised release under §3583(c), the same reasoning applies to the revocation of supervised release under §3583(e): The omission of §3553(a)(2)(A) from the list of factors means that courts may not consider it.

Text, structure, and precedent all point in the same direction : Congress’s decision to exclude §3553(a)(2)(A) from §3583(e)’s list of sentencing factors means that district courts cannot consider §3553(a)(2)(A) when deciding whether to revoke supervised release.

11

B

We are unpersuaded by the Government’s and the dissent ’s counterarguments. First, the Government and the dissent read the exclusion of §3553(a)(2)(A) as supporting a different inference—namely, that a district court must consider the enumerated factors but need not consider the unenumerated ones, such as §3553(a)(2)(A). Because “sentencing judges ‘enjo[y] discretion in the sort of information they may consider,’ ” the Government argues, Congress would have used more explicit language had it wanted to foreclose courts from considering retribution when revoking supervised release. Concepcion, 597 U. S., at 491 (quoting Dean v. United States, 581 U. S. 62, 66 (2017); alteration in original). For example, Congress could have said that courts may revoke supervised release “after considering only” the enumerated factors. Brief for United States 23; see also post, at 4–6 (ALITO, J., dissenting).

When interpretive disputes arise, it is easy to imagine how Congress could have drafted the statute to avoid them. But Congress cannot anticipate (much less account for) every future statutory skirmish—and even if it could, courts have no authority to hold Congress to a “perfect as we see it” standard of drafting. On the contrary, we have “routinely construed statutes to have a particular meaning” even when “Congress could have expressed itself more clearly.” Luna Torres v. Lynch, 578 U. S. 452, 472 (2016). What Congress said here gets its point across just fine: The itemized list in §3583(e) is exhaustive. As we explained, the provision allows district courts to consider all but two of the §3553(a) factors, and because one is inapplicable in this context (§3553(a)(3)), §3553(a)(2)(A) is the only consequential omission. So it makes sense to read §3583(e) as supplying the universe of factors that the district court can consider. And while the Government is right that district courts generally enjoy discretion over sentencing, see Brief for United States 21–22, Congress chose to limit that discretion here.

12

It makes much less sense to read §3583(e) as the Government does—that is, as listing only the factors that a court must consider, while leaving room for the court to consider an additional factor if it so chooses. This reading of §3583(e) trivializes the omission of §3553(a)(2)(A). After all, on the Government’s view, a court may “consider” an enumerated factor but give it no weight. Tr. of Oral Arg. 61. The difference between saying that a court must consider a given factor (but may give it no weight) and saying that a court may consider a given factor (if the court chooses) is negligible. Why would Congress have bothered to exclude §3553(a)(2)(A) from the list?

The dissent has no response to this point, so it instead analogizes to different—and inapposite—contexts. For example , the dissent says, suppose that a law clerk is directed to complete a certain number of bench memos before a deadline; surely that instruction would not carry the inference that the law clerk may not complete more work too. Post, at 5. But the important difference between that context and this one is that sentencing considerations, unlike bench memos, can be zero sum: Adding a new consideration to the mix reduces the role that the enumerated considerations play. In fact, the new consideration might lead to a different result altogether.

A play on the dissent’s own analogy reveals its weaknesses . Imagine that a judge tells her hypothetical law clerk to prepare a bench memo with a recommendation formed “after considering” statutory text, context, and structure, as well as any relevant precedent. Having looked at these sources, the law clerk is inclined to recommend that the judge reverse the court below. But the law clerk decides to look elsewhere—say, to social science studies— and she uses these sources to inform her recommendation. In the end, she submits a memo recommending that the judge vote to affirm the lower court’s decision. According to

13

the dissent, the law clerk has obeyed her boss’s instructions . Needless to say, we disagree.

More to the point, the dissent never explains why Congress omitted §3553(a)(2)(A) from the list of considerations applicable to a revocation proceeding. After all, in the dissent ’s view, district courts will almost inevitably consider that factor. See post, at 3–4, 8–11, 12–13, n. 7. So why did Congress exclude it? In the dissent’s view, the omission is just hortatory: “Congress may have sought to encourage sentencing courts to give greater weight to the remaining purposes of sentencing.” Post, at 12–13, n. 7.8 But omitting consideration of §3553(a)(2)(A) altogether would be an odd way of accomplishing that very limited goal. And one need not “live in a world of airy abstractions,” post, at 1, to recognize that Congress’s drafting decisions have significance.

Next, the Government and the dissent argue that Esteras’s reading of §3553(a)(2)(A) is unworkable, because consideration of the other enumerated factors will necessarily imply consideration of the retributive principles captured by §3553(a)(2)(A). The Government relies primarily on §3553(a)(1), which references “the nature and circumstances of the offense.” How, the Government asks, is a court to consider the nature and circumstances of the offense without also considering the “need for the sentence imposed” “to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense”? §3553(a)(2)(A).

The answer is straightforward. A court may consider the

14

nature and circumstances of the offense as relevant for the considerations set forth in §§3553(a)(2)(B), (C), and (D)— namely, deterrence, incapacitation, and rehabilitation—but a court cannot consider the nature and circumstances of the offense as relevant to §3553(a)(2)(A)’s retributive focus. For instance, if the defendant’s original offense was particularly violent, that fact might inform the court’s judgment as to whether revocation is necessary “to protect the public from further crimes of the defendant.” §3553(a)(2)(C). But the court could not revoke based on the view that, given the violent nature of the underlying offense, the defendant deserves additional punishment.9 The dissent, for its part, argues that we are simply wrong to understand §3553(a)(2)(A) as synonymous with retribution . Post, at 8–11. Under the dissent’s view, §3553(a)(2)(A) serves an extensive (if amorphous) role: The phrase “ ‘to promote respect for the law’ ” encompasses “the whole point of deterrence,” while the reference to the “ ‘seriousness of the offense’ ” invokes the “heightened need for deterrence and incapacitation.” Post, at 8–9. But this approach renders §3553(a)(2)(A) largely duplicative of the clear references to deterrence, incapacitation, and rehabilitation that follow in §§3553(a)(2)(B), (C), and (D). Recall, too, that there are four traditional purposes of sentencing. Given that §§3553(a)(2)(B), (C), and(D) set forth three of these purposes, it is no wonder that §3553(a)(2)(A) provides

15

the fourth. Not to mention that the provision speaks in retributive terms: the “seriousness of the offense,” the need “to promote respect for the law,” and, in particular, the need “to provide just punishment for the offense.” §3553(a)(2)(A). In any event, we say nothing about the meaning of §3553(a)(2)(A) today that we have not already said before. See Tapia, 564 U. S., at 326 (“retribution” is “the first purpose listed in §3553(a)(2)”); Concepcion, 597 U. S., at 494.

Finally, the Government suggests that §3583(g), which mandates revocation in certain circumstances, undermines Esteras’s interpretation of the statute. Section 3583(g) does not specify which factors apply when a court determines the length of the defendant’s postrevocation term of imprisonment .10 So, the Government reasons, reimprisonment under §3583(g) must be subject to the general rule for prison sentences articulated in §3582(a), and that provision directs courts to the full list of §3553(a) factors. It would be anomalous, the Government says, if a district court could consider retribution as part of a mandatory revocation under §3583(g) but not as part of a discretionary revocation under §3583(e)(3). Esteras, for his part, argues for consistency in the other direction: He argues that §3583(e)’s abbreviated list governs in both contexts.

We need not resolve the dispute because it does not affect our analysis. To begin, the Government’s reading of §3583(e) does not solve the problem that the Government has identified. As the Government sees it, if revocation is discretionary and thus governed by §3583(e), the court may consider §3553(a)(2)(A), but if revocation is mandatory and thus governed by §3583(g), the court must consider

16

§3553(a)(2)(A). So while the Government says that it would be “highly anomalous to have different regimes for determining the length of prison terms under Sections 3583(e)(3) and 3583(g),” that anomaly persists under the Government ’s own reading of the statute. Brief for United States 26. In any event, that §3583(e) might operate differently from §3583(g) is not reason to disregard §3583(e)’s plain meaning.

IV

At oral argument, the Government expressed concern that it would be difficult for appellate courts to determine whether a district court has impermissibly relied on §3553(a)(2)(A). See Tr. of Oral Arg. 49–50. So we conclude with a few observations about appellate review.

Much will turn on whether the defendant objects. If the defendant does not make the district court aware that it may be impermissibly relying on §3553(a)(2)(A), then the defendant’s appeal will be governed by plain-error review. See Fed. Rule Crim. Proc. 52(b). In that event, the district court’s order revoking supervised release and requiring reimprisonment will be affirmed unless it is “ ‘clear’ ” or “ ‘obvious ’ ” that the district court actually relied on §3553(a)(2)(A)—because it did so either expressly or by unmistakable implication. United States v. Olano, 507 U. S. 725, 734 (1993).

If the defendant does object to the district court’s reliance on §3553(a)(2)(A), we anticipate that the district court will recognize its potential error and clarify its revocation decision to make clear that it is not taking account of §3553(a)(2)(A). For example, the district court could withdraw any impermissible justification or explain that a stray reference to a §3553(a)(2)(A) factor was intended to bear on another §3553(a) factor or merely prefatory. If the court nonetheless considers the need to exact retribution for the defendant’s original criminal offense, and if the error was

17

not harmless, then the court of appeals should vacate the court’s order and remand for the court to apply the correct standard. See Fed. Rule Crim. Proc. 52(a).

This does not mean, as the Government suggests, that our reading of §3583(e) amounts to a “substance-free reverse magic-words requirement.” Brief for United States 37. The Government conflates the proper interpretation of §3583(e) with an appellate court’s ability to enforce that interpretation . The “requirement” here is very much substantive : District courts may not consider the retributive purpose of §3553(a)(2)(A) before revoking supervised release . We trust that district courts will heed that instruction regardless of the practical likelihood of reversal.11

* * *

District courts may revoke a term of supervised release after considering the factors enumerated in §3583(e). Because §3553(a)(2)(A) is excluded from that list, district courts may not consider it. The judgments of the Sixth Circuit are vacated, and the cases are remanded for further proceedings consistent with this opinion.

It is so ordered.

Footnotes

1. 1In addition to the Sixth Circuit, the First, Second, and Third Circuits have held that district courts may consider §3553(a)(2)(A). See United States v. Vargas-Dávila, 649 F. 3d 129, 131–132 (CA1 2011); United States v. Williams, 443 F. 3d 35, 47–48 (CA2 2006); United States v. Young, 634 F. 3d 233, 238–242 (CA3 2011). By contrast, the Fourth, Ninth, and Tenth Circuits have held that a court may not consider §3553(a)(2)(A). See United States v. Crudup, 461 F. 3d 433, 439 (CA4 2006); United States v. Miqbel, 444 F. 3d 1173, 1182 (CA9 2006); United States v. Booker, 63 F. 4th 1254, 1259–1260 (CA10 2023). The Fifth and Seventh Circuits have taken intermediate positions. See United States v. Sanchez, 900 F. 3d 678, 684, and n. 5 (CA5 2011) (while §3553(a)(2)(A) is an impermissible factor for a district court to consider, “the forbidden factor must be ‘dominant’ ” to constitute reversible error); United States v. Clay, 752 F. 3d 1106, 1108–1109 (CA7 2014) (allowing consideration of §3553(a)(2)(A) if the district court “relies primarily” on the enumerated factors). 2 Timothy Jaimez and Toriano Leaks joined Esteras’s petition for cer- tiorari. Jaimez and Leaks, like Esteras, argued that the District Courts in their revocation proceedings had improperly considered §3553(a)(2)(A). The Sixth Circuit held that these arguments were fore- closed by Esteras and Lewis. See United States v. Jaimez, 95 F. 4th 1004, 1007 (2024); United States v. Leaks, 2024 WL 2196795, *1 (Mar. 6, 2024). 3 The same list of factors also applies when a district court modifies the terms of supervised release, such as by terminating supervised release early, by extending the term of supervised release, or by altering the con- ditions of supervised release. §§3583(e)(1)–(2), (4). 4 The dissent offers a different view of §3553(a)(3). In its view, the ter- mination, extension, modification, and revocation of supervised release are all different “kinds of ” sentences. Post, at 7–8 (opinion of ALITO, J.). The Government never makes this argument—and for good reason. To begin, it is textually implausible. As the statute articulates, there are three “kinds of sentences available” for an individual “found guilty of an offense”: “a term of probation,” “a fine,” and “a term of imprisonment.” §§3551(b)(1)–(3); §3553(a)(3). In fact, a term of supervised release is not itself a “sentence” at all; it is a component of a defendant’s prison sen- tence. See §3583(a) (“The court, in imposing a sentence to a term of im- prisonment . . . , may include as a part of the sentence a requirement that the defendant be placed on a term of supervised release after imprison- ment” (emphasis added)). Thus, to the extent the dissent thinks that the termination, extension, modification, and revocation of supervised re- lease are different “kinds of sentences available” to the court, it is mis- taken. §3553(a)(3). The dissent’s interpretation is perplexing for another reason too: It treats Congress’s omission of §3553(a)(3) from the set of factors that gov- ern revocation proceedings as nothing more than an inexplicable drafting error; under its view, not only may district courts consider §3553(a)(3), they must do so. See post, at 7 (a judge “must consider exactly what is set out in §3553(a)(3)”). Silence is an unusual way to convey such an instruction. 5 Because §3553(a)(2)(A) speaks only to the “offense,” and “offense” here can mean only the underlying criminal conviction, we address only whether §3583(e) precludes the court from considering retribution for the underlying criminal conviction. “Offense” aside, the Government argues that §3553(a)(4)(B) and §3553(a)(5) authorize courts to consider policy statements issued by the Sentencing Commission that sound in retribu- tion for the violation of the conditions of the supervised release. See Brief for United States 11, 34; §3583(e) (requiring courts to consider §3553(a)(4) and §3553(a)(5)). The Commission has adopted the view that “the sentence imposed upon revocation [is] intended to sanction the vio- lator for failing to abide by the conditions of the court-ordered supervi- sion.” United States Sentencing Commission, Guidelines Manual ch. 7, pt. A, intro. 3(b) (Nov. 2024); see also ibid. (in revocation proceedings, a court may “sanction” an offender’s “breach of trust”). We take no position on whether this is a permissible consideration. 6 For those counting: the six factors enumerated in §§3553(a)(1), (3), (4), (5), (6), and (7), plus the four factors separately enumerated in §§3553(a)(2)(A)–(D). 7 The dissent suggests that when we said that “ ‘a court may not take account of retribution,’ ” we may have actually meant that “a judge, as a matter of discretion, may choose not to take retribution into account.” Post, at 14. We leave it to the reader to judge whether this is a plausible interpretation. 8 Alternatively, the dissent posits that Congress may have understood each of the three §3553(a)(2)(A) factors to “apply in some, though not all, hearings regarding the alteration of supervised release.” Post, at 12–13, n. 7. Which hearings make the cut? Frankly, we are not sure. Some parts of the dissent’s lengthy footnote suggest all; others suggest none; and still others suggest that it “depends on context.” Ibid. In our view, however, “[t]he Sentencing Reform Act does not place district judges in such a predicament.” Post, at 1–2. 9 The dissent suggests that this will be an impossible task for district courts. See post, at 1 (Courts must “engage in mind-bending exercises”); post, at 10 (They must “probe their mental processes to ensure that no thoughts of retribution entered in”); post, at 11 (We require a “soul- searching obligation”). Not so. We routinely require judges and juries to attend to some considerations while ignoring others. See, e.g., Samia v. United States, 599 U. S. 635, 646 (2023) (“Evidence at trial is often ad- mitted for a limited purpose, accompanied by a limiting instruction,” and “our legal system presumes that jurors will ‘ “attend closely the particu- lar language of [such] instructions in a criminal case and strive to under- stand, make sense of, and follow” ’ them” (alteration in original)). 10 Section 3583(g) states that, if the defendant violates certain condi- tions of supervised release, “the court shall revoke the term of supervised release and require the defendant to serve a term of imprisonment not to exceed the maximum term of imprisonment authorized under subsec- tion (e)(3).” (Emphasis added.) 11 The dissent takes a curious tack here. It suggests that we lack the courage of our convictions because we explain that defendants bear the burden of preserving objections, that district courts will generally comply with the law once an objection is raised, and that plain-error review ap- plies to unpreserved claims. See post, at 18–19. But there is nothing “remarkable” about our explanation, which simply articulates orthodox principles of party presentation and appellate review. Post, at 18. And it is surprising that the dissent opens its discussion by suggesting that we are insufficiently attuned to “the practical effec[t] of [our] holdin[g],” post, at 1, yet closes by complaining that we are overly sensitive to its consequences, post, at 18–19. 1 In some cases, the imposition of supervised release is mandatory. See, e.g., 18 U. S. C. §3583(k) (kidnapping offenses); §3583(j) (terrorism offenses); 21 U. S. C. §§841(b), 960(b) (drug-trafficking offenses).

Opinion of Justice Sotomayor

SUPREME COURT OF THE UNITED STATES

_________________

No. 23–7483

_________________

EDGARDO ESTERAS, PETITIONER v.

UNITED STATES

TIMOTHY MICHAEL JAIMEZ, FKA TIMOTHY M. WATTERS, PETITIONER v. UNITED STATES TORIANO A. LEAKS, JR., PETITIONER v.

UNITED STATES

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT [June 20, 2025]

JUSTICE SOTOMAYOR, with whom JUSTICE JACKSON joins, concurring in part and concurring in the judgment.

I join all but Part II–B of the Court’s opinion. I agree that district courts may not rely on the retributive sentencing factor set forth in 18 U. S. C. §3553(a)(2)(A) to revoke supervised release. In my view, however, district courts revoking a term of supervised release should not consider retribution for any purpose. The question in these cases is whether, despite Congress ’s omission in the supervised-release statute of any reference to retribution, courts may consider retribution when they decide whether to revoke a term of supervised release. The answer is no. In specifying the factors courts should consider before revoking a term of supervised release , Congress cross-referenced eight out of the ten familiar §3553(a) factors, including three of the four purposes of sentencing: deterrence, incapacitation, and rehabilitation. “Conspicuously missing” from the supervised-release statute , however, “is §3553(a)(2)(A),” which describes the re-

2

tributive purpose of punishment. Ante, at 1. The straightforward inference, as we have said twice before, is that “Congress has expressly precluded district courts from considering the need for retribution” in supervised release proceedings . Concepcion v. United States, 597 U. S. 481, 494 (2022); see also Tapia v. United States, 564 U. S. 319, 326 (2011). That resolves this case.

The majority approaches the question differently. “In the context of a revocation hearing,” it reasons, the word “ ‘offense ’ ” in §3553(a)(2)(A) refers to “the underlying crime of conviction, not the violation of the supervised-release conditions .” Ante, at 7. So (the majority continues) by excluding §3553(a)(2)(A) from the supervised-release statute, Congress precluded courts from accounting “for the need to exact retribution for the defendant’s underlying crime.” Ante, at 8.

The problem with that framing is that it relies on an interpretive question with no clear answer. The term “offense ” in §3553(a)(2)(A) has no meaning “[i]n the context of a revocation hearing,” because the supervised-release statute contains no reference to §3553(a)(2)(A) or its directive to “provide just punishment for the offense.” Cf. ante, at 7. At most, then, the majority could ask what Congress would have meant by “offense” if it had included a reference to §3553(a)(2)(A) in this statute. Yet such speculation is unnecessary to resolve this case. Congress studiously omitted all reference to retribution from the supervised-release statute, so retribution should play no role in revocation hearings.

Because the majority frames the question as one about retribution for the original offense, it never decides whether the supervised-release statute precludes courts from exacting retribution for the defendant’s supervised-release violation . Yet the answer to that question is straightforward. As the Court holds today, the supervised-release statute does not permit consideration of §3553(a)(2)(A). That

3

means courts may consider only the remaining eight enumerated factors, none of which contain any reference to retribution . See §3583(e). As the majority recognizes, moreover , the retributive interest in punishment is ill suited to supervised-release proceedings. Supervised release “ ‘fulfills rehabilitative ends’ and ‘provides individuals with postconfinement assistance.’ ” Ante, at 10 (quoting United States v. Johnson, 529 U. S. 53, 59–60 (2000)); see also S. Rep. No. 98–225, p. 124 (1983) (Senate Report noting that “the primary goal” of supervised release “is to ease the defendant’s transition into the community . . . or to provide rehabilitation”). Thus, “when a defendant violates the conditions of his supervised release, it makes sense that a court must consider the forward-looking ends of sentencing (deterrence , incapacitation, and rehabilitation), but may not consider the backward-looking purpose of retribution.” Ante, at 10.

In sum, when a court considers whether a supervised- release violation warrants revocation and reimprisonment (its primary task during revocation hearings), it must look only to the forward-looking ends mentioned in the statute. As to either a supervised-release violation or the underlying offense, the backward-looking end of retribution is out of bounds. Although I would have made this point explicit, nothing in the Court’s opinion is inconsistent with it. Accordingly , I otherwise join the Court’s opinion.

Opinion of Justice Jackson

SUPREME COURT OF THE UNITED STATES

_________________

No. 23–7483

_________________

EDGARDO ESTERAS, PETITIONER v.

UNITED STATES

TIMOTHY MICHAEL JAIMEZ, FKA TIMOTHY M. WATTERS, PETITIONER v. UNITED STATES TORIANO A. LEAKS, JR., PETITIONER v.

UNITED STATES

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT [June 20, 2025]

JUSTICE JACKSON, concurring in part and concurring in the judgment.

The question presented in this litigation is whether sentencing courts can reference and rely upon the retributive concerns outlined in 18 U. S. C. §3553(a)(2)(A) when revoking a previously imposed term of supervised release. “Esteras argues that district courts cannot consider the retributive purpose articulated in §3553(a)(2)(A) when revoking supervised release,” while “[t]he Government insists that courts can consider it, although they are not required to do so.” Ante, at 7. Like the Court, I agree with Esteras: Courts cannot consider this omitted sentencing purpose when revoking supervised release.

That answer is straightforward and responsive. But the majority goes further, appearing to opine as to the precise contours of the retributive concerns that Congress has taken off the table. Ante, at 7–8. Venturing into this territory is not necessary in the context of this litigation. As JUSTICE SOTOMAYOR explains, whatever §3553(a)(2)(A) calls for when Congress has authorized its use, all that the

2

Court properly decides today is that Congress has forbidden sentencing judges to rely on that same set of considerations in the supervised-release-revocation context.

I disagree with the Court’s discussion of the scope of “offense ” in §3553(a)(2)A) as it relates to §3583(e) because, in my view, that part of the opinion confuses more than it clarifies . As sentencing judges know, revocation of supervised release occurs in response to new—often criminal—conduct . It is entirely conceivable, and perhaps even likely, that when Congress removed from the list of permissible revocation considerations §3553(a)(2)(A)’s “need for the sentence imposed” to exact retribution, it sought to prevent the promotion of retribution with respect to whatever new offense triggered the call for revocation of supervised release. At the very least, that possibility counsels against definitively asserting that §3553(a)(2)(A)’s conspicuous omission only speaks to whether a district court can “account for the need to exact retribution for the defendant’s underlying crime.” Ante, at 8 (emphasis added).

In any event, because §3553(a) requires district courts to impose sentences that are “sufficient, but not greater than necessary” to promote certain purposes of punishment, the kind of hairsplitting that the majority’s opinion invites is entirely impractical. Sentences either “reflect,” “afford,” or “provide” for the listed purposes of punishment (retribution , deterrence, incapacitation, rehabilitation) or they don’t. §3553(a)(2). It is hard to fathom that Congress meant for retribution to be both operable and inoperable as it relates to the imposition of a single revocation sanction, and it is even harder to imagine how a sentencing court can effectively achieve that binary objective.

Part II–B of the majority’s opinion is both unnecessary to the outcome of this litigation and inscrutable in light of how revocation sentences are actually determined and imposed. Therefore, I cannot join that Part of the opinion.

Justice Alito, dissenting

SUPREME COURT OF THE UNITED STATES

_________________

No. 23–7483

_________________

EDGARDO ESTERAS, PETITIONER v.

UNITED STATES

TIMOTHY MICHAEL JAIMEZ, FKA TIMOTHY M. WATTERS, PETITIONER v. UNITED STATES TORIANO A. LEAKS, JR., PETITIONER v.

UNITED STATES

ON WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT [June 20, 2025]

JUSTICE ALITO, with whom JUSTICE GORSUCH joins, dissenting .

Veteran trial judges often complain that their appellate colleagues live in a world of airy abstractions and do not give enough thought to the practical effects of their holdings . Today’s decision is likely to earn the rank of Exhibit A in the trial bench’s catalog of appellate otherworldliness. The Court interprets the Sentencing Reform Act to mean that a federal district-court judge, when considering whether to impose or alter a term of supervised release, must engage in mind-bending exercises. The judge must take into account “the nature and circumstances” of a defendant ’s offense but is forbidden to consider “the seriousness of the offense.” 18 U. S. C. §3553(a). The judge must consider what is needed to “dete[r]” violations of the law or to rehabilitate a defendant, i.e., to cause him to lead a lawabiding life, but cannot be influenced by a desire “to promote respect for the law.” Ibid. The Sentencing Reform Act does not place district judges

2

in such a predicament. Neither the statutory text, the interpretive canon on which the Court relies, nor the structure of the Act supports the Court’s interpretation.

I

A

In order to understand these cases and the points on which the majority and I disagree, some background information about federal sentencing is required. When a defendant is “found guilty of an offense,” he may be sentenced to “a term of probation,” “a fine,” or “a term of imprisonment .” §3551(b). And in crafting a sentence, the court is required to consider the 10 factors that are set out in §3553(a). (For convenience, I will call these the sentencing factors.)

If a sentencing court imposes a term of imprisonment, the court generally has discretion to include “as a part of the sentence” a term of supervised release. §3583(a).1 “Supervised release is ‘a form of postconfinement monitoring’ that permits a defendant a kind of conditional liberty by allowing him to serve part of his sentence outside of prison.” Mont v. United States, 587 U. S. 514, 523 (2019) (quoting Johnson v. United States, 529 U. S. 694, 697 (2000)). In deciding whether to impose and in devising a term of supervised release, a judge is required to consider 8 of the 10 sentencing factors noted above, specifically, §§3553(a)(1), (a)(2)(B), (a)(2)(C), (a)(2)(D), (a)(4), (a)(5), (a)(6), and (a)(7). (For convenience, I will call these the supervised-release factors.)

During a term of supervised release, a defendant must comply with certain standard conditions, such as reporting to a probation officer, as well as any special conditions that

3

are “tailored specifically” to “the individual case.” Administrative Office of the U. S. Courts, Overview of Probation and Supervised Release Conditions 8 (2024). If a defendant fails to comply with these conditions, revocation of supervised release is sometimes mandatory, see, e.g., §3583(g), but in most cases, the court has discretion to choose the appropriate response. For example, it may “extend” supervised release “if less than the maximum authorized term was previously imposed”; “modify, reduce, or enlarge the conditions of supervised release”; or, more seriously, “revoke ” supervised release and “require the defendant to serve in prison all or part of the term of supervised release authorized by statute for the offense that resulted in such term of supervised release.” §§3583(e)(2), (3). In deciding whether to alter a term of supervised release, the court is required to consider the same eight factors that had to be taken into account in imposing the term of supervised release in the first place.

In these cases, the lead petitioner, Edgardo Esteras, violated a term of his supervised release by assaulting the mother of his children and shooting at her vehicle. The sentencing judge revoked supervised release and, in doing so, referred to Esteras’s criminal history and the need to “encourage [him] to be respectful of the law.” App. 96a. Citing these remarks, Esteras contends that the judge based his decision on an impermissible factor, namely, a desire to “promote respect for the law.” This factor, which appears in §3553(a)(2)(A), is one of the two sentencing factors omitted from the list of supervised-release factors. Esteras argues , and the Court now agrees, that the omission of §3553(a)(2)(A) from the list of supervised-release factors means that it cannot be considered. In other words, the judge in Esteras’s case was obligated to put out of his mind any desire “to promote respect for the law.” Before today, one might have thought that every aspect of a federal judge’s work should aim to “promote respect for the law.”

4

Apparently, that is no longer true.

B

The question in these cases is what to make of the omission of §3553(a)(2)(A) from the list of supervised-release factors . The Court infers that consideration of this factor is forbidden, and, in so doing, relies almost entirely on a canon of interpretation that bears the Latin name expressio unius est exclusio alterius. In English, that means “[t]he expression of one thing implies the exclusion of others.” A. Scalia & B. Garner, Reading Law: The Interpretation of Legal Texts 107 (2012) (Scalia & Garner). To modify a classic example slightly, suppose a sign at the entrance of a park said: “No gasoline- or electric-powered vehicles allowed.” Under the expressio unius canon, it would be fair to infer that standard bikes are permitted. See H. Hart, Positivism and the Separation of Law and Morals, 71 Harv. L. Rev. 593, 607 (1958).

Although this canon has its place, Reading Law emphasizes that it “must be applied with great caution, since its application depends so much on context.” Scalia & Garner 107. Our cases have made the same point and have often rejected use of the canon because of the context in which the expression appeared. See, e.g., NLRB v. SW General, Inc., 580 U. S. 288, 302 (2017) (“The expressio unius canon applies only when circumstances support[ ] a sensible inference that the term left out must have been meant to be excluded ” (internal quotation marks omitted)); Barnhart v. Peabody Coal Co., 537 U. S. 149, 168 (2003) (“[T]he canon expressio unius est exclusio alterius does not apply to every statutory listing or grouping”). Reading Law also contains this important warning: “Even when an all-inclusive sense seems apparent, one must still identify the scope of the inclusiveness (thereby limiting [the] implied exclusion).” Scalia & Garner 108 (emphasis added).

In these cases, I agree with the Court that the omission

5

of two of the sentencing factors from the list of supervised- release factors gives rise to a negative inference, but I disagree with the Court about the nature of that inference. The most likely—and in my view, the appropriate—inference is simply that §3583(e), the provision listing the supervised- release factors, sets out an exclusive list of mandatory factors . That is “the scope of the inclusiveness.” The Court, however, goes further and infers not just that there are no other mandatory factors but that there are no other permissible factors. That aggressive application of the expressio unius canon goes too far. While the inference produced by that canon is always dependent on context, the enumeration of things that must be done generally does not suggest that no other things may be done.

Consider this example. Suppose I tell my law clerks on a Monday that by the end of the week they must complete the bench memos in 6 of the 12 cases scheduled for argument during the next sitting. That statement would strongly imply that completion of other bench memos by the end of the week is not mandatory. But the statement would not suggest that finishing other bench memos is forbidden. Many other similar examples could be provided.

To salvage its far-reaching inference, the Court offers its own variation on the law-clerk hypothetical. In the Court’s hypothetical, a judge tells a clerk “to prepare a bench memo with a recommendation formed ‘after considering’ statutory text, context, and structure, as well as any relevant precedent .” Ante, at 12. As the Court sees it, if the law clerk also reviews secondary sources “to inform her recommendation,” the law clerk did not “obe[y] her boss’s instructions.” Ante, at 12–13.

But nothing about the Court’s hypothetical suggests the law clerk must exclusively consider those four factors. Would the clerk necessarily infer from the judge’s instruction that nothing else could be considered? I doubt it. And if the clerk knows that the judge thinks that other factors,

6

such as statutory purpose or legislative history, are relevant considerations, the inference is unlikely.

For this reason, the Court’s use of the expressio unius canon exceeds its customary and proper use, and the Court provides no good reason for its aggressive use of that interpretive tool. The Court finds it significant that Congress included in the list of supervised-release factors “almost the entire list” of sentencing factors but left out §3553(a)(2)(A). Ante, at 9. According to the Court, this makes “all the more glaring” Congress’s intention that sentencing courts should not consider §3553(a)(2)(A) when modifying a term of supervised release. Ibid.

This is illogical. When a provision sets out a rule and lists items to which the rule applies, the length of the list may bolster the strength of the inference that the list is exhaustive , but it does not alter the scope of the rule. Suppose I tell friends who are visiting Washington, D. C., for the first time that they must visit a list of enumerated sites. If the list is long, they might infer that I do not think that there are any other sites that they should necessarily try to squeeze in. But it is doubtful that they would infer that I am suggesting that they should necessarily avoid seeing anything else that might catch their interest.

In short, the expressio unius canon is insufficient to support the Court’s interpretation.

II

The Court’s interpretation is also inconsistent with the text and structure of the Sentencing Reform Act.

A

Starting with the text, I note three features that counsel against acceptance of the Court’s interpretation.

First, the relevant provisions of the Sentencing Reform Act lack language like that included in other provisions of the Act that limit a court’s consideration of specified factors.

7

For example, §3582(a)—the provision that directly precedes the provision setting out the mandatory supervised-release factors in §3583—expressly states that a judge’s decision whether to impose a term of imprisonment may not be based on a desire to “promot[e] correction and rehabilitation .” When “Congress includes particular language in one section of a statute but omits it in another section of the same Act, it is generally presumed that Congress acts intentionally and purposely in the disparate inclusion or exclusion .” Russello v. United States, 464 U. S. 16, 23 (1983) (internal quotation marks omitted). Or take §3565(b), which (as amended in 1988) contains an implicit prohibition on consideration of the sentencing factors. See 102 Stat. 4361. That provision mandates revocation of probation if certain conditions are met, and Congress apparently gave courts no discretion to even consider the sentencing factors. As a result, the absence of any even remotely similar language in §3583(e) is telling.

Second, we can infer that the Sentencing Reform Act allows a judge to consider all the sentencing factors because there are some situations in which factors omitted from the list of supervised-release factors must be considered. One such situation occurs when a judge decides how to respond to a defendant’s violation of the terms of supervised release because in that situation the judge must consider exactly what is set out in §3553(a)(3), which is one of the sentencing factors omitted from the list of supervised-release factors. That provision refers to “the kinds of sentences available,” and when a judge is deciding what to do about a defendant who has violated the terms of supervised release, the judge must consider the available options, which include termination , extension, modification, and revocation. §3583(e). Because a term of supervised release is “part of a sentence,” each one of these options changes the defendant’s sentence. And therefore, this list of options constitutes a list of “the

8

kinds of sentences available.”² Consideration of “the kinds of sentences available” may be required during a revocation proceeding for yet another reason. When a sentencing court revokes a term of supervised release and sends the defendant back to prison, the maximum period of additional confinement authorized depends on the class of the underlying offense. See §3583(e)(3). So a sentencing court determining the permissible range of reimprisonment will necessarily consider the “kinds of sentences” authorized under the statute of conviction .

Third, and most important, the Court’s interpretation cannot account for the inextricable relationship between the omitted sentencing factor at issue in these cases, i.e., §3553(a)(2)(A), and the sentencing factors that a judge must consider in deciding whether to alter a term of supervised release. Recall that §3553(a)(2)(A) requires, among other things, consideration of “the seriousness of the offense,” and according to the Court, taking that factor into account is prohibited. But it is undisputed that a judge must consider “the nature and circumstances of the offense” when altering a term of supervised release. §§3553(a)(1), 3583(e). Since “the seriousness of the offense” is part of “the nature of the offense,” the Court’s interpretation gives district judges irreconcilable instructions.

Another part of §3553(a)(2)(A) creates a similar problem for the Court. Section 3553(a)(2)(A) refers to “the need . . . to promote respect for the law,” and because that provision is omitted from the supervised-release factors, the Court

9

holds that consideration of this factor is not allowed either. But the Sentencing Reform Act demands consideration of factors that are hard to separate from “promot[ing] respect for the law.” In deciding whether to alter a term of supervised release, a judge must consider “the need . . . to afford adequate deterrence to criminal conduct,” §3553(a)(2)(B), and the whole point of deterrence is to encourage respect for and compliance with the law.

The Court struggles in vain to explain away this clash between what its interpretation forbids and what the Sentencing Reform Act demands. It entirely ignores the relationship between “afford[ing] adequate deterrence” (a mandatory consideration) and “promot[ing] respect for the law” (outlawed under its interpretation). And as for the conflict between requiring consideration of the “nature and circumstances of the offense” and banning consideration of “the seriousness of the offense,” the Court’s response is an elaborate theory that has no grounding in the statutory text.

The Court begins by claiming that §3553(a)(2)(A) relates solely to retribution and not to any of the other three permitted goals of punishment—deterrence, incapacitation, and rehabilitation. See ante, at 13–14. But that is simply not true with respect to two of the objectives to which §3553(a)(2)(A) refers, i.e., the need for a sentence to reflect “the seriousness of the offense” and “to promote respect for the law.” These two factors implicate other purposes of punishment too. For example, if the offense in question is very serious, there is a heightened need for deterrence and incapacitation, and rehabilitation may be harder. So the Court’s theory veers off course at the start.

After that flawed beginning, the Court tells us that the supervised-release factors operate as a kind of purpose filter . Section 3553(a)(2) enumerates four factors that, in the Court’s view, reflect the four recognized purposes of punishment . The Court tells us that the “nature and circumstances of the offense” must be considered in light of one of

10

those four purposes. See ante, at 13–14; accord, Brief for Petitioners 34–35. The supervised-release factors, however , include only three of those four purposes and exclude the goal of retribution supposedly encapsulated in §3553(a)(2)(A). So, in the Court’s view, the supervised- release statute filters out §3553(a)(2)(A) and thereby prohibits a court from any consideration of the “nature and circumstances of the offense” in light of §3553(a)(2)(A)’s supposedly “retributive focus” and concern with the “seriousness of the offense.” Ante, at 14 None of this has any textual support, and it does not solve the problem faced by a judge who is compelled to consider the nature and circumstances of an offense but forbidden to consider its seriousness. If taken seriously, it does not alter in any way the information that judges may consider. Instead , it merely calls on them to probe their mental processes to ensure that no thoughts of retribution entered in.³

11

Imposing such a soul-searching obligation as a requirement that may be enforced in litigation is utterly impractical.

B

No more is needed to reject the Court’s interpretation, but I note for good measure that the structure of the Sentencing Reform Act also weighs against the Court’s interpretation. The Act gives express instructions about the way in which the sentencing factors apply to decisions regarding the imposition of each of the Act’s three primary penalties (i.e., imprisonment, probation, fines). These instructions fall into one of two categories. Instructions in the first category either mandate consideration of all the sentencing factors4 or require their consideration to the extent they are relevant to the type of sentence in question.5 By contrast, instructions in the second category expressly prohibit consideration of specific sentencing factors.6

12

What is remarkable about §3583(e), which lists the mandatory supervised-release factors, is that it does not fall into either of these categories. It neither demands consideration of all relevant sentencing factors nor prohibits consideration of some or all of them. Instead, it requires consideration of some of the mandatory sentencing factors but says nothing about the rest. The appropriate inference is that §3583(e) is meant to have a different effect: namely, to give sentencing courts discretion to consider the omitted factors.

Such discretion is consistent with the Sentencing Reform Act’s general approach to supervised release, which confers broad discretion on trial judges. In most cases, sentencing courts “may include” a term of supervised release “as a part” of “a sentence to a term of imprisonment,” §3583(a), and “may” terminate, extend, or revoke supervised release, §3583(e).

This grant of discretion tracks the background principles against which the Act was drafted, namely, “the established tradition of district courts’ sentencing discretion.” Concepcion v. United States, 597 U. S. 481, 495 (2022). As we recently noted, “[t]he only limitations on a court’s discretion to consider any relevant materials at an initial sentencing or in modifying that sentence are those set forth by Congress in a statute or by the Constitution.” Id., at 494; see §3661. “Congress is not shy about placing such limits” on sentencing discretion, id., at 494, and its failure to do so in §3583(e) permits the inference that sentencing courts retain discretion to consider “any relevant materials,” including the omitted §3553(a) factors, ibid.7

13

III

A

The Court attempts to buttress its interpretation by citing statements made in two of our prior cases, but the holdings in those cases were far afield from the question presented here, and the comments in question simply provided background. In Tapia v. United States, 564 U. S. 319 (2011), we held that a sentence of imprisonment cannot be imposed or extended to foster a defendant’s rehabilitation. Before directly addressing that issue in Part III of the opinion , Part II provided “statutory background.” Id., at 323. After noting the four purposes of sentencing, the opinion stated that under the Sentencing Reform Act, “a particular

14

purpose may apply differently, or even not at all, depending on the kind of sentence under consideration.” Id., at 326. As an example, the opinion stated that under §3583(c), “a court may not take account of retribution (the first purpose listed in §3553(a)(2)) when imposing a term of supervised release.” Ibid. (emphasis in original). Because the wording of §3583(c) is similar to that of §3583(e)—the provision that controls alteration of a term of supervised release—the Court today treats Tapia’s statement as strong support for its decision.

This reads too much into what is plainly peripheral dicta. For one thing, it is not even clear that the statement means what the Court claims. The verb “may” sometimes denotes “permission” and sometimes denotes “possibility.” Random House Webster’s Unabridged Dictionary 1189 (2d ed. 2001). The Court reads the statement in Tapia to mean permission : a judge may not consider retribution because that is not allowed. But the statement could also mean “possibility ”: a judge, as a matter of discretion, may choose not to take retribution into account. It is debatable which reading is better, and in any event, as the opinion stated twice, the statement was merely “statutory background.” 564 U. S., at 323, 326. Moreover, Tapia “dealt with a different sentencing law” concerned with imposing imprisonment, not supervised release, and the Court’s drive-by analysis of §3583(c) did nothing to reckon with Congress’s lack of “explicit directions” as to the omitted sentencing factors. 88 F. 4th 1163, 1168 (CA6 2023). Such “[d]ictum settles nothing , even in the court that utters it.” Jama v. Immigration and Customs Enforcement, 543 U. S. 335, 351, n. 12 (2005).

The Court’s other precedent, Concepcion v. United States, 597 U. S. 481, is no better. The holding in that case—that sentencing courts adjudicating motions to reduce a sentence under the First Step Act have discretion to consider intervening changes in the law or relevant facts—is far re-

15

moved from the question here. After emphasizing that sentencing judges have generally exercised broad discretion regarding the facts considered in sentencing determinations, the opinion noted that Congress has sometimes imposed restrictions ; as an illustration, it simply cited the statement made in Tapia. See 597 U. S., at 494.

Whatever weight these dicta merit, they are surely insufficient to support today’s decision.

B

The Court attempts to justify its interpretation by drawing a hard line between the Act’s primary penalties— “[f]ines, probation, and imprisonment”—and supervised release . Ante, at 9. In its view, the former are a court’s main “tools for ensuring that a criminal defendant receives just deserts for the original offense.” Ibid. But “[w]hen a defendant violates the conditions of his supervised release,” the Court reasons, “it makes sense that a court must consider the forward-looking ends of sentencing (deterrence, incapacitation, and rehabilitation), but may not consider the backward-looking purpose of retribution.” Ante, at 10.

This argument exaggerates the distinction between the two classes of penalties to which the Court refers. Contrary to the Court’s suggestion, a decision about the revocation of supervised release may call for consideration of certain backward-looking factors. For example, under §3583(e)(3), the limits of the discretion enjoyed by a court considering the revocation of supervised release depend on the class (A, B, C, or D) of the “statute for the offense that resulted in [the] term of supervised release.” Section 3583(e) also mandates consideration of §3553(a)(6), which, in turn, requires consideration of “the need to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct.” In assessing that factor, a sentencing court must necessarily look back to the defendant’s offense conduct and also consider the

16

past offense conduct of similarly situated defendants.

A defendant’s original offense may also be important when a judge is considering whether a violation of a condition of supervised release is serious enough to justify sending the violator back to prison for a particular length of time. Consider a defendant convicted for the possession of child pornography who later violates a special condition of his release restricting his viewing of any sexually explicit materials, including material that depicts teenagers who are not actually under the age of 18. In such a case, the serious nature of the original crime may be thought to have a bearing on the implications of the later violation and the need for reimprisonment to deter the defendant from future backsliding, to protect society from the consequences of child pornography, or to provide the opportunity for further rehabilitation efforts in prison.

It is revealing that the Court only half-heartedly buys its forward-looking/backward-looking distinction. In a footnote , the Court refuses to take a position on the question whether a sentencing court may revoke supervised release in retribution “for the violation of the conditions of the supervised release.” Ante, at 8, n. 5 (emphasis deleted).

IV

In addition to all these faults, the Court’s interpretation, as I emphasized at the outset of this opinion, defies common sense.

The Court’s interpretation of which sentencing factors a court may consider in altering a term of supervised release must also apply when a court is initially imposing a term of supervised release. The relevant wording in §3583(c), which governs the imposition of supervised release, and §3583(e), which governs the alteration of supervised release , cannot be meaningfully distinguished. But applying the Court’s interpretation to the original imposition of supervised released creates a huge practical problem. A term

17

of imprisonment and a term of supervised release are often imposed at the same time. In imposing a prison term, a court must consider all of the 10 sentencing factors, but under today’s reading of the Sentencing Reform Act, the court is prohibited from considering 2 of those factors when it turns to the issue of supervised release. How this will work in practice is anyone’s guess. Must sentencing courts “adjourn the hearing after imposing” imprisonment and then “start over with a new unblemished inquiry” and without giving any thought to the omitted §3553(a) factors? 88 F. 4th, at 1168. The Sixth Circuit feared that might be required . Many district judges may soon face this problem because “[t]he vast majority (95.1%)” of offenders convicted of a felony or Class A misdemeanor are also “sentenced to terms of supervised release.” United States Sentencing Commission, Federal Offenders Sentenced to Supervised Release 50 (2010). Indeed, in 2024, supervised release was imposed in 50,865 cases. See United States Sentencing Commission, 2024 Sourcebook of Federal Sentencing Statistics 41.

Other daunting practical problems will arise when judges are deciding whether to alter a defendant’s term of supervised release. When sending a violator of supervised release back to prison, a court must carefully control every word that is uttered to explain the decision. Avoiding the language of §3553(a)(2)(A)—“the seriousness of the offense ,” “respect for the law,” and “just punishment for the offense”—is obviously essential. But what about synonyms and other arguably related words and phrases? Not only must sentencing judges avoid verbal missteps, they must do so while heeding the requirement to consider mandatory supervised-release factors that are essentially inseparable from the factors that are fatal. See supra, at 9.

The potential consequences of the Court’s interpretation are so alarming that many of the lower courts that have adopted today’s interpretation have squirmed to avoid its

18

effects. See, e.g., United States v. Miqbel, 444 F. 3d 1173, 1182 (CA9 2006) (“[A] sentence would be unreasonable if the court based it primarily on an omitted factor, such as a factor provided for in §3553(a)(2)(A). . . . We do not suggest that a mere reference to promoting respect for the law would in itself render a sentence unreasonable” (emphasis added)); United States v. Sanchez, 900 F. 3d 678, 684, n. 5 (CA5 2018) (“Of course, this is not to say that any use of words like ‘punish,’ ‘serious,’ or ‘respect’ automatically renders a revocation sentence void. Mere mention of impermissible factors is acceptable; to constitute reversible error . . . the forbidden factor must be ‘dominant’ ” (quoting United States v. Rivera, 784 F. 3d 1012, 1017 (CA5 2015))).

Today’s decision is similar. After adopting an impractical interpretation, the decision takes the remarkable step of outlining ways in which sentencing judges and courts of appeals can avoid strict compliance. At the trial level, if the defense objects that a comment made by the judge crosses the line, the Court counsels that the judge can easily ward off reversal. The judge can “withdraw any impermissible” remark or recast any “stray reference to a §3553(a)(2)(A) factor” as “merely prefatory.” Ante, at 16. In addition, the judge may “explain that a stray reference” to a prohibited factor was actually “intended to bear on” a mandatory factor with which it is closely related. Ibid.

After providing this roadmap for district judges, the Court turns to the courts of appeals. The Court reminds circuit judges that an appeal based on today’s decision “will be governed by plain-error review,” assuming the defendant fails to object in the district court. Ibid. And plain-error review, as we have long observed, is an “exacting” standard and is meant to correct “only particularly egregious errors” that “seriously affect the fairness, integrity, or public reputation of judicial proceedings.” United States v. Young, 470 U. S. 1, 15 (1985) (internal quotation marks omitted).

Reading between the lines, the Court’s message is clear.

19

Unless a district judge obdurately fails to backtrack as recommended when there is an objection based on today’s decision , there should be few problems. The Court’s outline shows that it has a low opinion of the value of its handiwork , and that should seal the case that the whole effort is a mistake.

* * *

For these reasons, I cannot agree with the Court’s decision , and I therefore respectfully dissent.

Footnotes

1. 2The Court argues that these options are not “kinds of sentences” be- cause the Sentencing Reform Act recognizes only three types of sen- tences: probation, fines, and imprisonment. See ante, at 6, n. 4. Accord- ing to the Court, supervised release is not a “ ‘kin[d] of sentence’ ” but is instead merely “a component of a defendant’s prison sentence.” Ibid. But the Sentencing Reform Act never says that there are only three “kinds” of sentences. See §3551(b).
2. 3Here is an example. Suppose a defendant’s offense of conviction was interstate travel in violation of a protective order prohibiting threats, harassment, or contact with his former wife. See 18 U. S. C. §2262(a). And suppose that the trial evidence showed the defendant engaged in conduct that caused his former wife very severe emotional distress. After serving his sentence, the defendant is given supervised release but quickly violates a condition of release by resuming the very conduct that landed him in prison in the first place. The judge must decide what to do. Heeding the requirement to consider “the nature and circumstances of the offense,” the judge calls to mind the trial evidence. Then, as re- quired by the Sentencing Reform Act, the judge considers whether this information suggests that reimprisonment is needed either to “afford ad- equate deterrence,” §3553(a)(2)(B) (i.e., to cause the defendant to leave his former wife alone), or to “protect the public from further crimes of the defendant,” §3553(a)(2)(C) (i.e., to keep the defendant’s former wife safe during the period of reimprisonment). The effect of the Court’s filtering scheme is to complicate the judge’s consideration of all these matters by demanding that he search his con- science and ask: “Have I let any thought of retribution enter into my thinking? I am inclined to send the defendant back to prison for x months, and I think I’m doing that to achieve deterrence and incapacita- tion. But was I somehow influenced by moral disapproval of the defend- ant’s conduct and the feeling that at least part of the x additional months is simply what he deserves?” 4 See §3572(a) (requiring the consideration of all sentencing factors in imposing fines); §3584(b) (same for the decision whether to run sentences concurrently or consecutively). 5 See §3562(a) (requiring the consideration of all sentencing factors “to the extent that they are applicable” in imposing an initial term of proba- tion); §3564(c) (same for early termination of probation); §3565(a) (same for the revocation of probation). 6 Section 3582(a) prohibits consideration of rehabilitation in deciding on a sentence of imprisonment. See Tapia v. United States, 564 U. S. 319, 332 (2011). Similarly, in §3563(b), Congress confined the discretion of sentencing courts to a particular subset of the §3553(a) factors. Under that provision, sentencing courts may set conditions of probation “to the extent that such conditions are reasonably related to the factors set forth in section 3553(a)(1) and (a)(2) and to the extent that such conditions involve only such deprivations of liberty or property as are reasonably necessary for the purposes indicated in section 3553(a)(2).” Congress’s use of limiting language (“to the extent that” and “as are reasonably nec- essary”) makes its restrictive intent clear. Another provision of the Act appears to go even further by removing a sentencing court’s discretion to consider the §3553(a) factors at all. Section 3565(b) provides that a sen- tencing court “shall revoke [a] sentence of probation” if the defendant possesses controlled substances or firearms, refuses to comply with drug testing, or fails a drug test more than three times over the course of a year. No mention of the §3553(a) factors is made. 7 Congress’s decision to omit §3553(a)(2)(A) specifically has its own logic. Congress may have sought to encourage sentencing courts to give greater weight to the remaining purposes of sentencing set forth in §3553(a)(2) to ensure that “[s]upervised release fulfills [its] rehabilitative ends.” United States v. Johnson, 529 U. S. 53, 59 (2000). But in doing so, Congress likely understood that sealing off any discretion to consider §3553(a)(2)(A)’s retributive factors would be impracticable. Supervised release is “a part of ” “a term of imprisonment for” the defendant’s of- fense, §3583(a), and when a sentencing court imposes a term of impris- onment for the offense, it must consider §3553(a)(2)(A), see §3582(a). Congress may have rendered the three §3553(a)(2)(A) factors discre- tionary because it understood that each factor may apply in some, though not all, hearings regarding the alteration of supervised release. The need “to promote respect for law” strikes me as an evergreen concern when a court is considering what remedy best enables a defendant to “learn to obey the conditions of his supervised release.” 88 F. 4th 1163, 1167 (CA6 2023) (internal quotation marks omitted). But the need “to provide just punishment for the offense” seems less relevant when a sentencing court addresses a violation of supervised release. And the relevance of the re- maining §3553(a)(2)(A) factor—the need “to reflect the seriousness of the offense”—depends on context. If, for example, a sentencing court decides to revoke supervised release following a defendant’s violation, it must, as already explained, examine how Congress categorized the offense’s “seriousness” to determine the authorized term of reimprisonment. See §3583(e)(3). But if the court instead opts for a less-severe sanction, such as the modification of a condition of release, it is not necessarily required to consider how Congress scored the underlying offense’s “seriousness.” See §3583(e)(2).