Pulsifer v. United States

P R E L I M I N A R Y P R I N T Volume 601 U. S. Part 1 Pages 124–186 O F F I C I A L R E P O R T S

T H E S U P R E M E C O U R T March 15, 2024

reporter of decisions

N O T I C E : This preliminary print is subject to formal revision before the bound volume is published. Users 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.

Syllabus

certiorari to the united states court of appeals for the eighth circuit No. 22 – 340. Argued October 2, 2023 —Decided March 15, 2024 Af ter pleading guilty to distributing at least 50 grams of methamphet- amine, petitioner Mark Pulsifer faced a mandatory minimum sentence of 15 years in prison. At sentencing, he sought to take advantage of the “safety valve” provision of federal sentencing law, which allows a sentencing court to disregard the statutory minimum if a defendant meets fve criteria. Among those is the requirement, set out in Para- graph (f )(1), that the sentencing court fnd that —

(1) the defendant does not have—

(A) more than 4 criminal history points, excluding any criminal history points resulting from a 1-point offense, as determined under the sentencing guidelines; (B) a prior 3-point offense, as determined under the sentencing guidelines; and (C) a prior 2-point violent offense, as determined under the sentenc- ing guidelines.

The Government argued that Pulsifer could not satisfy that requirement because he had two prior three-point offenses totaling six criminal- history points. In the Government's view, each of those prior offenses disqualifed him under Subparagraph B and the six total points disquali- fed him under Subparagraph A. But Pulsifer claimed he remained elig ible. He pointed out that his criminal record lacked a two-point violent offense, as specifed in Subparagraph C. And in his view, only the combination of the items listed in the subparagraphs could prevent him from getting safety-valve relief. The District Court agreed with the Government, and the Eighth Circuit affrmed.

Held: A defendant facing a mandatory minimum sentence is elig ible for safety-valve relief under 18 U. S. C. § 3553(f )(1) only if he satisfes each of the provision's three conditions—or said more specifcally, only if he does not have more than four criminal-history points, does not have a prior three-point offense, and does not have a prior two-point violent offense. Pp. 132 –153. (a) Each party offers a grammatically permissible way to read Para- graph (f )(1). Under Pulsifer 's reading, the word “and” joins three fea- tures of a defendant's criminal history into a single disqualifying charac-

Cite as: 601 U. S. 124 (2024) 125 Syllabus

teristic; accordingly, a defendant is inelig ible for the safety valve only if he has the items listed in Subparagraphs A, B, and C in combination.

In the Government's view, “and” connects three criminal-history condi- tions, all of which must be satisfed to gain safety-valve relief. In other words, the court must fnd the defendant does not have A, does not have B, and does not have C. Each of those readings is possible in the ab- stract. The choice between the two can sensibly be made only by exam- ining the content of Paragraph(f )(1)'s three subparagraphs, including what they say, how they relate to each other, and how they ft with other pertinent law. Pp. 132 –141. ( b) The text and context of Paragraph (f )(1), as read against the Guidelines, yield just one plausible statutory constr uction. The para- graph creates an elig ibility checklist, and specifes three necessary con- ditions for safety-valve relief. Reading the paragraph as Pulsifer does to set out a single condition—i. e., that the defendant not have the co m- b ina t io n of the characteristics listed in Subparagraphs A, B, and C— would create two statutory diffculties that the Government's reading does not. Pp. 141–149.

(1) Pulsifer 's reading would render Subparagraph A superfuous

because a defendant who has a three-point offense under Subparagraph B and a two-point offense under Subparagraph C will always have more than four criminal-history points under Subparagraph A. That reading leaves Subparagraph A with no work to do: removing it from the statute would make the exact same people elig ible (and inelig ible) for relief.

That kind of superfuity, in and of itself, refutes Pulsifer 's reading.

When a statutory constr uction “render[s] an entire subparagraph mean- ingless, ” this Court has noted, the canon against surplusage applies with special force. Na t io nal Assn. o f Mf rs. v. Depar tmen t o f De fense, 583

U. S. 109, 128. That is particularly tr ue when, as here, the subpara-

graph is so evidently desig ned to serve a concrete function. Pp. 142 –146.

(2) Pulsifer 's reading would also create a second problem related to

Paragraph (f )(1)'s gatekeeping function. The Guidelines presume that defendants with worse criminal records— exhibiting recidivism, lengthy sentences, and violence— deserve greater punishment. Under the Gov- e r n m e n t 's r e a d i n g, P a r a g r a p h ( f ) ( 1 ) s o r t s d e f e n d a n t s a c c o r d i n g l y.

When the defendant has committed multiple non-minor offenses, he can- not get relief (Subparagraph A). And so too when he has committed even a single serious offense punished with a lengthy prison sentence (Subparagraph B) or one involving violence (Subparagraph C). Pulsi- fer 's reading, by contrast, would allow safety-valve relief to defendants with more serious records while barring relief to defendants with less serious ones. A defendant with a three-point offense and a two-point

Syllabus

violent offense would be denied relief. But a defendant with multiple three-point violent offenses could get relief simply because he happens not to have a two-point violent offense.

Contrary to Pulsifer 's view, that anomalous result cannot be ig nored on the ground that a sentencing judge retains discretion to impose a lengthy sentence. If Congress thought it could always rely on sentenc- ing discretion, it would not have created a criminal-history requirement in the frst instance. Instead, it specifed a requirement that allows such discretion to operate only if a defendant's record does not reach a certain level of seriousness. Pulsifer 's constr uction of Paragraph (f )(1) makes a hash of that gatekeeping function. Pp. 146 –149. (c) The uncontested fact that Congress amended Paragraph (f )(1) as part of the First Step Act to make safety-valve relief more widely avail- able does not assist in interpreting the statutory text here. Both par- ties' views of the paragraph widen the opportunity for safety-valve re- lief, and Pulsifer 's interpretation is not better just because it would allow more relief than the Government's. “ [N]o law pursues its . . .

purpose[s] at all costs. ” Luna Perez v. Sturg is Publ ic Schoo ls, 598

U. S. 142, 150. Here, where Congress did not eliminate but only cur-

tailed mandatory minimums, the Court can do no better than examin- ing Paragraph (f )(1)'s text in context to determine the exact contours of the defendants to whom Congress extended safety-valve relief.

Pp. 151–152. (d) The Court re jects Pulsifer 's efforts to invoke the r ule of lenity.

Lenity applies only when a statute is genuinely ambiguous. For the reasons explained above, although there are two grammatically permis- sible readings of Paragraph (f )(1), in context its text is susceptible of only one possible constr uction. That leaves no role for lenity to play.

Pp. 152 –153.

39 F. 4th 1018, affrmed.

Kagan, J., delivered the opinion of the Court, in which Roberts, C. J., and Thomas, Alito, Kavanaugh, and Barrett, JJ., joined. Gorsuch, J., fled a dissenting opinion, in which Sotomayor and Jackson, JJ., joined, post, p. 155.

Shay D vo re tzky arg ued the cause for petiti oner. With h im on the br iefs were Parke r Ri de r-Lo ngmai d, Kyse r Bl akely, and J. Robe r t Bl ack.

Frede r ick Liu arg ued the cause for the United States.

With h im on the br ief were So l ici t o r Gene ral Prel ogar, As- Cite as: 601 U. S. 124 (2024) 127

Opinion of the Court

sistan t At t o r ney Gene ral Po l i te, Depu ty So l ici t o r Gene ral Fe i g in, and Francesco Valen t in i.*

Justice Kagan delivered the opini on of the Cour t.

The “safety va lve” provisi on of federa l sentencing law ex- empts cer tain defendants from mandatory minimum pena l- ties, thus enabling cour ts to g ive them lighter pr ison ter ms.

To qua lify for safety-va lve relief, a defendant must meet var- i ous cr iter ia, one of wh ich addresses h is cr imina l h istory.

That cr iter i on, in stylized for m, requires that a defendant “does not have A, B, and C”—where A, B, and C refer to three ways in wh ich past cr imina lity may suggest future dangerousness and therefore warrant a more severe sen- tence. In br ief (w ith details below), A, B, and C are “more than 4 cr imina l h istory points, ” a “ 3-point offense, ” and a “ 2-point vi olent offense. ” The questi on presented is how to understand the cr imina l- h istory requirement. The Gover nment contends that the phrase “does not have A, B, and C” creates a checklist w ith three distinct conditi ons. On that view, a defendant meets the requirement (and so is elig ible for safety-va lve relief ) if he does not have A, does not have B, and does not have C.

Or stated conversely, a person fails to meet the requirement (and so cannot get relief ) if he has any one of the three. The petiti oner here instead contends that the phrase “does not have A, B, and C” sets out a sing le, ama lgamated conditi on for relief. On h is reading, a defendant meets the require- ment (and is elig ible for relief ) so long as he does not have

*Briefs of amici cur i ae urg ing reversal were fled for the Americans for Prosperity Foundation by Michael Pepso n; for FAMM et. al by Dav id Debo ld, Je ff rey T. G reen, Dav id D. Co le, Brando n Buskey, Mar y Pr ice, and Peter Go ldberger; and for the National Association of Federal Defend- ers by Vincen t J. Br unkow and Michael Ho lley.

Ph i l ip D. Wi ll i amso n fled a brief for Thomas R. Lee et al. as amici cur i ae.

Opinion of the Court

the co mb in a t io n of A, B, and C. Or put conversely, he fails to meet the requirement (and cannot get relief ) only when he has a ll three. Today, we agree w ith the Gover nment's view of the cr imina l-h istory provisi on.

Congress sometimes establishes mandatory minimum pen- a lties for cr imes, including dr ug offenses. Those provisi ons put a lower limit on a cour t's sentencing discreti on, refecting Congress's judg ment that specifed conduct demands no less than a specifed punishment. For dr ug offenses of the k ind involved here, the existence and leng th of minimum pena lties typica lly depend on the type and quantity of the dr ug at issue, the har m resu lting from the cr ime, and (relevant here) the defendant's cr imina l h istory.

The safety-va lve provisi on, 18 U. S. C. § 3553(f ), offers some defendants convicted of dr ug offenses an escape from other w ise applicable mandatory minimums. Under the pro- visi on, a cour t is to sentence a defendant “w ithout regard to any statutory minimum” if it fnds that fve cr iter ia are met.

Ib i d. Three of the cr iter ia focus on character istics of the offense—in br ief, whether the defendant used vi olence; whether the cr ime resu lted in death or ser i ous in jury; and whether the defendant acted as a r ing leader. See § 3553(f )(2)–(4). One of the cr iter ia addresses the defend- ant's cooperati on w ith the Gover nment. See § 3553(f )(5).

And one—the frst listed and the most relevant here— concer ns the defendant's cr imina l h istory. See § 3553(f )(1).

The complete text of the safety-va lve provisi on is set out in th is opini on's appendix.

The cr imina l-h istory requirement —we usua lly ca ll it Para- graph (f )(1)—recently under went a substantia l revisi on, mak ing it easier for a defendant to meet. As or ig ina lly enacted, the paragraph limited safety-va lve relief to defend- ants who “d[id] not have more than 1 cr imina l h istory point, as deter mined under the sentencing g uidelines. ” Vi o-

Cite as: 601 U. S. 124 (2024) 129 Opinion of the Court

lent Cr ime Control and Law Enforcement Act of 1994, 108 Stat. 1985. What that meant in practice—we explain why just below—was that anyth ing more than a sing le minor cr ime barred a defendant from gaining relief. But that is no longer so. In the First Step Act of 2018, Congress re- laxed the safety-va lve provisi on's cr imina l-h istory require- ment, enabling defendants w ith more sig nifcant cr imina l records to qua lify. Pub. L. 115 – 391, 132 Stat. 5221. Today, Paragraph (f )(1) is met if “ the cour t fnds at sentencing ”

the defendant does not have— (A) more than 4 cr imina l h istory points, excluding any cr imina l h istory points resu lting from a 1-point of- fense, as deter mined under the sentencing g uidelines; (B) a pr i or 3-point offense, as deter mined under the sentencing g uidelines; and (C) a pr i or 2-point vi olent offense, as deter mined under the sentencing g uidelines.

And if the defendant a lso meets Secti on 3553(f )'s other four cr iter ia, he becomes exempt from a statutory minimum.

As the text makes clear, the new Paragraph (f )(1) ( like the old one) tur ns on the defendant's cr imina l-h istory points under the Guidelines. In genera l, the sever ity of Guidelines sentencing recommendati ons increases w ith the number of cr imina l-h istory points the defendant has (of ten ca lled h is cr imina l-h istory score). And the Guidelines assig n more points to more ser i ous pr i or offenses. There is a caveat to that r u le, wh ich w ill become per tinent later. See in f ra, at 143 –146. Some pr i or convicti ons, even if for ser i ous offenses, do not add any points to a defendant's score. That is tr ue, for example, if the convicti on is quite old or if it was rendered by a foreig n cour t. See U. S. Sentencing Guidelines (USSG) § 4A1.2(e)(3), ( h). But putting such excepti ons aside, convic- ti ons resu lting in longer pr ison sentences add more points to a defendant's tota l. As previewed above, the Guidelines

Opinion of the Court

award one point for minor offenses—specifca lly, for those resu lting in sentences of less than 60 days. In the new

Because of the “excluding ” phrase, they cannot, either a lone or in combinati on, prevent a defendant from gaining safety- va lve relief. But longer sentences, generating more points, fa ll w ith in the paragraph's notice. A pr ison sentence of be- tween 60 days and 13 months ear ns two points under the Guidelines. See § 4A1.1( b). So a convicti on punished w ith that sentence w ill count, for purposes of the paragraph, as a “pr i or 2-point offense, ” assuming it is a lso “vi olent” (as de- fned in a nearby provisi on). § 3553(f )(1)(C), (g). Moving up another notch, a sentence exceeding 13 months ear ns three points under the Guidelines. See § 4A1.1(a). So a convicti on g iving r ise to that greater pena lty (whether or not vi olent) w ill qua lify under the paragraph as a “pr i or 3-point offense. ” § 3553(f )(1)(B).

Th is case involves a dispute about whether Paragraph (f )(1) bars petiti oner Mark Pu lsifer from gaining safety-va lve relief. Pu lsifer pleaded g uilty in 2020 to distr ibuting at least 50 grams of methamphetamine. He faced a mandatory minimum of 15 years in pr ison unless the safety-va lve provi- si on came to h is aid. The Gover nment claimed it did not because Pu lsifer cou ld not meet its cr imina l-h istory require- ment. Pu lsifer had two relevant pr i or convicti ons, each for a three-point offense. In the Gover nment's view, that fact disqua lifed Pu lsifer from obtaining relief severa l times over.

He had not just one but two “pr i or 3-point offense[s], ” as specifed in Subparagraph B of the requirement. And be- cause three plus three equa ls six, he a lso had “more than 4 cr imina l h istory points, ” as specifed in Subparagraph A.

But Pu lsifer claimed that was still not enough. He pointed out that h is cr imina l record lacked a “ 2-point vi olent of- fense, ” as specifed in Subparagraph C. And in h is view, only the combinati on of the items listed in the three subpara-

Cite as: 601 U. S. 124 (2024) 131 Opinion of the Court

graphs—the fu ll package, as it were—cou ld prevent h im from getting safety-va lve relief.

The Distr ict Cour t re jected Pu lsifer 's arg ument, r u ling that a defendant is “ inelig ible for safety va lve” relief if he has any of the “ three th ings” specifed in Paragraph (f )(1).

App. to Pet. for Cer t. 35a – 36a. The mandatory minimum, the cour t concluded, thus applied to Pu lsifer 's sentence.

The Cour t of Appea ls for the Eighth Circuit affr med.

The cour t framed the questi on as “ in what sense the statute uses the word `and. ' ” 39 F. 4th 1018, 1021 (2022). In the abstract, the cour t stated, the phrase “ the defendant does not have (A), (B), and (C)” might be read in two different ways. It cou ld mean that the defendant does not have the combinati on of the “ three elements listed in (A), (B), and (C), ” as Pu lsifer urged. Ib i d. Or it cou ld mean, as the Gov- er nment arg ued, that the defendant does not have every one of those elements—in other words, that he does not have (A), does not have (B), and does not have (C). In choosing between those readings, the cour t found a “strong textua l basis” to prefer the Gover nment's. Ib i d. If Pu lsifer were r ight, the cour t explained, Subparagraph A wou ld be “ren- dered superfuous”—w ithout the slightest effect. Ib i d. “A defendant who has a pr i or three-point offense under [Sub- paragraph B] and a pr i or two-point vi olent offense under [Subparagraph C] wou ld alw ays meet the cr iter i on in [Sub- paragraph A], because he wou ld a lways have more than four cr imina l h istory points. ” Ib i d. That was reason enough to read Paragraph (f )(1) the other way—as an “elig ibility check- list” of three distinct conditi ons, each of wh ich the defendant must meet to qua lify for safety-va lve relief. Id., at 1022.

And on that view, the cour t concluded, Pu lsifer cou ld not escape a mandatory minimum: Because he had a pair of three-point offenses, it was simply “ immater ia l” that he did not a lso “ have a pr i or two-point vi olent offense. ” Id., at 1022 –1023.

Opinion of the Court

We granted cer ti orar i, 598 U. S. ––– (2023), because the Cour ts of Appea ls have split over how to read the safety- va lve provisi on's cr imina l-h istory requirement.1 Today, we adopt the Gover nment's view, and so affr m the decisi on below. A defendant is elig ible for safety-va lve relief under Paragraph (f )(1) only if he “does not have” a ll three of the items listed—or said more specifca lly, does not have four cr imina l-h istory points, does not have a pr i or three-point of- fense, and does not have a pr i or two-point vi olent offense.

The paragraph thus creates an elig ibility checklist, and de- mands that a defendant satisfy every one of its conditi ons.

We star t w ith Paragraph (f )(1)'s grammatica l str ucture, because Pu lsifer 's main arg ument (and initia lly the dissent's) is that it resolves th is case. See Br ief for Pu lsifer 16 – 20; post, at 161–162 (opini on of Gorsuch, J.). 2 Reca ll that the paragraph requires a cour t to fnd that the defendant does not have the features specifed in Subparagraphs A, B, and

C. “Because Congress used `and' to connect” those subpara-

graphs, Pu lsifer contends, “a defendant is inelig ible” for safety-va lve relief “only if he has the complete combo”—i. e., more than four cr imina l-h istory points plus a pr i or three- point offense plus a pr i or two-point vi olent one. Br ief for Pu lsifer 19. That resu lt follows, Pu lsifer claims, simply

1 Compare 39 F. 4th 1018 (CA8 2022) (case below) (holding that a defend- ant is elig ible for relief only if he does not have all three of the items listed); Un i ted Sta tes v. Palo mares, 52 F. 4th 640 (CA5 2022) (same); Un i ted Sta tes v. Haynes, 55 F. 4th 1075 (CA6 2022) (same); Un i ted Sta tes

v. Pace, 48 F. 4th 741 (CA7 2022) (same), with Un i ted Sta tes v. Jo nes, 60

F. 4th 230 (CA4 2023) (holding that a defendant is elig ible for relief so long

as he does not have any one of the items listed); Un i ted Sta tes v. Lopez, 998 F. 3d 431 (CA9 2021) (same); Un i ted Sta tes v. Garco n, 54 F. 4th 1274 (CA11 2022) (en banc) (same).

2 As later noted, in f ra, at 140, the dissent ultimately comes around to the view that the meaning of Paragraph (f )(1) depends on context, see post, at 168 –169.

Cite as: 601 U. S. 124 (2024) 133 Opinion of the Court from “what ordinary grammar says. ” Tr. of Ora l Arg. 3.

But in fact grammar does not say so much. There are two grammatica lly per missible ways to read Paragraph (f )(1).

Yes, one is Pu lsifer 's. But the other is the Gover nment's— that a defendant is inelig ible for relief unless he can satisfy each of the paragraph's three conditi ons. The choice be- tween the two, as th is Par t shows, is not a matter of gram- matica l r u les. It can sensibly be made only by examining, as the next Par t does, the paragraph's content, as read in con juncti on w ith the Guidelines. Or, as we usua lly say in statutory constr ucti on cases, by review ing text in context.

“And, ” in grammatica l ter ms, is of course a con juncti on— a word whose functi on is to connect specifed items. Both par ties here agree w ith that elementary propositi on. See Br ief for Pu lsifer 18; Br ief for United States 14. The word “and, ” each might say, means . . . well, and. Indeed, to the extent elaborati on is needed, both par ties select the same defniti on from the same dicti onary. “And, ” they recite in concer t, means “a long w ith or together w ith. ” Webster 's Th ird New Inter nati ona l Dicti onary 80 (1993); see Br ief for Pu lsifer 18; Br ief for United States 14.

Where th ings get more complicated is in fg ur ing out what goes a long or together w ith what —or other w ise said, what the “and” in Paragraph (f )(1) connects. As Pu lsifer reads the paragraph, the “and” joins three features of a defendant's cr imina l h istory into a sing le disqua lify ing character istic.

The con juncti on of Subparagraphs A, B, and C produces the th ing he labels “ the complete combo”; the questi on then be- comes whether the defendant has or “does not have” that fu ll package. Br ief for Pu lsifer 19; § 3553(f )(1). Some grade- school math notati on may help revea l the proposed order ing.

It is as if Pu lsifer inser ted parentheses into the paragraph, so that it asks whether “ the defendant does not have (A, B, and C). ” Much as a student wou ld solve “5 - (2 + 1)” by frst adding 2 and 1 and then subtracting the sum from 5, so Pu l- sifer wants a cour t frst to combine A, B, and C and then to

Opinion of the Court

deter mine whether the defendant has the tota l. By con- trast, the Gover nment reads the statute w ithout parenthe- ses, and so arr ives at a different conclusi on. On its view, the “does not have” lang uage operates on A, and on B, and on C consecutively, rather than on the three combined. So the “and” connects three cr imina l-h istory conditi ons, a ll of wh ich must be satisfed to gain safety-va lve relief. Or said another way, Paragraph (f )(1) requires that the defendant does not have A, and a lso does not have B, and fna lly does not have C. If he has even one, he cannot complete the req- uisite checklist and so cannot gain the safety va lve's benefts.

The Gover nment's view rests on a routine aspect of expressi on—that an introductory phrase ( here, “does not have”) may apply to, or modify, severa l ter ms coming af ter it, one by one by one. Suppose a person says af ter visiting a bookstore, “I bought a novel, a memoir, and a travel g uide. ” That is just a more effcient way of say ing “I bought a novel, bought a memoir, and bought a travel g uide. ” The verb in the sentence carr ies over—some grammar ians use the ter m “distr ibut[es] ”—to every item on the ensuing list. B. Gar- ner, Dicti onary of Lega l Usage 639 (3d ed. 2011). That prac- tice is pervasive, indeed inescapable, in every k ind of speech and wr iting. Consider th is, perhaps ha lf-remembered line from ch ildhood: “On Saturday he ate through one piece of chocolate cake, one ice-cream cone, one pickle, one slice of Sw iss cheese, one slice of sa lami, one lollipop, one piece of cherry pie, one sausage, one cupcake, and one slice of wa- ter melon. ” E. Car le, The Very Hungry Caterpillar 15 –16 (2018). The introductory words “ate through” apply inde- pendently and equiva lently to each of the ten foodstuffs that follow. Or if that example seems too tr ifing, take a couple from the Constituti on. Ar ticle III provides that “ [t]he judi- cia l Power sha ll extend to a ll Cases . . . ar ising under th is Constituti on, the Laws of the United States, and Treaties. ” § 2. That statement means—but says more concisely—that the judicia l power extends to cases ar ising under the Consti- tuti on; extends to cases ar ising under federa l law; and ex-

Cite as: 601 U. S. 124 (2024) 135 Opinion of the Court

tends to cases ar ising under treaties. The provisi on does not (as Pu lsifer 's view might suggest) limit judges to hear ing the few cases ar ising simu ltaneously under a ll three k inds of law. Similar ly, Ar ticle I of the Constituti on enables Con- gress “ [t]o reg u late Commerce w ith foreig n Nati ons, and among the severa l States, and w ith the Indian Tr ibes. ” § 8, cl. 3. That author izati on goes to commerce involving each k ind of entity, not just to commerce involving the three at once. So again, the verb phrase operates on each ter m se r i- a t im, not on the combinati on of the three. 3 Pu lsifer claims that verb phrases do not work the same way when “framed in the negative. ” Reply Br ief 2. One of h is favor ite examples is “don't dr ink and dr ive. ” Br ief for Pu lsifer 16. That “doesn't mean, ” he observes, “ that you shou ldn't dr ink and that you shou ldn't dr ive, but only [means] that you shou ldn't do both at the same time. ” Id., at 18 (emphasis deleted). So too, he says, for “don't clean the bathroom w ith bleach and ammonia. ” Id., at 12 (empha- sis deleted). The proh ibiti on does not go to bleach a lone and to ammonia a lone; instead, it goes only to the two in con juncti on. To retur n to math notati on, the statement is best understood as “don't clean w ith ( bleach and ammonia), ” rather than “don't clean w ith bleach and don't clean w ith am-

3 The dissent wrongly views this ordinary feature of language as a kind of uncanny trick. To understand a verb as applying to each of several ensuing terms (the dissent says) is to choose verbal “gymnastics” over “natural” meaning. Post, at 164. The dissent's primary proof is that such phrases can be rendered “with deleted words stricken and new ones added in bold. ” Post, at 163. Well, yes, but so what? It is tr ue, as the dissent might say, that “I bought (1) a novel, (2) a memoir, and (3) a travel guide” is equivalent in meaning to “I bought (1) bou g ht a novel, (2) bou g ht a memoir, and (3) bou g ht a travel guide. ” Cf. post, at 163 –164 (similarly representing Paragraph (f )(1)). But ordinary people still understand the verb to carry over to all the books in the sentence. The strikeouts and boldface, far from evidencing manipulation of meaning, just illustrate how expression can naturally work. And that is so, contra the dissent (post, at 167), when a sentence's subject is singular (rather than plural)—as shown by most of the sentences in the paragraph above, many similar ones to come, see in f ra, at 136, and the vast number a reader can make up on her own.

Opinion of the Court

Paragraph (f )(1) conditi ons relief on a cour t's fnding that a defendant “does not have (A, B, and C), ” rather than that he “does not have A, does not have B, and does not have C. ” But for every negative statement Pu lsifer offers up, an- other cuts the opposite way (suggesting, as we later discuss, that here grammar is not the pr imary deter minant of mean- ing). Consider two sentences discussed in The Cambr idge Grammar of the Eng lish Lang uage. If someone says “I'm not free on Saturday and Sunday, ” the Grammar notes, he most likely means “I'm not free on Saturday and I'm not free on Sunday”; he is not say ing that a lthough he cannot go away for a fu ll weekend, he can make plans on one of those days.

See R. Huddleston & G. Pu llum 1298 –1299 (2002) (emphasis deleted). Similar ly, if a person says, “I didn't like h is mother and father, ” he probably means “I didn't like h is mother and I didn't like h is father ”—not that he didn't like the two in combinati on, but thought that either a lone was fne. Ib i d. (emphasis deleted).4 Or take an example raised in ora l arg ument per taining, like Paragraph (f )(1), to an elig ibility requirement: A hospita l tells you that it can per- for m a medica l procedure only if you “don't eat, dr ink, and smoke for the preceding 12 hours. ” See Tr. of Ora l Arg.

6 – 8. Even Pu lsifer 's counsel agreed that he wou ld not feel free to have a steak and mar tini so long as he abstained from tobacco. See ib i d. The “don't” here, unlike in Pu lsif- er 's examples, carr ies over to each acti on on the list (eating, dr ink ing, and smok ing a like)—not just to the three in tandem.

And if those examples of negatively framed statements, both Pu lsifer 's and ours, seem a tad conversati ona l, consider

4 So too, a manual of contract draf ting observes that “ [t]he more natural meaning ” of “Acme shall not notify Able and Baker ” is “Acme shall not notify Able and shall not notify Baker, ” not that he shall not notify the two together, but may notify either one. K. Adams, A Manual of Style for Contract Draf ting § 11.16, p. 212 (3d ed. 2013) (emphasis deleted).

Cite as: 601 U. S. 124 (2024) 137 Opinion of the Court a statute str ik ing ly similar in for m to Paragraph (f )(1).

First retur n to that paragraph to remind yourself of how it looks and reads. See supra, at 129. Now check out 34

As used in th is secti on, the ter m “offenses against the United States” does not include—

(1) a cr imina l vi olati on of the Unifor m Code of Mili-

tary Justice (10 U. S. C. 801 et seq.);

(2) an offense against the laws of the Distr ict of Co-

lumbia; and

(3) an offense tr iable by an Indian tr iba l cour t or

Cour t of Indian Offenses.

The “does not include” lang uage at the top of course refers independently to cr imes satisfy ing (1), cr imes satisfy ing (2), and cr imes satisfy ing (3)—not to whatever cr imes manage to satisfy (1), (2), and (3) a ll at once. Or said other w ise, the statute means exactly what it wou ld mean if Congress had str ipped the phrase “does not include” from the prefatory line and repeated it three times in the subsequent list. Con- gress, we recog nize, just opted to draf t more concisely. And so too it cou ld have made that choice in draf ting Paragraph (f )(1)—w ith the “does not have” phrase referr ing to every item that follows. No grammatica l pr inciple precludes that understanding of what Congress wrote.

Pu lsifer protests that using the word “or ” ( instead of “and”) wou ld have better conveyed the Gover nment's read- ing, but that claim a lso fails. His basic objecti on (echoed in the dissent, see post, at 170 –171) is that Congress cou ld have expressed its intent more clear ly. “If the gover nment is r ight” about Paragraph (f )(1)'s meaning, Pu lsifer asks, “why didn't Congress just use `or '?”; doing so wou ld have shown “unequivoca lly” that a defendant must meet a ll three of the specifed conditi ons. Reply Br ief 16; Tr. of Ora l Arg. 10.

But to beg in w ith, we do not demand (or in tr uth expect) that Congress draf t in the most translucent way possible.

Opinion of the Court

We have “routinely constr ued statutes to have a par ticu lar meaning even as we acknowledged that Congress cou ld have expressed itself more clear ly. ” Lun a To r res v. L ynch, 578

U. S. 452, 472 (2016) (citing cases). And anyway, we doubt

that substituting “or ” for “and” wou ld have delivered us from interpretive controversy. Instead, we wou ld likely have confronted the mirror image of the dispute before us.

The Gover nment wou ld have read the requirement that a defendant “does not have A, B, or C” to mean that he “does not have (A, B, or C). ” So a defendant wou ld get safety- va lve relief only if he doesn't have any of the three listed cr imina l-h istory features. But Pu lsifer, we suspect, wou ld have read the same requirement to mean that a defendant “does not have A, does not have B, or does not have C. ” So he wou ld get safety-va lve relief as long as he doesn't have a sing le one of the listed features. That reading too is possi- ble when viewed only as a matter of abstract grammar, di- vorced from any ana lysis of A, B, and C's content. Even w ith Pu lsifer 's proposed redraf ting, then, the grammatica l back-and-for th wou ld continue.5 In fact, we can see why a Congress w ish ing to express the Gover nment's view might have chosen to use “and. ” Sup- pose that before putting words to the page, Congress had

5 That is not to say, of course, that any negative statement involving the word “or ” is realistically capable of two meanings; the point is only that context may drive such a statement in either direction. Consider two examples. First, suppose a restaurant chef decides to buy broccoli if his supplier “does not have spinach, eggplant, or caulifower. ” That most likely means the chef will buy broccoli only when the supplier is out of all three other vegetables, not when he is out of just one. But second, sup- pose the same chef typically places a food order if the restaurant “does not have meat, produce, or bread. ” That most likely means he'll place an order when the restaurant r uns out of one of those foodstuffs, not wait until it is lacking all three. The grammar in the two statements is identi- cal, but their most natural understanding is not. Here, content drives meaning, so that in the one sentence, the absence of three items—and in the other sentence, the absence of one item—triggers the relevant purchase.

Cite as: 601 U. S. 124 (2024) 139 Opinion of the Court

decided (as the Gover nment says) to create an elig ibility checklist, requir ing a defendant to meet three distinct condi- ti ons before getting safety-va lve relief. In the subse- quent draf ting process, an “and” cou ld well have seemed in- tuitive. Af ter a ll, on the Gover nment's “checklist” view, a defendant must meet every one of three conditi ons—th is one and th is one and th is one. Or said more concretely, the defendant must not have “more than 4 cr imina l h istory points” and must not have a “ 3-point offense” and must not have a “ 2-point vi olent offense. ” So why not use an . . . “and”? It serves to connect the three necessary conditi ons coming off the (effcient) prefatory lang uage. In other words, Congress might have thought that use of the con junc- tive word “and” wou ld refect the needed con juncti on of three requirements.

Consider, as a summary of a ll these points, Pu lsifer 's own main example, because it shows why Paragraph (f )(1)'s gram- matica l str ucture cannot decide th is case—and points to the k ind of ana lysis needed instead. Pu lsifer offers a college

All student-ath letes are elig ible for an academic scholar- sh ip, provided that the student dur ing the previ ous se- mester did not — (A) miss more than fve classes; (B) fail to submit a paper in the semester ly, campus- w ide wr iting competiti on; and (C) ear n less than a 3.0 GPA.

Br ief for Pu lsifer 19 (emphasis deleted). In Pu lsifer 's view, the policy is clear: A student may retain h is scholarsh ip un- less he funks “a ll three” of the conditi ons. Ib i d. So a stu- dent, Pu lsifer contends, is in the clear if he “submitted a paper in the wr iting competiti on and ear ned a 3.4 GPA . . . even though he missed seven classes. ” Ib i d. Wh ich sounds reasonable enough. But how about th is one: A student who misses only four classes, but fails to submit a competiti on

Opinion of the Court

paper and “ear ns” a 1.0 GPA. Or similar ly, a student who submits a (terr ible) paper, wh ile missing a ll h is classes and obtaining the same “D” average. Is it now so clear that the policy a llows a student to funk two of the conditi ons, rather than requir ing h im to satisfy a ll three? Or is it, at the least,

All student-ath letes are elig ible for an academic scholar- sh ip, provided that the student dur ing the previ ous se- mester did not — (A) fail a course; (B) commit plag iar ism; and (C) get arrested.

A student wou ld need a lot of confdence to arg ue that he remains scholarsh ip-elig ible when he (A) failed a course, and (B) committed plag iar ism, but (C) managed to evade arrest.

That reading—Pu lsifer 's reading—is grammatica lly possible.

But so too is the opposite—that a student must meet a ll three conditi ons. And when we th ink about the content of the policy—what (A), (B), and (C) actua lly say—against the backdrop of a ll we know (or perchance a ll the college hand- book tells us) about academic scholarsh ips, we cannot read the revised hypothetica l in Pu lsifer 's way.

The takeaway is th is: Paragraph (f )(1) cannot be constr ued in the abstract, as if a ll a reader has to go on is the str ipped- down phrase “ the defendant does not have A, B, and C. ” That might require the defendant not to have (A, B, and C)—

i. e., the combinati on of the three. Or it might require the

defendant not to have A, and not to have B, and not to have C—i. e., each of the three. Rea lly, it a ll depends. (It is no- table that even the dissent must in the end concede the point, noting that whether a speaker “ intend[s] for a listener to distr ibute words implicitly” depends on the context. See post, at 168; supra, at 132, and n. 2). The way a reader as- sig ns meaning to the phrase is to look at the substance of A, B, and C—the items on the list and the way they interact,

Cite as: 601 U. S. 124 (2024) 141 Opinion of the Court

as against relevant background understandings. Reca ll a couple of examples. See supra, at 135 –136. We interpret the in juncti on against dr ink ing and dr iving in Pu lsifer 's way—“do not (A and B)”—because the two activities are usua lly per ilous only in combinati on. We interpret the in- juncti on against eating and dr ink ing before surgery in the Gover nment's way—“do not A and do not B”—because each activity a lone is likely to have adverse consequence. Simi- lar ly here, the meaning of Paragraph (f )(1) may become clear if we examine the content of its three subparagraphs—what they say and how they relate to each other—as well as how they ft w ith other per tinent law. Or stated in the usua l lang uage of statutory constr ucti on, the answer may lie in consider ing the paragraph's text in its lega l context.

And indeed, that inquiry into text and context makes Para- graph (f )(1)'s meaning clear. The paragraph creates an eli- g ibility checklist. It specifes three necessary conditi ons for safety-va lve relief—that the defendant not have more than four cr imina l-h istory points, not have a pr i or three-point of- fense, and not have a pr i or two-point vi olent offense. Read- ing the paragraph instead to set out a sing le conditi on—that the defendant not have the co mb in a t io n of the listed char- acter istics—wou ld create two statutory diffcu lties. First, Subparagraph A wou ld become superfuous—w ithout any op- erative sig nifcance. That is because if a defendant has a three-point offense under Subparagraph B and a two-point offense under Subparagraph C, he w ill a lways have more than four cr imina l-h istory points under Subparagraph A.

Second, defendants' elig ibility for relief wou ld not corre- spond to the ser i ousness of their cr imina l records. Instead, a defendant w ith numerous vi olent three-point offenses cou ld get relief because he happens not to have a two-point offense.

The content of Subparagraphs A, B, and C, especia lly as read

Opinion of the Court

against the Guidelines, thus answers the statutory puzzle here—reducing two grammatica l possibilities to just one plausible constr ucti on.

Beg in w ith superfuity. Or actua lly w ith its absence— because there is none under the Gover nment's reading.

Each subparagraph does independent work, disqua lify ing de- fendants from relief even when the others wou ld not. Sub- paragraph A disqua lifes defendants who have more than four cr imina l-h istory points (excluding those from a one- point offense), even if they do not have a pr i or three-point offense or a pr i or two-point vi olent offense. So, for exam- ple, a defendant w ith three non-vi olent two-point offenses w ill be barred. Subparagraph B, in tur n, disqua lifes de- fendants who have any pr i or three-point offense, even if they do not have a two-point vi olent offense or more than four tota l points. And fna lly, Subparagraph C disqua lifes de- fendants who have a pr i or two-point vi olent offense, even if they do not have a three-point offense or more than four points. The paragraph thus excludes (A) var i ous repeat of- fenders, a long w ith anyone having even a sing le convicti on that (B) resu lted in a suffciently long pr ison sentence or (C) resu lted in a shor ter sentence but involved vi olence. Every par t of the paragraph has a functi on.

But that is not so under Pu lsifer 's reading, as a bit of ar ith- metic revea ls. Pu lsifer 's view, once again, is that Paragraph (f )(1) disqua lifes only defendants w ith the combinati on of the character istics in Subparagraphs A, B, and C—so more than four cr imina l-h istory points, a pr i or three-point offense, and a pr i or two-point vi olent one. But because 3 + 2 = 5, and because 5 is more than 4, a defendant w ith a three-point of- fense (Subparagraph B) and a two-point vi olent offense (Sub- paragraph C) w ill necessar ily have more than four h istory points (Subparagraph A). So Subparagraph A becomes meaning less: It does no independent work. Remove it from the statute, and what is lef t w ill make the exact same people elig ible (and inelig ible) for relief.

Cite as: 601 U. S. 124 (2024) 143 Opinion of the Court

And that k ind of superfuity, in and of itself, refutes Pu lsif- er 's reading. The problem here is no odd word or stray phrase, wh ich might have escaped Congress's notice. Pu lsif- er 's reading wou ld negate one of three—indeed, the frst of three—provisi ons in the very paragraph he is try ing to in- terpret. When a statutory constr ucti on thus “render[s] an entire subparagraph meaning less, ” th is Cour t has noted, the canon against surplusage applies w ith specia l force. Na- t io n al Assn. o f Mf rs. v. Depar tmen t o f De fen se, 583 U. S.

109, 128 (2018); see Ch icago v. Ful t o n, 592 U. S. 154, 159 (2021). And still more when the subparagraph is so evi- dently desig ned to serve a concrete functi on. In addressing elig ibility for sentencing relief, Congress specifed three par- ticu lar features of a defendant's cr imina l h istory—A, B, and

It wou ld then have enacted: B and C. But wh ile that is the paragraph Pu lsifer 's reading produces, it is not the para- graph Congress wrote.

To escape that quandary, Pu lsifer contends that under the Guidelines a three-point offense and a two-point offense do not a lways tota l fve cr imina l-h istory points. (The dissent reiterates Pu lsifer 's asser ti on. See post, at 175 –177.) The arg ument beg ins w ith a point not in dispute: Some pr i or con- victi ons, as noted ear lier, add zero points to a defendant's cr imina l-h istory score. See supra, at 129. That is tr ue if the convicti on is quite old; if it was rendered in a military, tr iba l or foreig n cour t; or if it merged into another convicti on because, for example, the two arose from “ the same charg ing instr ument. ” § 4A1.2(a)(2); see § 4A1.2(e), (g)–( i). The key move in Pu lsifer 's arg ument is the next one: He claims that a pr i or offense adding zero points to a h istory score can still be a three-point or two-point offense under Paragraph (f )(1).

That happens, he says, when the sentence g iven for the of- fense was long enough to o the r w ise add those points— meaning, in the absence of the trait that reduced points to zero. See Br ief for Pu lsifer 36 – 41. Take an example: A

Opinion of the Court

very old convicti on contr ibutes zero points to a defendant's h istory score, no matter how long the sentence. Still, Pu l- sifer contends, it is a three-point offense when the sentence g iven was suffciently long (over 13 months) to add three points in a case no t very old. And once that propositi on is accepted, Pu lsifer says, superfuity disappears. Suppose a defendant has, a long w ith the old convicti on just descr ibed, a newer two-point vi olent offense. The old convicti on, Pu l- sifer maintains, is a three-point offense satisfy ing Subpara- graph B. And the new convicti on satisfes Subparagraph C.

But the defendant has only two cr imina l-h istory points—zero from the old offense and two from the new—wh ich is not enough to satisfy Subparagraph A. So that subparagraph, Pu lsifer concludes, has an effect: It keeps such a defendant elig ible for safety-va lve relief.

But Pu lsifer 's arg ument craters because its key move is wrong: Contrary to h is view, there is no such th ing under the Guidelines as a three-point or two-point offense adding zero points. Under Subparagraphs B and C, the ter ms “ 3- point offense” and “ 2-point vi olent offense” are “as deter- mined under the sentencing g uidelines. ” § 3553(f )(1)(B)–(C).

And the Guidelines assig n points to an offense only in the context, and for the purpose, of “ [a]dd[ing] ” them to a de- fendant's “cr imina l h istory” “ tota l. ” § 4A1.1. So a convic- ti on becomes a three- or two-point offense only when—only because—it adds three or two points to a tota l h istory score.

Or said the other way round, only the additi on of three or two points to that score makes the offense a three- or two- point offense. The corollary is that a convicti on adding zero points—because, say, it is very old—cannot be a three- or two-point offense. It is (unsurpr ising ly) a zero-point offense—whatever wou ld be the case if the convicti on were newer. For that reason, such a convicti on cannot aid Pu lsif- er 's effor t to fnd a functi on for Subparagraph A. Because only the additi on of three or two points can make an offense

Cite as: 601 U. S. 124 (2024) 145 Opinion of the Court

a three- or two-point offense, a defendant who has a pr i or three-point offense and a pr i or two-point vi olent offense w ill a lways have (ar ithmetic again) more than four points tota l.

The Guidelines' mechanics thus foreclose Pu lsifer 's effor t to erase the superfuity h is reading creates. 6 Yet more, Pu lsifer 's effor t founders on the Guidelines' judg ments, refected in Paragraph (f )(1), about wh ich pr i or offenses warrant enhanced punishment. Consider what Pu l- sifer 's zero-to-three claim entails. Because an offense add- ing zero points can on h is account satisfy Subparagraph B or C, it can help prevent a defendant from gaining safety-va lve relief. But that resu lt ill compor ts w ith the Guidelines.

There are, af ter a ll, reasons why the Guidelines decline to assig n points to cer tain offenses. The specifcs vary, but each embodies a judg ment that some types of pr i or convic- ti ons shou ld not have the usua l weight in deter mining a cur-

6 The dissent tries to save Pulsifer 's effort by offering an account of the Guidelines' mechanics different from that g iven in the Guidelines them- selves. According to the dissent, the Guidelines “set forth a two-step process” in which a judge frst “assig ns points to the defendant's prior offenses” under § 4A1.1 and only then “computes the defendant's criminal history” score under § 4A1.2. Post, at 175. So, the dissent claims, a court may frst count two or three points for an old conviction (under § 4A1.1) and then exclude those points from the computation of a history score (under § 4A1.2). See post, at 175 –176. But in fact there are no “ two steps” under the Guidelines. As even the dissent concedes in one self- contradictory moment, see post, at 176, there is only a single calculation, with § 4A1.2 providing the “ [d]efnitions and [i]nstr uctions” for § 4A1.1's “add[ing] ” and “ total[ing]. ” (Given that relationship, the commentary ex- plains, the two provisions “must be read together. ” USSG ch. 4, pt. A, intro.) And among § 4A1.2's “ [i]nstr uctions” is that old convictions simply “not [be] counted”—rather than, as the dissent would have it, that they be frst counted and then uncounted. In other words, as § 4A1.2's com- mentary states, an old conviction should “receive[ ] no criminal history points” in the § 4A1.1 calculation. USSG § 4A1.2 comment., n. 3. So again: Neither Pulsifer nor the dissent can transform an old conviction into a two- or three-point offense.

Opinion of the Court

rent sentence. Maybe the pr i or convicti on is not as reliable as most. Or maybe it is not so good a measure of the de- fendant's future dangerousness. Whatever the precise ex- planati on, the Guidelines g ive zero points to an offense in order to ensure that it not increase a later punishment. Ex- cept that under Pu lsifer 's view it cou ld do just that: Offenses that the Guidelines deem irrelevant to future sentencing might end up tr igger ing a mandatory minimum. And like- w ise, Pu lsifer 's view conficts w ith a discrete feature of Para- graph (f )(1). Reca ll that under that paragraph, an offense ear ning o ne point cannot affect elig ibility for relief, either a lone or in combinati on w ith any other offense—presumably b e c a u s e a o n e - p o i n t o f f e n s e i s j u s t t o o m i n o r. S e e § 3553(f )(1); supra, at 129 –130. Yet Pu lsifer 's theory wou ld a llow an offense adding ze ro points to contr ibute to a fnding that relief is barred. That claim is, again, too out-of-sync w ith the statutory framework to offer an escape from h is “ 3 + 2 = 5” superfuity problem.

And beyond that problem lies a second, th is one relating to the way Paragraph (f )(1) precludes safety-va lve relief for defendants w ith ser i ous cr imina l h istor ies. The paragraph operates as a gatekeeper: It helps get some defendants into, and keeps other defendants out of, a wor ld free of mandatory minimums. And the cr iter ia for selecti on, evident on the paragraph's face, relate to just how bad a defendant's cr imi- na l record is. Pu lsifer h imself recog nizes that fact: In de- scr ibing Paragraph (f )(1), he notes that “subparagraph (A) targets recidivism”; that “subparagraph (B) targets ser i ous offenses” leading to leng thy pr ison ter ms; and that “subpara- graph (C) targets vi olent offenses” even though resu lting in lesser sentences. Br ief for Pu lsifer 25; see i d., at 44 (noting that each subparagraph addresses a “ type[ ] of behavi or sug- gestive of future dangerousness”). The paragraph thus focuses on the k inds of past cr imina l behavi or that under the Guidelines tr igger enhanced pena lties. Over and over, the Guidelines presume that defendants w ith worse cr imi-

Cite as: 601 U. S. 124 (2024) 147 Opinion of the Court

na l records— exh ibiting recidivism, leng thy sentences, and vi olence—are “deserving of greater punishment. ” USSG ch. 4, pt. A, intro. comment.; see § 4A1.1; USSG ch. 5, pt. A.

Paragraph (f )(1), in line w ith its repeated invocati on of the Guidelines, expresses the same understanding. Put simply, the paragraph sor ts defendants for relief (or not) based on the ser i ousness of their cr imina l h istory.

Under the Gover nment's reading, Paragraph (f )(1) per- for ms that functi on w ithout a h itch. When the defendant has committed mu ltiple non-minor offenses, he cannot get relief (Subparagraph A). And so too when he has com- mitted even a sing le offense punished w ith a leng thy pr ison sentence (Subparagraph B) or involving vi olence (Subpara- graph C). Only a defendant w ith none of those markers—a defendant who can check off every one of the three “does not have” requirements—is elig ible for relief. So the paragraph unerr ing ly separates more ser i ous pr i or offenders from less ser i ous ones, a llow ing only the latter through the gate.

That does not happen under Pu lsifer 's constr ucti on. To the contrary, h is reading wou ld a llow relief to defendants w ith more ser i ous records wh ile barr ing relief to defendants w ith less ser i ous ones. Or said other w ise, the sor ting ac- complished by Pu lsifer 's reading does not match what Para- graph (f )(1) and the Guidelines ca ll for. Consider two hypo- thetica l defendants. One has fve cr imina l-h istory points from a pr i or three-point offense and a pr i or two-point vi olent offense. The other has 15 cr imina l-h istory points from fve pr i or three-point offenses, every last one of a vi olent nature, but . . . has no two-point vi olent offense. (All h is cr imes were too ser i ous to w ind up in the two-point category.)

Wh ich of the two defendants is the more ser i ous pr i or of- fender? The latter of course: His record exh ibits greater recidivism, leng th ier sentences, and more vi olence. But under Pu lsifer 's view of Paragraph (f )(1), wh ich of the two defendants is excluded from relief ? The for mer a lone. For want of a two-point offense, the latter remains elig ible to

Opinion of the Court

avoid a mandatory minimum. The paragraph thus fails to divide, at the gate for safety-va lve relief, more from less ser i- ous pr i or offenders.7 And contrary to Pu lsifer 's view, that problem cannot be solved by resor t to a sentencing judge's discreti on. Notably, Pu lsifer does not arg ue that there is any rhyme or reason to mak ing our ser ia l three-point vi olent offender elig ible for safety-va lve relief. He says only that Congress “ had no rea- son to be concer ned” about that outcome because it knew “ that a sentencing cour t wou ld still have discreti on to impose a propor ti onate sentence. ” Br ief for Pu lsifer 25; see i d., at 45; see a lso post, at 182 –183 (Gorsuch, J., dissenting) (simi- lar ly rely ing on judges' ability, even w ithout mandatory min- imums, to impose leng thy sentences). But that “ tr ust in discreti on” claim cannot here work. If Congress thought it cou ld a lways rely on sentencing discreti on, it wou ld not have created a cr imina l-h istory requirement in the frst instance.

That requirement, by its ter ms, confnes such discreti on.

More specifca lly, it a llows discreti on to operate only when a defendant's record does not reach a cer tain level of ser i ous- ness. On the Gover nment's reading, the paragraph well perfor ms that gatekeeping functi on, separating more ser i ous from less ser i ous cr imina l h istor ies. On Pu lsifer 's reading, the paragraph does not: As just shown, it a llows and denies

7 The dissent labors unsuccessfully to fnd an explanation for this state of affairs. Here is what it comes up with: Sometimes Congress opts for “standardized formulas” or bright-line tests even though their “over- and under-inclusi[on] ” will produce statutory “anomalies. ” Post, at 181 (citing Ranso m v. FIA Card Ser v ices, N. A., 562 U. S. 61, 78 (2011)); see post, at 182 –183. That is tr ue enough, but has no application here. Pulsifer 's view of the statute is no more “standardized” than the Government's— and so no more predictable or administrable. The dissent thus remains without a plausible, or even cogent, explanation for the failure of its inter- pretation (and its interpretation alone) to perform the statute's gatekeep- ing role—in Pulsifer 's own words, to separate defendants whose criminal history is more “suggestive of future dangerousness” from defendants whose criminal history is less so. Brief for Pulsifer 44.

Cite as: 601 U. S. 124 (2024) 149 Opinion of the Court

relief in ways that do not correspond to the gravity of what a defendant has previ ously done. The need for a judge to correct those resu lts—wh ich Pu lsifer admits—shows that h is reading is wrong. Once again, h is constr ucti on of Para- graph (f )(1)—however grammatica l—makes a hash of the scheme Congress devised.

Pu lsifer tr ies to tell a competing story (wh ich the dissent mostly adopts, see post, at 162 –163, 171–173). Even suppos- ing the grammar of Paragraph (f )(1) is a wash, Pu lsifer con- tends that statutory context suppor ts h is view of what that provisi on means by “and. ” His arg ument invokes the “pre- sumpti on of consistent usage and the meaning fu l-var iati on canon. ” Br ief for Pu lsifer 22. Those are the ter ms of ten g iven to a genera lly usefu l—but still “defeasible”—interpre- tive pr inciple: In a g iven statute, the same ter m usua lly has the same meaning and different ter ms usua lly have different meanings. A. Sca lia & B. Gar ner, Reading Law 170 –171 (2012). The pr inciple is mostly applied to ter ms w ith some hef t and distinctiveness, whose use draf ters are likely to keep track of and standardize. See, e. g., IBP, Inc. v. Alv arez, 546

U. S. 21, 33 – 34 (2005) (constr uing the ter m “pr incipa l activity”

in the same way when used in neighbor ing provisi ons); Wis- co n sin Cen tral L td. v. Un i ted Sta tes, 585 U. S. 274, 279 (2018) ( holding that “money remunerati on” must mean someth ing different from “a ll remunerati on” when used in “compani on” statutes (emphasis deleted)). Pu lsifer breaks new ground in apply ing the pr inciple to words as ubiquitous and (as shown above) sometimes context-dependent as “and” and “or. ” See supra, at 132 –148 and th is page. He arg ues, more specifca lly, that another “and” plus an “or ” in Secti on 3553(f ) show that the “and” in Paragraph (f )(1) must be read h is way. But even accepting that a cour t can sometimes demand har moni- zati on of “and”s and “or ”s, Pu lsifer 's arg ument fails.

Take the other “and” frst. As noted ear lier, the cr imina l- h istory requirement is only one of fve conditi ons for safety-

Opinion of the Court va lve relief set out in Secti on 3553(f ). See supra, at 128 –

129. Those conditi ons appear in a list —Paragraphs (f )(1)

through (f )(5)—w ith an “and” link ing them, just as an “and” links Paragraph (f )(1)'s three subparagraphs. A look at the appendix may be helpfu l here. See in f ra, at 153. The styl- ized versi on of the list in Secti on 3553(f ) (w ith content re- moved) goes as follows: The safety va lve operates “ if the cour t fnds” 1, 2, 3, 4, “and” 5.

The problem for Pu lsifer is that the meaning of the “and” in Secti on 3553(f ) does not advance h is reading of Paragraph (f )(1). Everyone, including Pu lsifer, agrees that the “and” in Secti on 3553(f ) connects fve requirements for safety-va lve relief, a ll of wh ich a defendant must meet. In Pu lsifer 's view, the Gover nment has to read Paragraph (f )(1)'s “and” differently to make each one of its subparagraphs disqua lify- ing. See Br ief for Pu lsifer 21; Reply Br ief 14 –15. But that is just wrong. The “and” in Secti on 3553(f ) works identi- ca lly to the “and” in the Gover nment's reading of Paragraph (f )(1). Secti on 3553(f )'s “and” creates an elig ibility checklist. A defendant fu lflls that provisi on's requirements if the cour t fnds 1, fnds 2, fnds 3, fnds 4, and (fna lly) fnds 5. So the “and” joins fve individua lly necessary conditi ons for relief.

Likew ise, the “and” in the Gover nment's constr ucti on of Paragraph (f )(1) creates an elig ibility checklist. A defend- ant satisfes that paragraph's requirements if he does not have A, does not have B, and (fna lly) does not have C. So again, the “and” joins severa l individua lly necessary condi- ti ons for safety-va lve relief. Everyth ing is consistent, in meaning and operati on a like. It is actua lly Pu lsifer who in- troduces dissonance into the provisi on. As to the larger list, he acknowledges that a defendant cannot get relief w ithout check ing off 1, 2, 3, 4, and 5 individua lly. But as to the sma ller list, Pu lsifer changes the r u le. Now a defendant need not satisfy A, B, and C individua lly. Instead, he can get relief so long as he does not have A, B, and C combined.

In other words, in Pu lsifer 's wor ld, Secti on 3553(f ) is an elig i- bility checklist, but Paragraph (f )(1) is not.

Cite as: 601 U. S. 124 (2024) 151 Opinion of the Court

Pu lsifer 's deployment of another paragraph's “or ” fares no better. Secti on 3553(f )(4) conditi ons safety-va lve relief on a fnding that the “defendant was not an organizer, leader, manager, or supervisor of others in the offense. ” See supra, at 128. Congress thus used an “or ” to sig nify that being any one of those th ings is disqua lify ing. Invok ing the meaning fu l-var iati on canon, Pu lsifer arg ues that the differ- ent ter m “and” in Paragraph (f )(1) must mean the opposite: that only the combinati on of the listed th ings disqua lifes a defendant. See Br ief for Pu lsifer 21– 22. Reca ll that we have a lready re jected Pu lsifer 's unador ned view that the word “or ” is needed to convey the Gover nment's reading of Paragraph (f )(1). See supra, at 137–139. Pu lsifer 's addi- ti ona l reference to Paragraph (f )(4)'s “or ” does not streng then h is case. As we have shown throughout th is opini on, con juncti ons are versatile words, wh ich can work differently depending on context. Here is yet another ex- ample. As another g lance at the appendix w ill confr m, the relevant clause in Paragraph (f )(4) is markedly different in leng th and for matting from the mater ia l in Paragraph (f )(1), natura lly leading to different choices respecting the use of con juncti ons. And anyway, Congress draf ted the current versi ons of the paragraphs at different times—Paragraph (f )(4) in 1994, Paragraph (f )(1) in 2018. See Vi olent Cr ime Control and Law Enforcement Act of 1994, 108 Stat. 1985; First Step Act of 2018, Pub. L. 115 – 391, 132 Stat. 5221. It wou ld hardly be surpr ising if 24 years later, Paragraph (f )(1)'s draf ters did not perfectly har monize their con juncti on usage w ith a dissimilar-look ing nearby paragraph. There can be few better illustrati ons of the “defeasib[ility] ” of the meaning fu l-var iati on canon. Sca lia, Reading Law, at 171.

Fina lly, Pu lsifer and the dissent make a misg uided arg ument about leg islative purpose. As noted ear lier, Congress enacted the revised versi on of Paragraph (f )(1) as par t of the First Step Act, a sig nifcant sentencing refor m law. See supra, at 128 –129. Pu lsifer explains that the new provisi on was meant “ to make safety-va lve relief more w idely available. ”

Opinion of the Court

Br ief for Pu lsifer 22. And the dissent h igh lights how many more defendants wou ld get safety-va lve relief under Pu lsif- er 's reading than under the Gover nment's. See post, at 160 – 161; see a lso post, at 155 –161, 173. We do not doubt the points. But they do not assist in interpreting the statutory text before us. Both views of the paragraph—Pu lsifer 's and the Gover nment's—sig nifcantly w iden the oppor tunity for safety-va lve relief ; reca ll that under the pr i or provisi on, any- th ing more than a sing le cr imina l-h istory point precluded deviati on from a mandatory minimum. See supra, at 128 –

129. And Pu lsifer 's interpretati on is not better just because

it wou ld go fur ther than the Gover nment's. “ [N]o law pur- sues its . . . purpose[s] at a ll costs. ” Lun a Pe rez v. Sturg is Publ ic Schoo ls, 598 U. S. 142, 150 (2023). So here, Congress did not eliminate but only cur tailed mandatory minimums— did not extend safety-va lve relief to a ll defendants, but only to some. And to deter mine the exact contours of that class, we can do no better than examine Paragraph (f )(1)'s text in context. For a ll the reasons g iven, that scr utiny revea ls that Pu lsifer 's view goes too far.

Yet Pu lsifer ( joined again by the dissent, see post, at 184 –

185) asser ts we are not done. At the least, he claims, the

meaning of the cr imina l-h istory requirement is uncer tain.

And because it is uncer tain, he must w in. The r u le of lenity, he says, requires cour ts to read “ambig uous cr imina l stat- utes in favor of liber ty. ” Br ief for Pu lsifer 47.

The problem is that we do not view Paragraph (f )(1) as genuinely ambig uous.8 There are, to be sure, two grammat- ica lly per missible readings of the statute when viewed in the abstract. It may be read Pu lsifer 's way—as stating that a defendant can get safety-va lve relief so long as he does not

8 For that reason, we have no need to address the Government's argu- ment that the r ule of lenity does not apply to Paragraph (f )(1) because it is not properly considered a “penal law. ” Brief for United States 46 – 47.

Cite as: 601 U. S. 124 (2024) 153 Appendix to opinion of the Court

have the combinati on (A, B, and C). Or it may be read the Gover nment's way—as stating that a defendant can get safety-va lve relief only if he does not have A, does not have B, and does not have C. But the diffcu lty in choosing be- tween those two constr ucti ons fa lls away once we consider the content of Subparagraphs A, B, and C: more than four cr imina l-h istory points (excluding points from a one-point of- fense), a pr i or three-point offense, and a pr i or two-point vi o- lent offense, a ll as deter mined under the Sentencing Guide- lines. Then we discover that Pu lsifer 's view creates g lar ing superfuity, whereas the Gover nment's view does not. And we discover that only the Gover nment's view renders the provisi on capable of sor ting more ser i ous from less ser i ous cr imina l records, consistent w ith both the statute's and the Guidelines' desig ns. The two possible readings thus reduce to one—leaving no role for lenity to play.

In sum, Paragraph (f )(1)'s cr imina l-h istory requirement sets out an elig ibility checklist. A defendant is elig ible for safety-va lve relief only if he satisfes each of the paragraph's three conditi ons. He cannot have more than four cr imina l- h istory points. He cannot have a pr i or three-point offense.

And he cannot have a pr i or two-point vi olent offense. Be- cause Pu lsifer has two pr i or three-point offenses tota ling six points, he is not elig ible. It makes no difference that he does not a lso have a pr i or two-point vi olent offense. Ac- cording ly, we affr m the judg ment of the Cour t of Appea ls for the Eighth Circuit.

It is so o rde red.

§ 3553. I mposit ion of a sent ence . . . . . (f ) Limitation on Applicability of Statutory Mini- mums in Certain Cases.—Notw ithstanding any other pro- visi on of law, in the case of an offense under secti on 401, 404, or 406 of the Controlled Substances Act (21 U. S. C. 841, 844,

Appendix to opinion of the Court

846), secti on 1010 or 1013 of the Controlled Substances Im- por t and Expor t Act (21 U. S. C. 960, 963), or secti on 70503 or 70506 of title 46, the cour t sha ll impose a sentence pursu- ant to g uidelines promu lgated by the United States Sentenc- ing Commissi on under secti on 994 of title 28 w ithout regard to any statutory minimum sentence, if the cour t fnds at sen- tencing, af ter the Gover nment has been afforded the oppor- tunity to make a recommendati on, that —

(1) the defendant does not have—

(A) more than 4 cr imina l h istory points, excluding any cr imina l h istory points resu lting from a 1-point offense, as deter mined under the sentencing g uidelines; (B) a pr i or 3-point offense, as deter mined under the sen- tencing g uidelines; and (C) a pr i or 2-point vi olent offense, as deter mined under the sentencing g uidelines;

(2) the defendant did not use vi olence or credible threats

of vi olence or possess a frear m or other dangerous weapon (or induce another par ticipant to do so) in connecti on w ith the offense;

(3) the offense did not resu lt in death or ser i ous bodily

in jury to any person;

(4) the defendant was not an organizer, leader, manager,

or supervisor of others in the offense, as deter mined under the sentencing g uidelines and was not engaged in a continu- ing cr imina l enterpr ise, as defned in secti on 408 of the Con- trolled Substances Act; and

(5) not later than the time of the sentencing hear ing, the

defendant has tr uthfu lly provided to the Gover nment a ll in- for mati on and evidence the defendant has concer ning the of- fense or offenses that were par t of the same course of con- duct or of a common scheme or plan, but the fact that the defendant has no relevant or usefu l other infor mati on to pro- vide or that the Gover nment is a lready aware of the infor ma- ti on sha ll not preclude a deter minati on by the cour t that the defendant has complied w ith th is requirement.

Cite as: 601 U. S. 124 (2024) 155 Gorsuch, J., dissenting

Infor mati on disclosed by a defendant under th is subsecti on may not be used to enhance the sentence of the defendant unless the infor mati on relates to a vi olent offense.

Justice Gorsuch, w ith whom Justice Sotomayor and Justice Jackson join, dissenting.

The First Step Act of 2018 may be “ ` the most sig nifcant cr imina l justice refor m bill in a generati on. ' ” Br ief for Sen.

Richard J. Durbin et a l. as Amici Cur i ae in Te r r y v. Un i ted Sta tes, O. T. 2020, No. 20 – 5904, p. 9. Through the 1980s and 1990s, Congress adopted an ever-increasing number of ever- longer mandatory minimum pr ison sentences. In par t due to these policies, the federa l pr ison popu lati on grew by more than 100% in less than a decade. In the First Step Act, Con- gress sought to reca librate its approach. It did so by prom- ising more individua ls the chance to avoid one-size-fts-a ll mandatory minimums and receive instead sentences that ac- count for their par ticu lar circumstances and cr imes.

Th is dispute concer ns who is elig ible for individua lized sentencing and who remains subject to mandatory minimums af ter the First Step Act. Before the Act, a defendant seek- ing to avoid a mandatory minimum had to satisfy fve str in- gent statutory tests. Af ter the Act, a ll those tests remain, only the frst is now less demanding. As revised, it provides that a defendant may be elig ible for individua lized sentenc- ing if he “does not have” three traits: (A) more than 4 cr imi- na l h istory points, (B) a 3-point offense, and (C) a 2-point vi olent offense. In lower cour t proceedings, the gover n- ment admitted that th is new test is “most natura l[ ly] ” read to mean what it says: A defendant may be elig ible for indi- vidua lized sentencing unless he possesses a ll three listed traits—A, B, and C. Br ief for United States in No. 19 – 50305 (CA9), p. 7 (Gover nment CA9 Br ief ); i d., at 10 –11; ac- cord, Br ief for United States in No. 21–1609 (CA8), p. 11 (Gover nment CA8 Br ief ). Despite its admissi on, however, the gover nment urges us to adopt a different constr ucti on.

Gorsuch, J., dissenting

It asks us to read the First Step Act as promising a defend- ant a chance at individua lized sentencing only when he does not have any of the three listed traits—A, B, o r C.

If th is difference seems a sma ll one, it is anyth ing but.

Adopting the gover nment's preferred interpretati on g uaran- tees that thousands more people in the federa l cr imina l jus- tice system w ill be denied a chance—just a chance—at an individua lized sentence. For them, the First Step Act offers no hope. Nor, it seems, is there any r u le of statutory inter- pretati on the gover nment won't set aside to reach that re- su lt. Ordinary meaning is its frst victim. Contextua l clues follow. Our traditi ona l practice of constr uing pena l laws str ictly fa lls by the wayside too. Replacing a ll that are policy concer ns we have no business consider ing. Respect- fu lly, I wou ld not indu lge any of these moves.

In approach ing the dispute before us, some background helps. Before the 1980s, federa l judges genera lly en joyed broad discreti on at sentencing. Of ten, they cou ld impose punishments rang ing from probati on up to statutor ily speci- fed maximum pr ison ter ms. M istre t ta v. Un i ted Sta tes, 488

U. S. 361, 363 (1989). In exercising that discreti on, judges

had to “consider every convicted person as an individua l” and pick punishments that “ft the offender and not merely the cr ime. ” Peppe r v. Un i ted Sta tes, 562 U. S. 476, 487– 488 (2011).

Today, many defendants still receive individua lized sen- tences. In the mine r un of federa l cases, a cour t w ill star t w ith sentencing g uidelines the United States Sentencing Commissi on has prepared at Congress's directi on. The g uidelines help a cour t identify a range of presumptively rea- sonable sentences tailored to the defendant and h is cr ime.

See Ri ta v. Un i ted Sta tes, 551 U. S. 338, 347 (2007). That Cite as: 601 U. S. 124 (2024) 157

Gorsuch, J., dissenting

range depends on an “offense level, ” a fg ure that takes into account the ser i ousness of the defendant's cr ime and h is role in it, as well as the defendant's “cr imina l h istory” score, a ta lly ing that accounts for h is past misconduct. United States Sentencing Commissi on, Guidelines Manua l §§ 1B1.1, 4A1.1– 4A1.2, ch. 5, pt. A (Nov. 2023) (USSG); see Rosales- M ireles v. Un i ted Sta tes, 585 U. S. 129, 133 –134 (2018). The g uidelines, however, are just that. A sentencing judge may sometimes depar t or vary from the g uidelines' recommended range, pick ing a lower or h igher sentence if it best fts the defendant and broader penolog ica l goa ls Congress has in- str ucted cour ts to consider. See Gall v. Un i ted Sta tes, 552

U. S. 38, 46, 49 – 50 (2007); 18 U. S. C. § 3553(a).

In the 1980s and 1990s, Congress pursued a different ap- proach for cer tain dr ug offenses. See Anti-Dr ug Abuse Act of 1986, 100 Stat. 3207– 2 to 3207– 4; Anti-Dr ug Abuse Act of 1988, 102 Stat. 4370, 4377– 4378. It required cour ts to im- pose mandatory minimum pr ison ter ms based only on the k ind and quantity of the dr ugs involved in the defendant's cr ime. A cour t “was required to send the offender to pr ison” for a set per i od of years “no matter how minor the offender 's par ticipati on in the offense may have been, and no matter what mitigating circumstances might have been present. ” J. Rakoff, Why the Innocent Plead Guilty and the Guilty Go Free 13 (2021). Under th is reg ime, for example, a defendant distr ibuting 5 grams of crack cocaine faced a 5- year mandatory pr ison ter m, and one w ith 50 grams faced a 10-year ter m. 100 Stat. 3207– 2 to 3207– 3. Meanwh ile, a defendant found w ith powder cocaine confronted those same pr ison ter ms only if he distr ibuted 100 times those amounts. Ib i d.

In shor t order, the federa l pr ison popu lati on exploded. In 1986, federa l pr isoners numbered 30,104, approximately 37.7% of whom were serving time for dr ug offenses. Dept. of Justice, Sourcebook of Cr imina l Justice Statistics 519 (31st

Gorsuch, J., dissenting

ed. 2003). By 1994, the federa l pr ison popu lati on reached a lmost 74,000, w ith approximately 61.3% of inmates serving time for a dr ug offense. Ib i d.

Ca lls for refor m came quickly and grew w ith time. See,

e. g., U. S. Sentencing Commissi on, Specia l Repor t to the

Congress: Mandatory Minimum Pena lties in the Federa l Cr imina l Justice System i i i (1991); i d., App. G (collecting statements from the Judicia l Conference and 12 circuits).

Eventua lly, Congress responded to these ca lls in var i ous ways. In one refor m, for example, it prospectively reduced the crack-cocaine dispar ity from 100:1 to 18:1. See Fair Sen- tencing Act of 2010, 124 Stat. 2372. In another, it adopted § 3553(f ), a provisi on that came to be ca lled the “safety va lve” and that lies at the hear t of today's case. See Vi olent Cr ime Control and Law Enforcement Act of 1994, 108 Stat.

1985 –1986.

As or ig ina lly enacted in 1994, the safety va lve provided modest relief. It exempted defendants who cou ld meet fve statutory cr iter ia from other w ise-applicable mandatory min- imums, directing instead that they shou ld receive individua l- ized sentences. Ib i d. (codifed as amended at 18 U. S. C. § 3553(f )). But the frst of the safety va lve's fve cr iter ia, codifed in paragraph (f )(1), was especia lly demanding. It precluded relief for any individua l w ith “more than 1 cr imi- na l h istory point”—meaning that a defendant cou ld fnd h im- self inelig ible for individua lized sentencing if h is background included even a sing le 60-day pr ison ter m or two pr i or of- fenses involving no pr ison ter m at a ll. 108 Stat. 1985; see § 3553(f )(1) (1994 ed.); USSG §§ 4A1.1( b)–(c), 4A1.2 (Nov.

1994).

In the First Step Act of 2018, Congress adopted an array o f f u r t h e r r e f o r m s . P u b . L . 1 1 5 – 3 9 1 , 1 3 2 S t a t . 5 1 9 4 .

Passed w ith over whelming major ities in both chambers of Congress and w ith presidentia l suppor t, the Act reduced the leng th of some mandatory minimums by 25%. See § 401, i d., at 5220 – 5221. It narrowed the circumstances

Cite as: 601 U. S. 124 (2024) 159 Gorsuch, J., dissenting

under wh ich a cour t cou ld “stack ” cer tain mandatory mini- mums on top of one another. See § 403(a), i d., at 5221– 5222;

One Year of Implementati on 5 (2020). And it made Con- gress's ear lier amendment to the crack-cocaine dispar ity ret- roactive, a llow ing individua ls sentenced before that amend- ment's adopti on a chance at resentencing. See § 404, 132 Stat. 5222.

The First Step Act a lso revised the safety va lve's frst pro- visi on. Where paragraph (f )(1) once barred a defendant w ith even a sing le cr imina l h istory point from receiving an individua lized sentence, Congress now chose a different course. As amended, the fu ll safety va lve today instr ucts a cour t to afford an individua lized sentence “ if [it] fnds at sentencing . . . that —” “(1) the defendant does not have— “(A) more than 4 cr imina l h istory points, excluding any cr imina l h istory points resu lting from a 1-point of- fense, as deter mined under the sentencing g uidelines; “(B) a pr i or 3-point offense, as deter mined under the sentencing g uidelines; and “(C) a pr i or 2-point vi olent offense, as deter mined under the sentencing g uidelines; “(2) the defendant did not use vi olence or credible threats of vi olence or possess a frear m or other danger- ous weapon (or induce another par ticipant to do so) in connecti on w ith the offense; “(3) the offense did not resu lt in death or ser i ous bodily in jury to any person; “(4) the defendant was not an organizer, leader, man- ager, or supervisor of others in the offense, as deter- mined under the sentencing g uidelines and was not engaged in a continuing cr imina l enterpr ise, as defned in . . . the Controlled Substances Act; and “(5) not later than the time of the sentencing hear ing, the defendant has tr uthfu lly provided to the Gover n- ment a ll infor mati on and evidence the defendant has

Gorsuch, J., dissenting

concer ning the offense or offenses that were par t of the same . . . common scheme or plan . . . . ” 18 U. S. C. § 3553(f ).

The questi on we face concer ns how the amended safety va lve works. Everyone agrees that a defendant must still clear fve daunting statutory hurdles. But the par ties dis- agree what the frst entails af ter the First Step Act. Ob- serving that the word “and” connects each of the subpara- graphs (f )(1)(A), (B), and (C), Mark Pu lsifer arg ues that the safety va lve's frst provisi on now operates to render inelig i- ble one k ind of defendant —a defendant who bears a ll three enumerated traits, A, B, and C. Because he does not have a ll three, Mr. Pu lsifer submits, he is elig ible for safety-va lve relief as long as he can satisfy the law's four remaining provi- si ons. Meanwh ile, on the gover nment's telling, paragraph (f )(1) renders three k inds of defendants inelig ible for relief— any defendant who has trait A, B, o r C. And because Mr. Pu lsifer has at least one of those traits, the rest of the safety va lve is irrelevant; paragraph (f )(1) a lone renders h im inelig ible for relief.

Disputes about the amended safety va lve's operati on have simmered for years in the lower cour ts and y ielded confict- ing resu lts.1 At least one th ing, though, is clear: The dispute before us matters profoundly. According to a Sentencing Commissi on ana lysis based on 2021 data, about 33% of dr ug offenders were elig ible for safety-va lve relief under the law's old ter ms. See 88 Fed. Reg. 7186 (2023). Under Mr. Pu lsif-

1 The Fourth, Ninth, and Eleventh Circuits have taken Mr. Pulsifer 's approach. See Un i ted Sta tes v. Jo nes, 60 F. 4th 230 (CA4 2023); Un i ted Sta tes v. Lopez, 998 F. 3d 431 (CA9 2021); Un i ted Sta tes v. Garco n, 54 F. 4th 1274 (CA11 2022) (en banc). The Fif th, Sixth, Seventh, and Eighth Circuits have taken the government's view. See Un i ted Sta tes v. Palo- mares, 52 F. 4th 640 (CA5 2022); Un i ted Sta tes v. Haynes, 55 F. 4th 1075 (CA6 2022); Un i ted Sta tes v. Pace, 48 F. 4th 741 (CA7 2022); 39 F. 4th 1018 (CA8 2022) (case below).

Cite as: 601 U. S. 124 (2024) 161 Gorsuch, J., dissenting

er 's understanding of the First Step Act, about 66% wou ld become elig ible for individua lized sentencing. See ib i d.

By contrast, under the gover nment's reading of the Act, that number wou ld shr ink to around 44%. See ib i d. Our deci- si on today thus promises to affect the lives and liber ty of thousands of individua ls.

Unless some feature of the law suggests that one or an- other of its ter ms bears a specia lized meaning, our duty is to interpret Congress's work as an ordinary reader wou ld.

See Ni z-Chav ez v. Garl and, 593 U. S. 155, 163 (2021). At the hear t of today's dispute lies no specia lized ter m but per- haps the most ordinary of words: Everyth ing tur ns on what work the word “and” perfor ms in paragraph (f )(1), where a sentencing cour t is tasked w ith deter mining whether “ the defendant does not have” three traits—A, B, “and” C.

In tak ing up the par ties' dispute, star t w ith a few simple and uncontested observati ons. First, as the Cour t agrees, “and” is “a con juncti on—a word whose functi on is to connect specifed items. ” An te, at 133; see J. Opdycke, Harper 's Eng lish Grammar 200 (rev. ed. 1966).

Second, and more specifca lly, “and” is an “additive” con- juncti on, one of ten indicating that the words it connects shou ld be added together. Id., at 200; The Ch icago Manua l of Style § 5.183, p. 191 (15th ed. 2003). As the Cour t ex- plains, when “and” perfor ms that role, it means “ [t]ogether w ith, ” “a long w ith, ” “ in additi on to, ” or “as well as. ” Amer- ican Her itage Dicti onary 66 (5th ed. 2018); see an te, at 133.

Th ird, in paragraph (f )(1) “and” connects a list in a nega- tive conditi ona l statement (“ if . . . the defendant does not have”). Negative conditi ona l “ if . . . not” statements of ten functi on like the word “unless. ” See R. Huddleston & G.

Pu llum, The Cambr idge Grammar of the Eng lish Lang uage § 14.3, p. 755 (2002). Consider the mother who tells her

Gorsuch, J., dissenting

ch ild, “If you do not have any homework lef t, you can go play w ith your fr iends. ” The ch ild wou ld understand that he cou ld play w ith h is fr iends un less he had homework lef t to do.

Now apply those observati ons to paragraph (f )(1). Given the meaning of “and, ” an ordinary reader wou ld natura lly understand that a defendant is elig ible for individua lized sentencing if he “does not have” trait A, trait B, t oge the r w i th trait C. Add to the mix what we know about the inter- changeability of “ if . . . not” and “unless”: A defendant may receive g uidelines sentencing un less he has trait A, trait B, t oge the r w i th trait C. Put the points together, and the stat- ute indicates that a cour t may issue an individua lized sen- tence unless the defendant has a ll three traits listed in § 3553(f )(1), just as Mr. Pu lsifer contends.

What the lang uage of paragraph (f )(1) suggests, surround- ing context confr ms. When Congress uses different ter ms in a statute, we nor ma lly presume it does so to convey differ- ent meanings. So u thw est Airl ines Co. v. Saxo n, 596 U. S.

450, 457– 458 (2022). We sometimes ca ll th is presumpti on the “meaning fu l-var iati on canon. ” Id., at 457. Here, we see just such a meaning fu l var iati on. When Congress sought a sing le word to indicate that one trait among many is suffcient to disqua lify an individua l from safety-va lve re- lief, it chose an obvi ous soluti on: not the con junctive “and, ” but the disjunctive “or. ” In fact, Congress used “or ” th is way no fewer than three times. Paragraph (f )(2) specifes that, for a defendant to be elig ible for individua lized sentencing, a cour t must fnd that “ the defendant did not use vi olence o r credible threats of vi olence o r possess a frear m o r other dangerous weapon (o r induce another par ticipant to do so) in connecti on w ith the offense. ” (Emphasis added.) Paragraph (f )(3) premises eli- g ibility on a fnding that a defendant's “offense did not resu lt

Cite as: 601 U. S. 124 (2024) 163 Gorsuch, J., dissenting

in death o r ser i ous bodily in jury to any person. ” (Emphasis added.) And paragraph (f )(4) provides that elig ibility for relief tur ns on whether the defendant “was not an organizer, leader, manager, o r supervisor of others in the offense. ” (Emphasis added.)

The fact that Congress repeatedly used “or ” when it wanted relief to tur n on a sing le trait among many suggests that the “and” in paragraph (f )(1) perfor ms different work.

Even the gover nment once acknowledged as much, conceding below that the “and” in paragraph (f )(1) is “most natura l[ ly] ” read as requir ing a sentencing cour t to fnd that a defendant possesses a ll three listed traits before holding h im inelig ible for relief. Gover nment CA9 Br ief 7; i d., at 10 –11; accord, e.g ., Gover nment CA8 Br ief 11. Nor is the gover nment a lone in th is unsurpr ising understanding: A study involving ordinary Amer icans found that the largest share of par tici- pants understood a sentence track ing paragraph (f )(1)'s str ucture to tr igger inelig ibility only if a ll three conditi ons are satisfed. See Br ief for Thomas R. Lee et a l. as Amici Cur i ae 15, 18.

The gover nment disputes none of th is evidence about the law's ordinary meaning. Instead, it beg ins w ith a theory.

Maybe, the gover nment says, there is another “per missible” way to read paragraph (f )(1). An te, at 133; Br ief for United States 18, 37. Maybe Congress implicitly wanted a reader to “distr ibut[e] ” the “verb phrase” “does not have” among each subparagraph. An te, at 134 –136 ( inter na l quotati on marks omitted); Br ief for United States 14 –18. Maybe, then, we shou ld effectively read the statute to work th is way,

(1) the defendant does not have —

(A) does not have more than 1 cr imina l h istory point . . . ;

Gorsuch, J., dissenting

(B) does not have a pr i or 3-point offense . . . ; and (C) does not have a pr i or 2-point vi olent offense.

Yes, the gover nment's implicit distr ibuti on theory requires a reader to delete words before the em dash. Yes, it requires a reader to reinser t them in three different places where they do not appear. But maybe, the gover nment suggests, Congress implicitly intended for a reader to do a ll that.

Even though what it wrote is susceptible to a far more natu- ra l constr ucti on requir ing none of these g ymnastics.

That is not how statutory interpretati on usua lly works.

Statutes aren't games or puzzles but “ instr uments of a prac- tica l nature, founded on the common business of human life, . . . and ftted for common understandings. ” 1 J. Story, Com- mentar ies on the Constituti on of the United States § 451,

p. 437 (1833). For that reason, we usua lly presume that

Congress “employed words in their natura l sense, and . . . intended what [it] said. ” G ibbo n s v. Ogden, 9 Wheat. 1, 188 (1824). And once we have identifed the most natura l sense of the law's ter ms, as we have here, our interpretive task is usua lly at an end. See, e.g ., Bar nhar t v. Si gmo n Co al Co.,

The gover nment's implicit distr ibuti on theory is so far from the most natura l reading of the law that its many and able lawyers didn't even stumble on it until late in the game.

In litigati on below, the gover nment star ted by arg uing pr i- mar ily that paragraph (f )(1) “must be read in the disjunc- tive”—a fancy way of say ing that “and” means “or. ” Gov- er nment CA8 Br ief 4; see Un i ted Sta tes v. Garco n, 54 F. 4th 1274, 1280 (CA11 2022) (en banc). In ear ly cases, that was the gover nment's only arg ument. See, e. g., Sentencing Tr. in No. 3:19 – cr– 207 (ED Tenn.), ECF Doc. 176, p. 4 (“I th ink the Depar tment of Justice's positi on as well as our positi on here today is . . . that it shou ld be read disjunctively”); see a lso Tr. of Ora l Arg. 103. Only af ter a resounding loss on that arg ument, see Un i ted Sta tes v. Lo pez, 998 F. 3d 431, 435 – 443 (CA9 2021), did the gover nment sh if t to its implicit

Cite as: 601 U. S. 124 (2024) 165 Gorsuch, J., dissenting

distr ibuti on theory, stressing that its new offer ing does not require cour ts to “ transfor m” “and” into “or, ” see Br ief for United States 42 – 43.

The gover nment's implicit distr ibuti on theory may be a “convenient litigating positi on, ” Bo w en v. Geo rge t o w n Un i v.

Hosp i tal, 488 U. S. 204, 213 (1988), but it does not come close to respecting the most natura l constr ucti on of the law. It may have the beneft of leaving “and” a lone, but it comes at the cost of rearrang ing so much else in the statute. One way or another, the gover nment cannot get where it w ishes to go w ithout tinker ing w ith the law. And to know that much shou ld be enough to br ing th is case to a close: “Cr imes are supposed to be defned by the leg islature, not by clever prosecutors r iffng on equivoca l lang uage. ” D ub in

v. Un i ted Sta tes, 599 U. S. 110, 129 –130 (2023) ( inter na l quo-

tati on marks and a lterati on omitted).

How does the gover nment reply? It insists that contex- tua l clues suppor t its implicit distr ibuti on theory. These clues are so compelling, it says, any other constr ucti on of the law isn't “plausible” or “possible. ” An te, at 142, 153; Br ief for United States 18 –19. It is a bold claim, not only because the gover nment over looks a ll the evidence of the statute's meaning outlined above, but a lso because it over looks one piece of contextua l evidence af ter another weigh ing against its implicit distr ibuti on theory.

Star t w ith th is one: The statute before us stands far afeld from classic cases that invite questi ons about implied distr i- buti on. In everyday speech, the gover nment stresses, a lis- tener may appreciate the need to “distr ibut[e] ” what th is Cour t has ca lled “severa l antecedents” to “severa l conse- quents. ” Encino M o t o rcars, LLC v. Nav ar ro, 584 U. S. 79, 87 (2018) (quoting 2A N. Singer & S. Singer, Suther land Stat- utes and Statutory Constr ucti on § 47:26, p. 448 (rev. 7th ed.

2014) ( inter na l quotati on marks omitted)). In its br ief be-

Gorsuch, J., dissenting

fore us, the gover nment leads w ith th is example of the phe- nomenon: someone who says, “I sell red, wh ite, and blue caps. ” See Br ief for United States 14. That statement, the gover nment observes, contains an ambig uity. One listener might th ink that the seller offers caps bear ing a ll three col- ors. But another listener might wonder if the seller implic- itly means to “distr ibute” different colors to different caps— so that she rea lly means to say she sells red caps, she sells blue caps, and she sells wh ite caps. Only context, the gover nment insists, can resolve the ambig uity and revea l wh ich understanding best refects the seller 's meaning. Id., at 16.

If context suggests anyth ing, however, it is that th is obser- vati on has little to offer when it comes to the statute before us. The First Step Act does not contain severa l “anteced- ents” (many caps, for example) that might or might not dis- tr ibute among severa l “consequents” (say, colors). Instead, paragraph (f )(1) speaks of a sing le person—“ the defendant” presently before the sentencing cour t —who must not have three specifed traits (A, B, and C). And that “sing u lar ” constr ucti on “ tends to avoid the ambig uity” about distr ibu- ti on that a “plura l” constr ucti on can invite. M. Kirk, Lega l Draf ting: The Ambig uity of “And” and “Or, ” 2 Tex. Tech.

Grammar § 1.3.1, at 1280 –1281.

Draf ting exper ts illustrate the point w ith th is phrase: “char itable and educati ona l instituti ons. ” R. Dickerson, The Fundamenta ls of Lega l Draf ting § 6.2, pp. 109 –110 (2d ed.

1986); Kirk, 2 Tex. Tech. L. Rev., at 239 – 241. The phrase is ambig uous. The mu ltiple “ instituti ons” might distr ibute across the mu ltiple listed traits to descr ibe both “char itable instituti ons and educati ona l instituti ons. ” Dickerson, Fun- damenta ls of Lega l Draf ting § 6.2, at 110; Kirk, 2 Tex. Tech.

tr ibute, so the phrase descr ibes only instituti ons that are both char itable and educati ona l. Id., at 240 – 241. But if there is just a sing le “ instituti on, ” any ambig uity dissipates:

Cite as: 601 U. S. 124 (2024) 167 Gorsuch, J., dissenting

“A char itable and educati ona l instituti on” is an instituti on w ith both traits. The same holds tr ue when a sa leswoman offers “ the red, wh ite, and blue cap”: In that case, a buyer knows w ith cer tainty that the seller offers one k ind of cap bear ing a ll three colors.

Th is contextua l clue poses the gover nment w ith a ser i ous problem. When Congress wrote paragraph (f )(1), it em- ployed a sing u lar constr ucti on that tends to avoid the ambi- g uity about distr ibuti on that plura l constr ucti ons invite.

The statute before us thus bears no resemblance to the gov- er nment's lead illustrati on involving mu ltiple caps and colors.

Nor does it bear any resemblance to the gover nment's var i- ous illustrati ons from statutory and constituti ona l law in- volving mu ltiple “offenses” that fa ll into mu ltiple classes, see an te, at 137; Br ief for United States 17–18 (discussing 34

U. S. C. § 20101(f )); mu ltiple “Cases” that meet mu ltiple de-

scr ipti ons, see an te, at 134 –135; Br ief for United States 40 (quoting Ar t. III, § 2, cl. 1); or the many k inds of “Commerce” Congress can reg u late, see an te, at 135; Br ief for United States 39 – 40 (quoting Ar t. I, § 8, cl. 3).2 Sensing the gover nment's diffcu lty, the Cour t str ugg les for an example of its own involving a sing u lar person or th ing that does generate an ambig uity about distr ibuti on.

Eventua lly, it lands on Er ic Car le's story about a caterpillar who “ `ate through' ” (among so many other th ings) “ `one sau- sage, one cupcake, and one slice of water melon. ' ” An te, at 134; see a lso an te, at 135, n. 3. Missi on accomplished: One ch ild might implicitly distr ibute the phrase “ate through” to each foodstuff, wh ile another might read the list w ithout im- plicit distr ibuti on to mean the caterpillar ate through a “com- binati on” that includes them a ll. An te, at 135.

2 Although at frst blush “Commerce” might appear to be a singular noun, this term in fact describes “a noncountable abstraction, ” Ni z-Chav ez

v. Garland, 593 U. S. 155, 163 (2021), that this Court has said sweeps in

“every species of commercial intercourse, ” G ibbo ns v. Ogden, 9 Wheat. 1, 193 (1824).

Gorsuch, J., dissenting

But what does that prove? “ [T]o acknowledge ambig uity is not to conclude that a ll interpretati ons are equa lly plausi- ble. ” G w al tney o f Smi th f eld, L td. v. Chesapeake Bay Fo unda t io n, Inc., 484 U. S. 49, 57 (1987). And an example of ambig uity about distr ibuti on in a ch ildren's book does noth ing to prove that the federa l cr imina l statute before us is most plausibly read to require implicit distr ibuti on. Add some of paragraph (f )(1)'s sa lient features into the illustra- ti on and that much becomes clear. As the story goes, the caterpillar is in the process of becoming a butterfy. So sup- pose the story said the caterpillar “w ill remain a caterpillar if he does not eat (A) one sausage, (B) one cupcake, and (C) one slice of water melon. ” I suspect most ordinary readers (and ch ildren) wou ld have little trouble concluding that the sentence means that the caterpillar w ill remain a caterpillar unless he eats a ll three th ings; one a lone w ill not do.

Here's another problem w ith the gover nment's theory: If in some contexts a speaker might intend for a listener to distr ibute words implicitly, the context before us counsels against attr ibuting any such intenti on to Congress. It does because a carefu l look at the safety-va lve statute revea ls that, when Congress wanted to distr ibute a phrase in th is law, it did not leave the matter to implicati on. It did not depend on the reader 's imag inati on. Instead, Congress dis- tr ibuted phrases expressly.

Tw ice, in fact. In paragraph (f )(4), Congress took the trouble to distr ibute expressly the phrase “was not, ” per mit- ting relief only if “ the defendant w as no t an organizer, leader, manager, or supervisor of others . . . and w as no t engaged in a continuing cr imina l enterpr ise. ” (Emphasis added.) Likew ise, in paragraph (f )(1) itself Congress ex- pressly distr ibuted the phrase “as deter mined under the sen- tencing g uidelines” three times, in each of subparagraphs (A), (B), and (C). All the contextua l evidence before us thus

Cite as: 601 U. S. 124 (2024) 169 Gorsuch, J., dissenting

suggests that, in a statute carry ing grave cr imina l conse- quences, Congress was carefu l w ith its words and concer ned w ith clar ity. It did not leave ambig uities about distr ibuti on to be resolved by implicati on. Instead, it resolved them ex- pressly, even at the cost of repetiti on.

Once more, the gover nment's examples only serve to illus- trate its problem. It imag ines a speaker who says, “ `I didn't like h is mother and father. ' ” An te, at 136; Br ief for United States 39. The gover nment suggests that a listener wou ld “probably” understand the sentence as implicitly distr ibuting the phrase “I didn't like h is, ” so that it rea lly means, “I didn't like h is mother and I d i dn' t l i ke h is father. ” An te, at 136 (emphasis added); Br ief for United States 39. But as the hedge (“probably”) indicates, an ambig uity lurks here. The sentence cou ld a lso be understood w ithout any distr ibuti on to convey the idea that “I didn't like h is mother and father ” as a couple, even if I liked each individua lly well enough.

See Huddleston, Cambr idge Grammar § 2.2.2, at 1298 –1299.3 Only context, the gover nment concedes, can clar ify wh ich meaning is more apt. See an te, at 141; Br ief for United States 16. Yet somehow, the gover nment neg lects that same message when it comes to the statute before us—where context revea ls that Congress did not leave questi ons of dis- tr ibuti on to implicati on but resolved them expressly.

Context exposes yet another faw in the gover nment's im- plicit distr ibuti on theory. If, as the gover nment imag ines,

3 Same goes for the government's example “I'm not free on Saturday and Sunday. ” An te, at 136; Brief for United States 39. In some contexts, the sentence might be understood to distribute the phrase “I'm not free on” and mean “I'm not free on Saturday and I'm no t f ree o n Sunday. ” In others (suppose you were asked for help with a 2-day home renovation project), it might mean “I'm not free on Saturday and Sunday” as a combi- nation, even if I am free one day or the other. See Huddleston, Cam- bridge Grammar § 2.2.2, at 1298 –1299.

Gorsuch, J., dissenting

Congress was deter mined to fnd an “effcient” way to dis- qua lify a defendant bear ing any one of the three traits listed in paragraph (f )(1), an te, at 134, 139; see Br ief for United States 18, it had an obvi ous soluti on before it: the word “or. ” As we have seen, Congress employed exactly that approach three times in the safety va lve: Paragraphs (f )(2), (f )(3), and (f )(4) a ll premise disqua lifcati on for relief on the presence of one trait o r another. See Par t II–B, supra. In th is way, too, context confr ms that, when Congress wanted to make one trait among many disqua lify ing, it proceeded expressly (and of ten effciently)—but never by implicati on.

Af ter disregarding others, the gover nment at least ac- knowledges th is par ticu lar complicati on for its theory. It responds th is way: Even substituting “or ” for “and, ” it says, wou ld not “delive[r] us from interpretive controversy. ” An te, at 138; Br ief for United States 26. It wou ld not be- cause replacing “and” w ith “or ” in paragraph (f )(1) still wou ld not answer the questi on whether a sing le trait a lone is enough to render a defendant inelig ible for relief. An te, at 138; Br ief for United States 26. As evidence of the ma lle- ability of the word “or ” in some contexts, the Cour t cooks up var i ous illustrati ons involving a hypothetica l chef. An te, at 138, n. 5.

It is a remarkable response. At arg ument, the gover n- ment acknowledged that “or ” “might have been a clearer way to express” that a sing le trait is disqua lify ing in para- graph (f )(1). Tr. of Ora l Arg. 98. Below, the gover nment initia lly pushed for treating “and” as meaning “or ” precisely because it knew that doing so wou ld mean that a defendant is inelig ible for relief if he has even one of its listed traits.

See i d., at 101; Gover nment CA9 Br ief 11–13; Gover nment CA8 Br ief 7– 8. And everyone, the Cour t included, concedes that Congress's use of the word “or ” in paragraph (f )(4) means that a defendant meeting any one of severa l cr iter ia is disqua lifed from relief. An te, at 151. Simply put, “we wou ldn't be sitting here if Congress had used the word `or ' ”

Cite as: 601 U. S. 124 (2024) 171 Gorsuch, J., dissenting

in paragraph (f )(1). Tr. of Ora l Arg. 97. Whatever ambig u- ity “or ” might carry in other contexts, it carr ies none in § 3553(f ). Throughout the safety va lve, Congress used it to indicate that a sing le trait among many is disqua lify ing.4 Finding the gover nment on its back foot, the Cour t again comes to its defense, th is time by try ing to change the r u les of play. Perhaps, the Cour t specu lates, Congress's choice of “and” rather than “or ” in paragraph (f )(1) was the product of careless d ra f t i n g. Se e a n t e, at 13 7– 13 8. Pe rh ap s, to o, those two con juncti ons are “versatile” words not entitled to the respect we usua lly pay Congress's var iati ons in usage— a respect, the Cour t suggests, that is due only “ to ter ms w ith some hef t and distinctiveness, whose use draf ters are likely to keep track of and standardize. ” An te, at 149, 151.

Consider how far we have now retreated. Lower cour ts re- jected the gover nment's and-means-or arg ument. In re- sponse, the gover nment introduced its implicit distr ibuti on theory. Before us, the gover nment stresses that its new the- ory does not depend on “ transfor m[ing] ” “and” into “or. ” Br ief for United States 42; see a lso i d., at 15, 25. At frst, the Cour t seems to proceed on the same premise. See an te, at 133 –134. But now it reverses course. Resuscitating an arg u-

4 Alternatively, the government suggests, Congress might have used “and” in paragraph (f )(1) rather than “or ” as it did in paragraphs (f )(2)–

(4) because of something to do with the length or format of these provi-

sions. To that end, the government invites us to compare paragraphs (f )(1) and (f )(4). Tr. of Oral Arg. 64; see also an te, at 151. But, as it turns out, those paragraphs are almost the same length: 49 words and 40 words, respectively. See §§ 3553(f )(1), (4). Nor can much be made of the formatting. The main difference is paragraph (f )(1)'s use of an em dash to set off the listed traits. But even the government has declined to make much of the em dash, and for good reason. It simply “mark[s] an interr up- tion in the str ucture of a sentence, ” substituting here for a colon. B.

Garner, Modern English Usage 750 (4th ed. 2016). No party before us suggests that this em dash is so versatile that it can transform an interr up- tion into an implied distribution. See Brief for United States 38 – 39 (con- ceding that an em dash “ is neither necessary nor suffci en t for a distribu- tive interpretation” (emphasis added)).

Gorsuch, J., dissenting

ment the gover nment itself has abandoned, the Cour t con- tends not just that the ter ms “and” and “or ” are interchange- able, but that we need not even rely on our usua l r u les of interpretati on when faced w ith them.

Th is arg ument was a loser below and it shou ld be here.

When Congress employs “differ ing lang uage in . . . two sub- secti ons, ” we star t from a presumpti on that it meant to con- vey a difference in meaning, not a presumpti on that it made “a simple mistake in draf tsmansh ip. ” Russell o v. Un i ted Sta tes, 464 U. S. 16, 23 (1983). Never, to my knowledge, has th is Cour t suggested that we may tur n our back on th is ap- proach when con juncti ons or other putatively “ indistinctive” words are in play. Nor have we deployed that approach for “ hef ty” words a lone—as if we were pick ing paper towels in- stead of interpreting statutes. To the contrary, our cases beg in (and of ten end) w ith the presumpti on that Congress is carefu l in a ll its word choices and afford var iati ons between ter ms like “and” and “or ” the same respect due others. See, e.g ., Un i ted Sta tes v. Pal o mar-San t i ago, 593 U. S. 321, 326 (2021) (reversing the lower cour t for failing to g ive effect to a statute's use of “ the con junctive `and' ”); Encino M o t o r- cars, 584 U. S., at 87 (resting a reading of the relevant statute on “ the ordinary, disjunctive meaning of `or ' ”); Lo ughr in v.

Un i ted Sta tes, 573 U. S. 351, 357 (2014) (re jecting an arg u- ment that wou ld “disregard what `or ' customar ily means”); Re i te r v. So no t o ne Co r p., 442 U. S. 330, 338 – 339 (1979) (simi- lar); Rum sfeld v. Pad i ll a, 542 U. S. 426, 434 (2004) (g iving weight to the federa l habeas statute's “consistent use of the defnite ar ticle”). 5 Nor cou ld the premise latent in the Cour t's arg ument be fur ther from the tr uth. The difference between words like “and” and “or ” of ten cannot be easily dismissed as meaning-

5 Even the cases the Court cites, see an te, at 149, describe the presump- tion of meaning ful variation without the qualifcation it now imag ines.

See IBP, Inc. v. Alv arez, 546 U. S. 21, 34 (2005); Wisco nsin Cen tral Ltd.

v. Un i ted Sta tes, 585 U. S. 274, 279 (2018).

Cite as: 601 U. S. 124 (2024) 173 Gorsuch, J., dissenting

less when it comes to settling lega l r ights. Just imag ine if the Sixth Amendment gave the accused a “r ight to a speedy o r public tr ia l. ” Rather than getting a both timely and transparent tr ia l, a defendant wou ld be forced to choose wh ich feature he prefers. Because the difference between “and” and “or ” so reg u lar ly proves dispositive of impor tant lega l r ights, draf ting manua ls for lega l text from contracts to congressi ona l leg islati on war n about the need to deploy the ter ms w ith care. See, e.g ., Senate Offce of the Leg isla- tive Counsel, Leg islative Draf ting Manua l 64 – 65 (1997); K.

Adams, A Manua l of Style for Contract Draf ting §§ 11.9 – 11.11, p. 211 (3d ed. 2013). And here, of course, the differ- ence between “and” and “or ” affects the lives of thousands, see supra, at 160 –161—a fact so inconvenient for the Cour t that the Cour t says to ig nore it as well, see an te, at 152.6

So far, th ings look bleak for the gover nment. Mr. Pu lsifer offers a perfectly natura l reading of the law. In response, the gover nment offers a theory that it says r ises or fa lls based on context. See an te, at 140 –141; Br ief for United States 11, 16. Yet, as it tur ns out, not one but three contex- tua l clues array against its theory.

6 The Court offers still one more guess, again premised on careless draf t- ing, about why Congress used “and” rather than “or. ” Maybe, the Court posits, when Congress amended paragraph (f )(1) in 2018 it failed to notice that it had used “or ” when draf ting paragraphs (f )(2)–(f )(4) in 1994. An te, at 151. Normally, though, we assume “ that Congress is aware of existing law when it passes leg islation. ” Mi les v. Apex Mar ine Cor p., 498 U. S. 19, 32 (1990). And it beggars belief to suppose that Congress didn't bother to review the rest of the safety valve when it amended one of its provisions— particularly when it knew that defendants, prosecutors, and judges would necessarily read all fve safety-valve provisions together as part of a single “elig ibility checklist. ” An te, at 150; cf. G wal tney o f Smi th f eld, Ltd. v.

Chesapeake Bay Founda t io n, Inc., 484 U. S. 49, 57, and n. 2 (1987) (point- ing to meaning ful variation between the statutory language at issue and other, later enacted statutory provisions to counter the assertion that the choice of language was “a `careless accident' ”).

Gorsuch, J., dissenting

Unable to muster a convincing response to any of that, the gover nment pivots. Even if its implicit distr ibuti on theory suffers so many faws, the gover nment urges us to adopt it anyway because Mr. Pu lsifer 's reading of the law wou ld in- troduce a superfuity into the safety-va lve statute. It is a resourcefu l reply. The gover nment has many. But it, too, fa lls fat.

Without questi on, the canon against superfuity can be a usefu l tool when seek ing the meaning of a statute. It rests

the presumpti on that Congress is a carefu l draf ter and each word it chooses “ is there for a reason. ” Advoca te Heal th Care Ne two rk v. Staple t o n, 581 U. S. 468, 477 (2017). But that fact a lso makes the gover nment's choice to rest its case on the superfuity canon a cur i ous one. As we have seen, the gover nment's implicit distr ibuti on theory depends on the assumpti on that Congress was no t a carefu l draf ter. It re- quires us to assume Congress lef t a distr ibuti on implicit in one secti on of paragraph (f )(1), even as it made others ex- press elsewhere in paragraphs (f )(1) and (f )(4). It requires us to assume Congress meant for “and” in paragraph (f )(1) to do the same work as “or ” in paragraphs (f )(2)–(f )(4).

Sometimes, it seems, we are supposed to assume Con- gress was sloppy, other times carefu l. The only common thread seems to be what benefts the gover nment in the moment.

Even putting that sma ll irony aside, the gover nment has a bigger problem: Mr. Pu lsifer 's reading leaves no provisi on in th is statute superfuous. As the gover nment sees it, a defendant who has both the pr i or 3-point offense required by subparagraph (B), and the 2-point vi olent offense required by subparagraph (C), w ill necessar ily have more than the 4 cr imina l h istory points required by subparagraph (A). Be- cause of th is, the gover nment submits, subparagraph (A) has no work to perfor m on Mr. Pu lsifer 's reading: “Remove it

Cite as: 601 U. S. 124 (2024) 175 Gorsuch, J., dissenting

from the statute, and what is lef t w ill make the exact same people elig ible (and inelig ible) for relief. ” An te, at 142; Br ief for United States 19 – 20. Only its implicit distr ibuti on theory, the gover nment contends, can cure the problem by a llow ing one subparagraph to “disqua lif [ y] defendants from relief even when the others wou ld not. ” An te, at 142; Br ief for United States 19 – 20.

It's a nice arg ument, but it rests on a fau lty premise. As it happens, a defendant who has a 3-point offense under sub- paragraph (B) and a 2-point vi olent offense under subpara- graph (C) of ten w ill no t have “more than 4 cr imina l h istory points . . . under the sentencing g uidelines” for purposes of subparagraph (A). And in cases like that, subparagraph (A) perfor ms vita l work under Mr. Pu lsifer 's reading of the law by ensur ing that the defendant remains elig ible for relief.

There is simply no surplus here for the gover nment's implicit distr ibuti on theory to cure.

To appreciate why th is is so, consider the sentencing g uidelines Congress cross-referenced in subparagraphs (A) through (C). They set for th a two-step process for ca lcu lat- ing a defendant's cr imina l h istory. At the frst step, dis- cussed in § 4A1.1 of the g uidelines, a judge assig ns points to the defendant's pr i or offenses. Usua lly, the points corre- spond to the leng th of the defendant's previ ous sentences.

So, for example, three points nor ma lly attach to an offense carry ing a sentence longer than 13 months, two points to an offense w ith a sentence shor ter than that but at least 60 days long, and one point to any other sentence.

At the second step, descr ibed in § 4A1.2 of the g uidelines, a judge then computes the defendant's cr imina l h istory. But dur ing th is process, a judge doesn't just tote up a ll the points assig ned to each offense. Under a var iety of circumstances, the g uidelines instr uct a judge no t to count points assig ned to one offense or another. Points associated w ith h itchh ik- ing, public intoxicati on, and fsh and game offenses, for exam- ple, “are never counted. ” § 4A1.2(c)(2). Nor are points as-

Gorsuch, J., dissenting

sociated w ith sentences imposed by a cour t-mar tia l, a foreig n cour t, or a tr iba l cour t. §§ 4A1.2(g)–( i). The g uidelines a lso instr uct judges not to count points associated w ith of- fenses of a cer tain age. So, by way of illustrati on, if the defendant fnished h is sentence for a 3-point offense more than 15 years ago, those points are not counted. Like- w ise, if the defendant fnished h is sentence for a 2-point of- fense more than 10 years ago, those points do not count. §§ 4A1.2(e)(1)–(3). Cour ts thus perfor m “a sing le ca lcu la- ti on” of a defendant's cr imina l h istory score. An te, at 145,

n. 6. But in doing so, they routinely disting uish between

the points an offense carr ies and a defendant's u ltimate, countable cr imina l h istory points.7 Now retur n to subparagraph (A). It provides that the de- fendant must not have “more than 4 cr imina l h istory points, exclud ing any cr imin al h ist o r y po in ts resul t ing f ro m a 1- po in t o ffen se, as deter mined under the sentencing g uide- lines. ” § 3553(f )(1)(A) (emphasis added). As the ita licized lang uage demonstrates, when adopting the First Step Act Congress fu lly appreciated the distincti on between what points an offense carr ies and whether those points contr ibute to a defendant's cr imina l h istory score. And because of that very distincti on, it is possible for a defendant to have a pr i or 3-point offense and a pr i or 2-point vi olent offense w ithout

7 See, e.g ., Un i ted Sta tes v. Nesby, 2020 WL 4933657, *2 (SD Ill., Aug.

24, 2020) (defendant had “accumulated 34 criminal history points, many of which were not countable in his criminal history calculation”); Jo nes v.

Un i ted Sta tes, 2019 WL 365715, *3 (D NJ, Jan. 30, 2019) (“ The Sentencing Guidelines only permit a maximum of four one-point offenses to count toward a defendant's criminal history”); Un i ted Sta tes v. Johnso n, 2023 WL 4944732, *1 (WD Pa., Aug. 3, 2023) (same); Damero n v. Un i ted Sta tes, 2007 WL 893050, *4, n. 1 (ND Ohio, Mar. 21, 2007) (“ The criminal convic- tions above produce a subtotal criminal history score of 10, and it is noted that 3 of the defendant's 7 points under U.S.S.G. § 4A1.1(c) were not count- able”); Un i ted Sta tes v. Dal to n, 2010 WL 455239, *3 (D SC, Feb. 2, 2010) (noting 45 uncountable criminal history points “ in addition to the ff teen countable criminal history points”).

Cite as: 601 U. S. 124 (2024) 177 Gorsuch, J., dissenting

having more than 4 cr imina l h istory points. Most obvi ously, as Ch ief Judge Pryor, for mer Acting Chair of the Sentencing Commissi on, has observed, a defendant may have a 3-point offense and a 2-point vi olent offense but both offenses are so old that he scores no cr imina l h istory points at a ll. See USSG §§ 4A1.2(e)(1)–(3); Garco n, 54 F. 4th, at 1281. As Judge Wood has noted, there are a var iety of other situati ons as well in wh ich a defendant w ill have both a 3-point offense and a 2-point vi olent offense but still not have more than four cr imina l h istory points. See Un i ted Sta tes v. Pace, 48

F. 4th 741, 763 –764 (CA7 2022) (dissenting opini on).

To know that is to know no superfuity problem exists— and thus no need to resor t to the gover nment's implicit dis- tr ibuti on theory to solve it. On Mr. Pu lsifer 's reading of the law, a cour t apply ing subparagraph (A) w ill consu lt the sentencing g uidelines' methodolog y for scor ing cr imina l h is- tory points set for th in § 4A1.2. In doing so, the cour t may fnd that, wh ile the defendant has a pr i or 3-point offense and a pr i or 2-point vi olent offense for purposes of subparagraphs (B) and (C), one or another is too old or suffers from some other faw so that he does not have more than four cr imina l h istory points. In a ll these cases, subparagraph (A) does sig nifcant work by mak ing clear that, despite having a pr i or 3-point offense and a pr i or 2-point vi olent offense, the de- fendant remains elig ible for relief. See Garco n, 54 F. 4th, at 1281–1282; Pace, 48 F. 4th, at 763 (opini on of Wood, J.).

The gover nment does not contest the centra l observati on that defeats its superfuity arg ument. It admi ts that cer- t a i n p a s t o f f e n s e s “ a d [ d ] z e r o p o i n t s t o [a d e f e n d a n t 's] cr imina l-h istory score. ” Br ief for United States 32, n. 2; an te, at 143. So what exactly is the problem here?

To complain about a superfuity problem, it tur ns out the gover nment must create one. It does so th is way. As wr it- ten, subparagraphs (B) and (C) require a sentencing cour t

Gorsuch, J., dissenting

to ask whether the defendant “ ha[s] ” a “ 3-point offense, as deter mined under the sentencing g uidelines, ” and “a 2-point vi olent offense, as deter mined under the sentencing g uide- lines. ” But, the gover nment suggests, we shou ld read those provisi ons differently. We shou ld read them to require a sentencing cour t to ask the fur ther questi on whether the de- fendant's offenses also score cr imina l h istory points. As the gover nment candidly admits, its superfuity arg ument de- pends on reading subparagraphs (B) and (C) as “car[ing] only about offenses that do score . . . cr imina l-h istory points. ” Br ief for United States 28 – 29; an te, at 145 –146. Only then might subparagraph (A) be lef t w ithout work to perfor m, for indeed an offense that scores three cr imina l h istory points under subparagraph (B) and a vi olent offense that scores two cr imina l h istory points under subparagraph (C) w ill a lways score more than four cr imina l h istory points under subpara- graph (A).

Put plainly, for the gover nment's superfuity arg ument to gain any tracti on, we must read st i ll mo re words into the First Step Act, constr uing it now th is way:

(1) the defendant does not have —

(A) does not have more than 4 cr imina l h istory points, excluding any cr imina l h istory points resu lting from a 1-point offense, as deter mined under the sentenc- ing g uidelines; (B) does not have a pr i or 3-point offense that scores 3 cri m i n a l hi st or y poi nt s, as deter mined under the sentencing g uidelines; and (C) does not have a pr i or 2-point vi olent offense that scores 2 cri m i n a l hi st or y poi nt s, as deter mined under the sentencing g uidelines.

It is one more remarkable request. Last I heard, the canon again st assuming Congress has adopted superfuous words is not a license for judges to crea te a superfuity by inser ting new words into a law. Let a lone do so simply to

Cite as: 601 U. S. 124 (2024) 179 Gorsuch, J., dissenting

help the gover nment make its implicit distr ibuti on theory seem just a little less implausible.

At th is stage, the gover nment w ithdraws to its fna l re- doubt: a policy arg ument. In the gover nment's view, the only “functi on” Congress gave paragraph (f )(1) was the task of separating “more from less ser i ous pr i or offenders. ” An te, at 147–148; Br ief for United States 21. Affording the statute's ter ms their ordinary meaning, the gover nment asser ts, wou ld not a llow the law to perfor m that “purpose” adequately. Id., 20. By contrast, its implicit distr ibuti on theory wou ld enable the law to fu lfll its intended “role” “unerr ing ly. ” An te, at 147, 148, n. 7; Br ief for United States 21.

If th is policy arg ument sounds familiar, it is because we have time and again re jected ones just like it. We do not presume that a law perfor ms only one “functi on” or “role, ” but recog nize that a lmost every piece of leg islati on seeks to serve many competing purposes. See Lun a Pe rez v. Sturg is Publ ic Schoo ls, 598 U. S. 142, 150 (2023); Bar nhar t, 534 U. S., at 461; Ch icago v. Env iro nmen tal De fen se Fund, 511 U. S.

328, 339 (1994). We do not suppose that a law pursues any of those competing purposes to its log ica l end, acknowledg- ing instead that a lmost every law is the product of compro- mise. Lun a Pe rez, 598 U. S., at 150. And we do not dis- place ordinary statutory ter ms w ith judicia l “specu lati on as to Congress['s] intent, ” Magwood v. Pa t te rso n, 561 U. S. 320, 334 (2010), because the Amer ican people have consented to be gover ned by the wr itten laws their elected representa- tives adopt, not by the con jecture of others, see Un i ted Sta tes v. Bass, 404 U. S. 336, 348 (1971). For a ll these rea- sons and more, “ it is quite mistaken to assume, ” as the gov- er nment does, “ that whatever might appear to fur ther the statute's pr imary objective must be the law. ” Hen so n

Gorsuch, J., dissenting

v. San tande r Co n sume r USA Inc., 582 U. S. 79, 89 (2017)

( inter na l quotati on marks and a lterati ons omitted).

Perhaps reca lling our frequent admoniti on that policy ta lk cannot overcome plain text, the gover nment tr ies a bit of rebranding. Although it refers occasi ona lly to the First Step Act's “purpose, ” Br ief for United States 20 – 21, 34, 48, for the most par t it frames its arg ument in ter ms of rati ona lity. When we measure the competing interpre- tati ons before us against how well they perfor m the sta- tute's only job, the gover nment insists, we w ill fnd that the law's ordinary meaning invites “arbitrar[ y] ” resu lts and “nonsensica l” implicati ons. Id., at 22, 34, 36, 48. The Cour t buys into th is th inly disg uised policy appea l, see an te, at 146 –148, and n. 7, even as it forcefu lly (and w ithout a trace of irony) fau lts Mr. Pu lsifer for appea ling to statutory “pur- pose, ” an te, at 151. New framing or old, however, we have n o b u s i n e s s e n t e r t a i n i n g t h e g o v e r n m e n t 's r a m s h a c k l e arg ument.

If anyth ing, the gover nment's attempt at rebranding only makes matters worse for it. When a statute produces a tr u ly irrati ona l resu lt, we have a doctr ine to dea l w ith the dilemma: absurdity. In narrow circumstances, a simple and “eas[ily] ” fxed statutory error that “no reasonable per- son cou ld intend” may be amenable to judicia l correcti on under th is Cour t's traditi ona l absurdity doctr ine. See A.

Sca lia & B. Gar ner, Reading Law: The Interpretati on of Lega l Texts 234, 237 (2012); Story, Commentar ies § 427, at

411. It is a h igh ly demanding doctr ine— deliberately so, for

judges have no license to rewr ite a law's ter ms just because they happen to th ink different ones more sensible. And, telling ly, no one th inks th is law produces anyth ing like an absurd resu lt that might ca ll for a judicia l remedy. In fact, the gover nment affr matively d isavo ws any reliance on ab- surdity doctr ine. See Br ief for United States 36. Instead, it only gestures vag uely in the directi on of “nonsensica l” re- su lts and asks us to r un w ith the idea. As if we cou ld tinker

Cite as: 601 U. S. 124 (2024) 181 Gorsuch, J., dissenting

w ith Congress's work on the basis of some newly fash i oned “absurdity-lite” doctr ine.

There is a reason why the gover nment does not attempt an arg ument actua lly grounded on absurdity doctr ine. Its core complaint is that the natura l reading of the law does not, w ith suffcient precisi on, separate “more from less ser i- ous pr i or offenders. ” An te, at 148; Br ief for United States

21. But, of necessity, Congress of ten deploys “standardized

for mu la[s] ” or checklists, like the one found in paragraph (f )(1), that “are by their nature over- and under-inclusive. ” Ran so m v. FIA Card Se r v ices, N. A., 562 U. S. 61, 78 (2011). And because Congress may rati ona lly prefer these approaches for var i ous reasons, including their ease of administrati on, th is Cour t has long held that we w ill not second-g uess them merely because they may produce some “oddit[ies] ” or “anoma lies. ” Ib i d.; see Rodr i guez v. Un i ted Sta tes, 480 U. S. 522, 526 (1987) (pe r cur i am ) (“Deciding what competing va lues w ill or w ill not be sacr ifced to the ach ieve- ment of a par ticu lar objective is the very essence of leg isla- tive choice”).8 If, as the gover nment supposes, a seeming ly anoma lous resu lt a lone cou ld unsettle a statute, it wou ld face its own troubles, too. Under its implicit distr ibuti on theory, an indi- vidua l who previ ously committed a nonvi olent offense and received a sentence longer than 13 months (i. e., a 3-point offense) is categor ica lly inelig ible for relief. Meanwh ile, an individua l who committed the same cr ime but received a sen- tence equa l to or one day less than 13 months (i. e., a 2-point

8 In a footnote, the Court concedes that bo th sides read paragraph (f )(1) as announcing a standardized formula or checklist that inevitably produces some “anomalies. ” An te, at 148, n. 7. Yet the Court proceeds to re ject Mr. Pulsifer 's reading. Why? Only because it thinks that interpretation is just worse at performing the paragraph's “role. ” Ib id. Once more, the Court resorts to policy and purpose to escape its interpretive dilemma.

And once more, it fails to heed its own advice to Mr. Pulsifer that one “ interpretation is not better ” than another “ just because it would go fur- ther ” in advancing some view about the law's “role. ” An te, at 152.

Gorsuch, J., dissenting

offense) thanks to a slightly more lenient sentencing judge remains elig ible for relief. Rather than “unerr ing ly” enable the safety va lve to “separat[e] more ser i ous pr i or offenders from less ser i ous ones, ” an te, at 147, the gover nment's approach thus leaves much to happenstance and luck— an anoma lous resu lt indeed.

Retur n, then, to our actu al absurdity doctr ine and con- sider the gover nment's arg ument in its light. The gover n- ment worr ies that respecting paragraph (f )(1) as wr itten wou ld treat “more ser i ous” offenders too leniently. But in doing so, the gover nment ig nores what follows. A defend- ant who satisfes paragraph (f )(1) must still go on to satisfy paragraphs (f )(2)–(f )(5). And those provisi ons collectively operate to deny relief to vir tua lly anyone whose current of- fense involves any trace of vi olence.

Even if a “more ser i ous” offender cou ld somehow thread h is way through a ll those needles, too, another wou ld await.

The safety va lve instr ucts a sentencing cour t to fash i on a sentence “pursuant to [the] g uidelines. ” § 3553(f ). The g uidelines expressly account for a defendant's cr imina l h is- tory, and few wou ld accuse them of leniency toward those w ith a h istory of ser i ous offenses. In fact, defendants w ith sig nifcant cr imina l h istor ies of ten w ind up w ith a recom- mended g uidelines sentence h i ghe r than the other w ise- applicable mandatory minimum. See Br ief for Nati ona l Associati on of Federa l Defenders as Amicus Cur i ae 7– 8.

Sentencing cour ts may have the discreti on to vary or depar t from the g uidelines' recommended ranges. But Congress cou ld have rati ona lly tr usted cour ts to exercise that discre- ti on w ith an appreciati on for the fact that individua ls w ith ser i ous cr imina l h istor ies—such as the gover nment's hypo- thetica l defendant w ith many pr i or three-point vi olent of- fenses, see an te, at 147–148; Br ief for United States 23 — warrant equa lly ser i ous sentences. So, look ing to the law as a whole (as we must) and appreciating that Congress of ten leg islates using standardized for mu las or checklists that may

Cite as: 601 U. S. 124 (2024) 183 Gorsuch, J., dissenting

be over- and under-inclusive (again, as we must), there is noth ing approach ing an absurdity that might license us to rewr ite the First Step Act.9 In a fna l effor t to bolster the gover nment's case, the Cour t professes an entirely different concer n of its own. It claims to worry that the natura l reading of the law wou ld some- times be too harsh in operati on. Holding a defendant ineli- g ible for safety-va lve relief based on offenses that score “zero points, ” we are told, wou ld be “out-of-sync ” w ith the law's purpose. An te, at 146. But there is noth ing ab- surd here either. Subparagraph (A) provides that defend- ants are elig ible for relief as long as their past convicti ons do not y ield more than four cr imina l h istory points—a ca lcu- lati on that, as we have seen, does not include points associ- ated w ith old cr imes and cer tain other offenses. Subpara- graphs (B) and (C) provide that other defendants w ith more than four cr imina l h istory points are elig ible for relief too as long as they don't have anywhere in their past a ser i ous (3- point) offense and a weighty (2-point) vi olent offense— even if those offenses are (say) too old to contr ibute to their cr imi- na l h istory scores. So whatever unfair ness the Cour t may perceive in one par t of the safety va lve ( here, subparagraphs (B) and (C)) is diminished when considered in light of another ( here, subparagraph (A)). Some might prefer a different a r r a n g e m e n t , b u t t h e o n e C o n g r e s s o r d a i n e d i s h a r d l y absurd.

9 Nor is it clear that a “more serious” offender could even make it past paragraph (f )(1) to beg in with. The government seems to worry that a 3-point violent offense would not count as a 2-point violent offense under subparagraph (C), thus allowing some violent offenders to satisfy para- graph (f )(1) under its most natural constr uction. An te, at 147–148; Brief for United States 22 – 23. But, while Mr. Pulsifer has not pursued the point in his case and so it is not at issue before us, some lower courts have held that an ordinary person would read subparagraph (C)'s reference to a 2-point violent offense to embrace a violent offense carrying a t least that many points. See, e.g ., Lopez, 998 F. 3d, at 440 – 441, n. 10; see also Pace,

48 F. 4th, at 765 (Wood, J., dissenting).

Gorsuch, J., dissenting

If any law demonstrates the w isdom of our usua l r u les against elevating policy appea ls over plain text, it is th is one.

Under the ordinary meaning of the statute, it is possible some “more ser i ous” offenders may make it past paragraph (f )(1), and perhaps even end up receiving an individua lized sentence under g uidelines that hardly exh ibit solicitude for those w ith “more ser i ous” cr imina l h istor ies. Under the im- plicit distr ibuti on theory, in contrast, the availability of indi- vidua lized sentencing may depend on the happenstance of one extra day in pr ison. In the end, attempting to pick be- tween these two outcomes proves noth ing more than the futility of the exercise. However ar tfu lly the gover nment frames its dissatisfacti on w ith the text of the statute, we have neither the instituti ona l competence nor the constitu- ti ona l mandate “ to assess the relative mer its of different approaches” Congress cou ld have taken. Har tfo rd Unde r- w r i te rs In s. Co. v. Un io n Pl an te rs Bank, N. A., 530 U. S. 1, 13 (2000). Our role is a more modest one: “ [W]hen the stat- ute's lang uage is plain” and “ the dispositi on required by the text is not absurd, ” “ the sole functi on of the cour ts . . . is to enforce it according to its ter ms. ” Id., at 6 ( inter na l quota- ti on marks omitted). Because that is undoubtedly the case here, we must apply the safety va lve as wr itten.

As I see it, the gover nment hasn't come close to supply ing a lawfu l basis for depar ting from the law's ordinary meaning.

Suppose, though, at the end of th is long march through its inventive theor ies you remain unsure. Suppose you are lef t w ith a reasonable doubt about whether Mr. Pu lsifer or the gover nment has the better reading of the law. In circum- stances like that, another r u le of constr ucti on supplies an answer. It is lenity.

The r u le of lenity “ is perhaps not much less old than con- str ucti on itself. ” Un i ted Sta tes v. Wi l tbe rge r, 5 Wheat. 76, 95 (1820) (Marsha ll, C. J.); see Wooden v. Un i ted Sta tes, 595

U. S. 360, 388 (2022) (Gorsuch, J., concurr ing in judg ment)

Cite as: 601 U. S. 124 (2024) 185 Gorsuch, J., dissenting (citing The Adv en ture, 1 F. Cas. 202, 204 (No. 93) (CC Va.

1812) (Marsha ll, C. J.)). It requires cour ts to interpret am- big uous “pena l laws, ” including those concer ning sentencing, in favor of liber ty, not punishment. Wi l tbe rge r, 5 Wheat., at 95; Un i ted Sta tes v. Ba tchelde r, 442 U. S. 114, 121 (1979); Bi f ulco v. Un i ted Sta tes, 447 U. S. 381, 387 (1980).

Th is r u le enforces weighty constituti ona l va lues. Cour ts constr ue ambig uous pena l laws w ith lenity because a free nati on operates against a background presumpti on of indi- vidua l liber ty. See Wooden, 595 U. S., at 391– 392 (opini on of Gorsuch, J.). We resolve doubts about a cr imina l law's reach in favor of lenity, too, because in our federa l gover n- ment only the people's elected representatives, not their judges, are vested w ith the power to “defne a cr ime, and ordain its punishment. ” Wi l tbe rge r, 5 Wheat., at 95; accord, Crando n v. Un i ted Sta tes, 494 U. S. 152, 158 (1990); Wooden, 595 U. S., at 391– 392 (opini on of Gorsuch, J.).

Lenity protects vita l due process interests, as well, by en- sur ing individua ls fair notice of the consequences of their acti ons. Un i ted Sta tes v. L an i e r, 520 U. S. 259, 266 (1997); see M cBo y le v. Un i ted Sta tes, 283 U. S. 25, 27 (1931); Wooden, 595 U. S., at 389 – 391 (opini on of Gorsuch, J.). And lenity perfor ms still fur ther work, g uarding against the pos- sibility that judges might condemn unpopu lar individua ls to punishment on the streng th of their own views about com- mon sense, good public policy, or “no more than a g uess as to what Congress intended. ” L adne r v. Un i ted Sta tes, 358

So suppose you thought a reasonable doubt remained about how best to constr ue the First Step Act. In those circumstances, the answer cannot be to adopt an interpreta- ti on that restr icts safety-va lve relief to thousands more indi- vidua ls. The only per missible answer is one that favors liber ty.

Today, the Cour t does not hedge its doubts in favor of lib- er ty. Instead, it endorses the gover nment's implicit distr i-

Gorsuch, J., dissenting

buti on theory and elevates it over the law's ordinary and most natura l meaning.

It is a regrettable choice that requires us to abandon o n e p r i n c i p l e o f s t a t u t o r y i n t e r p r e t a t i o n a f t e r a n o t h e r.

We must read words into the law; we must delete others.

We must ig nore Congress's use of a constr ucti on that tends to avoid, not invite, questi ons about implicit distr ibuti on.

We must dismiss Congress's var iati ons in usage as sloppy mistakes. Never mind that Congress distr ibuted phrases expressly when it wanted them to repeat in the safety va lve.

Never mind that Congress used “or ” when it sought an eff- cient way to h inge elig ibility for relief based on a sing le char- acter istic. We must then read even more words yet into the law to manufacture a superfuity problem that does not exist.

We must elevate unexpressed congressi ona l purposes over statutory text. Fina lly, rather than resolve any reasonable doubt about statutory meaning in favor of the individua l, we must prefer a more punitive theory the gover nment only re- cently eng ineered.

Today, the Cour t indu lges each of these moves. All to what end? To deny some individua ls a chance—just a chance—at relief from mandatory minimums and a sentence that fts them and their circumstances. It is a chance Con- gress promised in the First Step Act, and it is a promise th is Cour t shou ld have honored. Respectfu lly, I dissent.

Reporter’s Note

The attached opinion has been revised to refect the usual publication and citation style of the United States Reports. The revised pag ination makes available the offcial United States Reports citation in advance of publication. The syllabus has been prepared by the Reporter of Decisions for the convenience of the reader and constitutes no part of the opinion of the Court. A list of counsel who argued or fled briefs in this case, and who were members of the bar of this Court at the time this case was argued, has been inserted following the syllabus. Other revisions may include adjustments to formatting, captions, citation form, and any errant punctuation. The following additional edits were made:

None