with whom Justice Thomas joins, concurring in part and concurring in the judgment.
In United States v. Booker, 543 U. S. 220 (2005), five Justices of this Court, I among them, held that our previous decision in Blakely v. Washington, 542 U. S. 296 (2004), applied to sentences imposed under the Federal Sentencing Guidelines because those Guidelines were mandatory and binding on judges. See 543 U. S., at 233-234, 243-244. We thus reaffirmed that “[a]ny fact (other than a prior conviction) which is necessary to support a sentence exceeding the maximum authorized by the facts established by a plea of guilty or a jury verdict must be admitted by the defendant or proved to a jury beyond a reasonable doubt.” Id., at 244. In response to this constitutional holding, a different majority of five Justices held that the appropriate remedy was to make the Guidelines nonmandatory in all cases and to review sentences on appeal only for reasonableness. See id., at 258-265. I disagreed with the Court’s remedial choice, believing instead that the proper remedy was to maintain the mandatory character of the Guidelines and simply to require, for that small category of cases in which a fact was legally essential to the sentence imposed, that the fact be proved to a jury beyond a reasonable doubt or admitted by the defendant. See id., at 272-291 (Stevens, J., joined by Scalia and Souter, JJ., dissenting in part).
I do not mean to reopen that debate. As a matter of statutory stare decisis, I accept Booker’s remedial holding that district courts are no longer bound by the Guidelines and that appellate courts should review the sentences imposed for reasonableness. As should be clear from our need to decide the case today, however, precisely what “reasonableness” review entails is not dictated by Booker. As I lamented then, “[t]he worst feature of the scheme is that no one knows — and perhaps no one is meant to know — how advisory Guidelines and ‘unreasonableness’ review will function in practice.” Id., at 311 (Scalia, J., dissenting in part).
*369Earlier this Term, the Court intensified its silence when it declined to flesh out what it had in mind in the face of an argument that the form of reasonableness review had constitutional implications. In Cunningham v. California, 549 U. S. 270 (2007), Justice Alito defended the constitutionality of California’s sentencing system in part by arguing that, even post-Booker, some federal sentences will be upheld as reasonable only if the judge makes additional findings of fact beyond those encompassed by the jury verdict or guilty plea. 549 U. S., at 309, and n. 11 (dissenting opinion). The Cunningham majority’s response, much like the Booker remedial opinion, was cryptic. While the Court did not explain why Justice Alito was incorrect, it strongly intimated that his premise was wrong: that he had erroneously “ ‘anticipate!/!]’ ” how “reasonableness review operates in practice.” Cunningham, 549 U. S., at 293, n. 15. Because that question is squarely presented in this case that was then pending, the Court found it “neither necessary nor proper ... to join issue with Justice Alito on this matter,” suggesting that all would be revealed in the opinion we issue today. See id., at 288, n. 13.
Today has arrived, and the Court has broken its promise. Nothing in the Court’s opinion explains why, under the advisory Guidelines scheme, judge-found facts are never legally necessary to justify the sentence. By this I mean the Court has failed to establish that every sentence which will be imposed under the advisory Guidelines scheme could equally have been imposed had the judge relied upon no facts other than those found by the jury or admitted by the defendant. In fact, the Court implicitly, but quite plainly, acknowledges that this will not be the case, by treating as a permissible post-jBooker claim petitioner’s challenge of his within-Guidelines sentence as substantively excessive. See ante, at Part IV. Under the scheme promulgated today, some sentences reversed as excessive will be legally authorized in later cases only because additional judge-found facts are *370present; and, as Justice Alito argued in Cunningham, some lengthy sentences will be affirmed (i e., held lawful) only because of the presence of aggravating facts, not found by the jury, that distinguish the case from the mine run. The Court does not even attempt to explain how this is consistent with the Sixth Amendment.
No explanation is given because no explanation is possible. The Court has reintroduced the constitutional defect that Booker purported to eliminate. I cannot acquiesce in this course. If a sentencing system is permissible in which some sentences cannot lawfully be imposed by a judge unless the judge finds certain facts by a preponderance of the evidence, then we should have left in place the compulsory Guidelines that Congress enacted, instead of imposing this jerry-rigged scheme of our own. In order to avoid the possibility of a Sixth Amendment violation, which was the object of the Booker remedy, district courts must be able, without finding any facts not embraced in the jury verdict or guilty plea, to sentence to the maximum of the statutory range. Because, therefore, appellate courts cannot reverse within-range sentences for being too high; and because no one would contend that Congress intended that sentences be reviewed only for being too low; I would hold that reasonableness review cannot contain a substantive component at all. I believe, however, that appellate courts can nevertheless secure some amount of sentencing uniformity through the procedural reasonableness review made possible by the Booker remedial opinion.
I
A
The Sixth Amendment requires that “[a]ny fact (other than a prior conviction) which is necessary to support a sentence exceeding the maximum authorized by the facts established by a plea of guilty or a jury verdict must be admitted by the defendant or proved to a jury beyond a reasonable doubt.” *371Booker, 543 U. S., at 244. Two hypotheticals will suffice to reveal why the notion of excessive sentences within the statutory range, and the ability of appellate courts to reverse such sentences, inexorably produces, in violation of the Sixth Amendment, sentences whose legality is premised on a judge's finding some fact (or combination of facts) by a preponderance of the evidence.
First, consider two brothers with similar backgrounds and criminal histories who are convicted by a jury of respectively robbing two banks of an equal amount of money. Next assume that the district judge finds that one brother, fueled by racial animus, had targeted the first bank because it was owned and operated by minorities, whereas the other brother had selected the second bank simply because its location enabled a quick getaway. Further assume that the district judge imposes the statutory maximum upon both brothers, basing those sentences primarily upon his perception that bank robbery should be punished much more severely than the Guidelines base level advises, but explicitly noting that the racially biased decisionmaking of the first brother further justified his sentence. Now imagine that the appellate court reverses as excessive only the sentence of the non-racist brother. Given the dual holdings of the appellate court, the racist has a valid Sixth Amendment claim that his sentence was reasonable (and hence lawful) only because of the judicial finding of his motive in selecting his victim.1
Second, consider the common case in which the district court imposes a sentence within an advisory Guidelines range that has been substantially enhanced by certain judge-found facts. For example, the base offense level for robbery under the Guidelines is 20, United States Sentencing Commission, Guidelines Manual §2B3.1(a) (Nov. 2006), which, *372if the defendant has a criminal history of I, corresponds to an advisory range of 33-41 months, id., eh. 5, pt. A, Sentencing Table. If, however, a judge finds that a firearm was discharged, that a victim incurred serious bodily injury, and that more than $5 million was stolen, then the base level jumps by 18, §§2B3.1(b)(2), (3), (7), producing an advisory range of 235-293 months, id., ch. 5, pt. A, Sentencing Table. When a judge finds all of those facts to be true and then imposes a within-Guidelines sentence of 293 months, those judge-found facts, or some combination of them, are not merely facts that the judge finds relevant in exercising his discretion; they are the legally essential predicate for his imposition of the 293-month sentence. His failure to find them would render the 293-month sentence unlawful. That is evident because, were the district judge explicitly to find none of those facts true and nevertheless to impose a 293-month sentence (simply because he thinks robbery merits seven times the sentence that the Guidelines provide) the sentence would surely be reversed as unreasonably excessive.
These hypothetical are stylized ways of illustrating the basic problem with a system in which district courts lack full discretion to sentence within the statutory range. Under such a system, for every given crime there is some maximum sentence that will be upheld as reasonable based only on the facts found by the jury or admitted by the defendant. Every sentence higher than that is legally authorized only by some judge-found fact, in violation of the Sixth Amendment. Appellate courts’ excessiveness review will explicitly or implicitly accept those judge-found facts as justifying sentences that would otherwise be unlawful. The only difference between this system and the pre-Booker mandatory Guidelines is that the maximum sentence based on the jury verdict or guilty plea was specified under the latter but must be established by appellate courts, in case-by-case fashion, under the former. Ibis is, if anything, an additional constitutional disease, not a constitutional cure.
*373To be clear, I am not suggesting that the Sixth Amendment prohibits judges from ever finding any facts. We have repeatedly affirmed the proposition that judges can find facts that help guide their discretion within the sentencing range that is authorized by the facts found by the jury or admitted by the defendant. See, e. g., Booker, supra, at 233; Apprendi v. New Jersey, 530 U. S. 466, 481 (2000). But there is a fundamental difference, one underpinning our entire Apprendi jurisprudence, between facts that must be found in order for a sentence to be lawful, and facts that individual judges choose to make relevant to the exercise of their discretion. The former, but not the latter, must be found by the jury beyond a reasonable doubt in order “to give intelligible content to the right of jury trial.” Blakely, 542 U. S., at 305.2
1 am also not contending that there is a Sixth Amendment problem with, the Court’s affirmation of a presumption of reasonableness for within-Guidelines sentences. I agree with the Court that such a presumption never itself makes judge-found facts legally essential to the sentence imposed, since it has no direct relevance to whether the sentence would have been unreasonable in the absence of any judge-found facts. See ante, at 352-354.3 Nor is my claim that the Sixth *374Amendment was violated in this case, for petitioner cannot demonstrate that his relatively low sentence would have been unreasonable if the District Court had relied on nothing but jury-found or admitted facts.
Rather, my position is that there will inevitably be some constitutional violations under a system of substantive reasonableness review, because there will be some sentences that will be upheld as reasonable only because of the existence of judge-found facts. Booker itself reveals why that reality dooms the construct of reasonableness review established and applied by today’s opinion. Booker made two things quite plain. First, reasonableness is the standard of review implicitly contained within the Sentencing Reform Act of 1984 (SRA). 543 U. S., at 260-261. Second, Congress wanted a uniform system of sentencing review, rather than different schemes depending on whether there were Sixth Amendment problems in particular cases. Id., at 265-267. Thus, if the contours of reasonableness review must be narrowed in some cases because of constitutional concerns, then they must be narrowed in all cases in light of Congress’s desire for a uniform standard of review. The Justices composing today’s Court were in total agreement with this principle of statutory interpretation the day Booker was decided:
“[W]hen deciding which of two plausible statutory constructions to adopt, a court must consider the necessary *375consequences of its choice. If one of them would raise a multitude of constitutional problems, the other should prevail — whether or not those constitutional problems pertain to the particular litigant before the Court.” Clark v. Martinez, 543 U. S. 371,380-381 (2005) (opinion for the Court by Scalia, J., joined by, inter alios, Stevens, Kennedy, Ginsburg, and Breyer, JJ.).
Yet they now adopt substantive reasonableness review without offering any rebuttal to my charge of patent constitutional flaw inherent in such review. The one comfort to be found in the Court’s opinion — though it does not excuse the failure to apply Martinez’s interpretive principle — is that it does not rule out as-applied Sixth Amendment challenges to sentences that would not have been upheld as reasonable on the facts encompassed by the jury verdict or guilty plea. Ante, at 353-354; ante, at 365-366 (Stevens, J., joined by Ginsburg, J., concurring).4
B
Had the Court bothered to frame objections to the constitutional analysis undertaken above, there are four conceivable candidates.
1
The most simplistic objection is that the Sixth Amendment is not violated because the judge-found facts are made le*376gaily necessary by the decision of appellate courts rather than the decision of Congress. This rebuttal errs both in premise and in conclusion.
The premise is wrong because, according to the remedial majority in Booker, the facts that excessiveness review renders legally essential are made such by Congress. Reasonableness is the standard of review implicitly contained within 18 U.S.C. §3742 (2000 ed. and Supp. IV). See Booker, supra, at 260-261. But the Sixth Amendment would be violated even if appellate courts really were exercising some type of common-law power to prescribe the facts legally necessary to support specific sentences. Neither Apprendi nor any of its progeny suggests that violation of the Sixth Amendment depends upon what branch of government has made the prescription. To the contrary, Booker flatly rejected the argument that the mandatory Guidelines were constitutional because it was the Sentencing Commission rather than Congress that specified the facts essential to punishment. See 543 U. S., at 237-239. And for good reason. The Sixth Amendment is “a reservation of jury power.” Blakely, 542 U. S., at 308. It makes no difference whether it is a legislature, a Sentencing Commission, or an appellate court that usurps the jury’s prerogative. Were it otherwise, this Court could prescribe that the only reasonable sentences are those consistent with the same mandatory Guidelines that Booker invalidated. And the California Supreme Court could effectively reverse our decision in Cunningham simply by setting aside as unreasonable any trial-court sentence that does not conform to pre-Cunningham California law.
2
The next objection minimizes the extent to which excessiveness review makes judge-found facts legally essential to punishment. If appellate courts will uphold, based only on the facts found by the jury, a district court’s decision to impose all but the lengthiest sentences, then the number of *377sentences that are legally dependent on judge-found facts will be quite small. Thus, the argument goes, there is no reason to prohibit substantive reasonableness review altogether: Absent a claim that such review creates a constitutional problem in a given case, why prohibit it? I have already explained why this line of defense is inconsistent with established principles of statutory interpretation. See supra, at 374-375. But even on its own terms, the defense is inconsistent with Booker because reasonableness review is an improper and inadequate remedial scheme unless it ensures that judge-found facts are never legally necessary to justify the sentence imposed under the advisory Guidelines.
The mandatory Guidelines system that was invalidated in Booker had the same attribute of producing unconstitutional results in only a small proportion of cases. Because of guilty pleas and Guidelines ranges that did not depend on judge-found facts, the overwhelming majority of sentences imposed under the pre-Booker federal system were perfectly in accord with the Sixth Amendment. See Booker, 543 U. S., at 248; id., at 275-277 (Stevens, J., dissenting in part). Booker nevertheless excised key statutory provisions governing federal sentencing, in order to eliminate constitutional violations entirely. If our conjured-up system does not accomplish that goal, then by what right have we supplanted the congressionally enacted mandatory Guidelines?
If it is true that some sentences under today’s Court-prescribed system will still violate the Sixth Amendment, nonetheless allowing the system to go forward will produce chaos. Most cases do not resemble my stylized hypothetical, and ordinarily defendants and judges will be unable to figure out, based on a comparison of the facts in their case with the facts of all of the previously decided appellate cases, whether the sentence imposed would have been upheld as reasonable based only on the facts supporting the jury verdict or guilty plea. That will not stop defendants from mak*378ing the argument, however, and the Court certainly has not foreclosed them from trying. See supra, at 375, and n. 4. Judges will have in theory two options: create complicated charts and databases, based on appellate precedents, to ascertain what facts are legally essential to justify what sentences; or turn a deaf ear to these claims, though knowing full well that some of them are justified. I bet on the latter.5 Things were better under the mandatory Guidelines system, where every judge could readily identify when the Sixth Amendment was being violated, and could rule accordingly.
3
Proponents of substantive reasonableness review could next argue that actual sentencing involves the consideration of dozens of different facts in order to make an individualized determination about each defendant. In the real world, they would contend, it is difficult, if not impossible, to determine whether any given fact was legally essential to the punishment imposed. But identifying the particular fatal fact is not necessary to identifying a constitutional violation. In the second hypothetical given above, for example, it is not possible to say which single fact, or which combination of facts, sufficed to bring the sentence within the bounds of the “reasonable.” But it is possible to say (indeed, it must be said) that some judge-found fact or combination of facts had that effect — and that suffices to establish a Sixth Amendment violation.
“Whether the judge’s authority to impose an enhanced sentence depends on finding a specified fact . . . , one *379of several specified facts ... , or any aggravating fact ... , it remains the case that the jury’s verdict alone does not authorize the sentence. The judge acquires that authority only upon finding some additional fact.” Blakely, 542 U. S., at 305.
4
The last conceivable defense of the Guidelines-light would be to wrap them in the mantle of history and tradition.
“[W]hen a practice not expressly prohibited by the text of the Bill of Rights bears the endorsement of a long tradition of open, widespread, and unchallenged use that dates back to the beginning of the Republic, we have no proper basis for striking it down. Such a venerable and accepted tradition is not to be laid on the examining table and scrutinized for its conformity to some abstract principle of [constitutional] adjudication devised by this Court. To the contrary, such traditions are themselves the stuff out of which the Court’s principles are to be formed.” Rutan v. Republican Party of Ill., 497 U. S. 62, 95-96 (1990) (Scalia, J., dissenting) (footnote omitted).
This consideration has no application here. In the federal system, prior to the SRA, substantive appellate review of a district court’s sentencing discretion essentially did not exist. See, e. g., Dorszynski v. United States, 418 U. S. 424, 431 (1974) (noting “the general proposition that once it is determined that a sentence is within the limitations set forth in the statute under which it is imposed, appellate review is at an end”); id., at 443 (“[W]ell-established doctrine bars review of the exercise of sentencing discretion”). As for state appellate review of sentences, as late as 1962, at least 39 States did not permit appellate courts to modify sentences imposed within the statutory limits. See Appellate Review of Sentences, A Symposium at the Judicial Conference of the *380United States Court of Appeals for the Second Circuit, 32 F. R. D. 249, 260 (1962). It would be an exaggeration to say that history reflects an established understanding that appellate review of excessive sentences conflicts with the Sixth Amendment. But it would also be an exaggeration to say that the historical pedigree of substantive appellate review of sentencing is so strong and clear as to overcome the basic principle underlying the jury-trial right applied by this Court in Apprendi, Blakely, Booker, and Cunningham.
C
A final defense of substantive reasonableness review would be to invoke the intent of Congress or of the Booker remedial opinion. As for congressional intent: Of course Congress intended that judge-found facts be legally essential to the punishment imposed; that was the whole reason the mandatory Guidelines violated the Sixth Amendment. If we are now to indulge a newfound respect for unconstitutional congressional intent, we should reimpose the mandatory Guidelines system. The quasi-Guidelines system the Court creates today manages to contravene both congressional intent and the Sixth Amendment.
As for the “intent” of the Booker remedial opinion: That opinion purported to be divining congressional intent in light of what the Sixth Amendment compelled. See 543 U. S., at 263-265. Absent some explanation of why substantive reasonableness review does not cause judge-found facts to justify greater punishment than the jury’s verdict or the defendant’s guilty plea would sustain, I fail to understand how such review could possibly have been intended by all five Justices who composed the Booker remedial majority. After all, at least one of them did not intend “to override Blakely, and to render academic the entire first part of Booker itself,” and has confirmed that “[t]here would have been no majority in Booker for the revision of Blakely es*381sayed in [Justice Auto’s Cunningham] dissent.” Cunningham, 549 U. S., at 293, n. 15 (opinion for the Court by Ginsburg, J.).
II
Abandoning substantive reasonableness review does not require a return to the pre-SRA regime that the Booker remedial opinion sought to avoid. See 543 U. S., at 263-265. As I said at the outset, I believe it is possible to give some effect to the Booker remedial opinion and the purposes that it sought to serve while still avoiding the constitutional defect identified in the Booker merits opinion. Specifically, I would limit reasonableness review to the sentencing procedures mandated by statute.
A
A central feature of the Booker remedial opinion was its conclusion that the SRA was not completely inseverable. See id., at 258-265. As a result, the Sentencing Commission “remains in place, writing Guidelines, collecting information about actual district court sentencing decisions, undertaking research, and revising the Guidelines accordingly.” Id., at 264. Likewise, sentencing courts remain obligated to consider the various factors delineated in 18 U. S. C. § 3553(a) (2000 ed., Supp. IV), including the now-advisory Guidelines range. 543 U. S., at 259-260. And they are still instructed by that subsection to “impose a sentence sufficient, but not greater than necessary, to comply with the purposes set forth in paragraph (2) of [that] subsection.” Significantly, § 3553(c) (2000 ed. and Supp. IV) continues to require that district courts give reasons for their sentencing decisions, a requirement the requisite detail of which depends on whether the sentence is: (1) within the advisory Guidelines range; (2) within an advisory Guidelines range that spans more than 24 months; or (3) outside the advisory Guidelines *382range. These explanations, in turn, help the Commission revise the advisory Guidelines to reflect actual sentencing practices consistent with the statutory goals. See Booker, supra, at 264 (citing 28 U. S. C. § 994 (2000 ed. and Supp. IV)).
Booker’s retention of these statutory procedural provisions furthered the congressional purpose of “iron[ing] out sentencing differences,” 543 U. S., at 263, and “avoiding] excessive sentencing disparities,” id., at 264. It is important that appellate courts police their observance. Booker excised the provision of the SRA containing the standards for appellate review, see id., at 260 (invalidating 18 U. S. C. § 3742(e) (2000 ed. and Supp. IV)), but the remedial majority’s creation of reasonableness review gave appellate courts the necessary means to reverse a district court that: appears not to have considered § 3553(a); considers impermissible factors; selects a sentence based on clearly erroneous facts; or does not comply with §3553(c)’s requirement for a statement of reasons.6 In addition to its direct effect on sentencing uniformity, this procedural review will indirectly produce, over time, reduction of sentencing disparities. By ensuring that district courts give reasons for their sentences, and more specific reasons when they decline to follow the advisory Guidelines range, see § 3553(c)(2) (2000 ed., Supp. IV), appellate courts will enable the Sentencing Commission to perform its function of revising the Guidelines to reflect the desirable sentencing practices of the district courts. See Booker, supra, at 264 (citing 28 U. S. C. § 994 (2000 ed. and Supp. IV)). And as that occurs, district courts will have less *383reason to depart from the Commission’s recommendations, leading to more sentencing uniformity.7
One possible objection to procedural review that the Booker remedial opinion appears not to have considered is 18 U. S. C. § 3742(f) (2000 ed., Supp. IV), which limits appellate courts to reversing sentences that are imposed “in violation of law” or “as a result of an incorrect application of the sentencing guidelines,” §3742(f)(1), or that fall in certain categories and are either “too high” or “too low,” § 3742(f)(2).8 But, as I noted in Booker, § 3742(e) and § 3742(f) are inextricably intertwined: Having excised § 3742(e)’s provisions setting forth the standards for appellate review, it is nonsensical to continue to apply § 3742(f )’s provisions governing the “Decision and Disposition” of appeals, which clearly track those now-excised standards. See 543 U. S., at 306-307 (Scalia, J., dissenting in part). I would hold that § 3742(f) is “incapable of functioning independently” of the provisions excised in Booker, and is thus inseverable from them. See Alaska *384Airlines, Inc. v. Brock, 480 U. S. 678, 684 (1987); 2 N. Singer, Sutherland Statutes and Statutory Construction § 44:4, p. 576 (6th ed. 2001) (“Even where part of an act is independent and valid, other parts which are not themselves substantively invalid but have no separate function to perform independent of the invalid portions of the act are also held invalid”).
B
Applying procedural review in this case does not require much further discussion on my part. I join Part III of the Court’s opinion. See ante, at 356-359.
* * *
The Court’s decision today leaves unexplained why the mandatory Guidelines were unconstitutional, but the Court-created substantive-review system that contains the same potential for Sixth Amendment violation is not. It is irresponsible to leave this patent inconsistency hanging in the air, threatening in the fiiture yet another major revision of Guidelines practices to which the district courts and courts of appeals will have to adjust. Procedural review would lay the matter to rest, comporting with both parts of the Booker opinion and achieving the maximum degree of sentencing uniformity on the basis of judge-found facts that the Constitution permits.
Of course, it may be that some fact other than racial animus would also have sufficed to sustain the increased sentence. But it is undeniable that in the case at hand the judicial finding of racial animus filled that role. See Blakely v. Washington, 542 U. S. 296, 305 (2004).
For similar reasons, I recognize that the Sixth Amendment problem with reasonableness review is created only by the lack of district court discretion to impose high sentences, since eliminating discretion to impose low sentences is the equivalent of judicially creating mandatory minimums, which are not a concern of the Sixth Amendment. See Harris v. United States, 536 U. S. 545, 568-569 (2002). But since reasonableness review should not function as a one-way ratchet, United States v. Booker, 543 U. S. 220, 257-258, 266 (2005), we must forswear the notion that sentences can be too low in light of the need to abandon the concept that sentences can be too high.
For this reason, I do not join Justice Souter’s dissent. He wishes to give “district courts [assurance] that the entire sentencing range set by statute is available to them.” Post, at 391. That is a proper goal—indeed, an essential one to prevent the Booker remedy from effectively overturning Apprendi and Blakely. But eliminating the presumption of rea*374sonableness will not achieve it. In those Circuits that already decline to employ the presumption, a within-Guidelines sentence has never been reversed as substantively excessive, Brief for New York Council of Defense Lawyers as Amicus Curiae 5, refuting the belief that mere elimination of the presumption will destroy the “gravitational pull,” post, at 390 (Souter, J., dissenting), to stay safely within the Guidelines. The only way to assure district courts that they can deviate from the advisory Guidelines, and to ensure that judge-found facts are never legally essential to the sentence, is to prohibit appellate courts from reviewing the substantive sentencing choices made by district courts.
The Court suggests that my reliance on hypothetical indicates that its interpretation of reasonableness will not create a multitude of constitutional problems. Ante, at 353-354; see also ante, at 366 (Stevens, J., concurring). Setting aside the question whether the volume of constitutional violations has any relevance to the application of Martinez’s interpretive principle, the Court is wrong to think that the constitutional problem today’s opinion ignores is hypothetical, merely because I have used hypotheticals to describe it. It is all too real that advisory Guidelines sentences routinely change months and years of imprisonment to decades and centuries on the basis of judge-found facts — as Booker itself recognized, see 543 U. S., at 236-237 (citing, inter alia, a case in which a defendant’s sentence increased from 57 months to 155 years).
Perhaps I am too cynical. At least one conscientious District Judge has decided to shoulder the burden of ascertaining what the maximum reasonable sentence is in each case based only on the verdict and appellate precedent, correctly concluding that this is the only way to eliminate Sixth Amendment problems after Cunningham v. California, 549 U. S. 270 (2007), if Booker mandates substantive reasonableness review. See United States v. Griffin, 494 F. Supp. 2d 1, 12-14 (D. Mass. 2007) (Young, J.) (Sentencing Memorandum).
“Substance” and “procedure” are admittedly chameleon-like terms. See Sun Oil Co. v. Wortman, 486 U. S. 717, 726-727 (1988). As the text indicates, my use of the term “procedure” here includes the limiting of sentencing factors to permissible ones — as opposed to using permissible factors but reaching a result that is “substantively” wrong. I therefore disagree with Justice Stevens that a district court which discriminates against Yankees fans is acting in a procedurally “impeccable” way. Ante, at 365 (concurring opinion).
Courts must resist, however, the temptation to make procedural review more stringent because substantive review is off the table. The judicial role when conducting severability analysis is limited to determining whether the balance of a statute that contains an unconstitutional provision is capable “of functioning independently.” Alaska Airlines, Inc. v. Brock, 480 U. S. 678, 684 (1987). Courts have no power to add provisions that might be desirable now that certain provisions have been excised. Thus, when engaging in reasonableness review to determine whether the district court has complied with the various procedures in § 3553, an appellate court cannot subject the district court to any greater requirements than existed pre-Booker.
I say “possible” because one could claim that the failure to comply with 18 U. S. C. §3553’s procedural requirements results in a sentence imposed in violation of law, and thereby covered by § 3742(f)(1). But § 3742(f)(l)’s applicability to such procedural errors is called into question by § 3742(f)(2) (2000 ed., Supp. IV), which specifically addresses sentences where “the district court failed to provide the required statement of reasons [mandated by § 3553(c)(2)].” For the reasons specified in the text, however, I see no need to grapple, post-Booker, with the proper interpretation of § 3742(f).