*623Opinion

MOSK, J.

In this proceeding for writ of prohibition we are called upon to decide whether an unborn but viable fetus is a “human being” within the meaning of the California statute defining murder (Pen. Code, § 187). We conclude that the Legislature did not intend such a meaning, and that for us to construe the statute to the contrary and apply it to this petitioner would exceed our judicial power and deny petitioner due process of law.

The evidence received at the preliminary examination may be summarized as follows: Petitioner and Teresa Keeler obtained an interlocutory decree of divorce on September 27, 1968. They had been married for 16 years. Unknown to petitioner, Mrs. Keeler was then pregnant by one Ernest Vogt, whom she had met earlier that summer. She subsequently began living with Vogt in Stockton, but concealed the fact from petitioner. Petitioner was given custody of their two daughters, aged 12 and 13 years, and under the decree Mrs. Keeler had the right to take the girls on alternate weekends.

On February 23, 1969, Mrs. Keeler was driving on a narrow mountain road in Amador County after delivering the girls to their home. She met petitioner driving in the opposite direction; he blocked the road with his car, and she pulled over to the side. He walked to her vehicle and began speaking to her. He seemed calm, and she rolled down her window to hear him. He said, “I hear you’re pregnant. If you are you had better stay away from the girls and from here.” She did not reply, and he opened the car door; as she later testified, “He assisted me out of the car. . . . [I]t wasn’t roughly at this time.” Petitioner then looked at her abdomen and became “extremely upset.” He said, “You sure are. I’m going to stomp it out of you.” He pushed her against the car, shoved his knee into her abdomen, and struck her in the face with several blows. She fainted, and when she regained consciousness petitioner had departed.

Mrs. Keeler drove back to Stockton, and the police and medical assistance were summoned. She had suffered substantial facial injuries, as well as extensive bruising of the abdominal wall. A Caesarian section was performed and the fetus was examined in útero. Its head was found to be severely fractured, and it was delivered stillborn. The pathologist gave as his opinion that the cause of death was skull fracture with consequent cerebral hemorrhaging, that death would have been immediate, and that the injury could have been the result of force applied to the mother’s abdomen. There was no air in the fetus’ lungs, and the umbilical cord was intact.

*624Upon delivery the fetus weighed five pounds and was 18 inches in length. Both Mrs. Keeler and her obstetrician testified that fetal movements had been observed prior to February 23, 1969. The evidence was in conflict as to the estimated age of the fetus;1 the expert testimony on the point, however, concluded “with reasonable medical certainty” that the fetus had developed to the stage of viability, i.e., that in the event of premature birth on the date in question it would have had a 75 percent to 96 percent chance of survival.

An information was filed charging petitioner, in count I, with committing the crime of murder (Pen. Code, § 187) in that he did “unlawfully kill a human being, to wit Baby Girl Vogt, with malice aforethought.” In count II petitioner was charged with wilful infliction of traumatic injury upon his wife (Pen. Code, § 273d), and in count III, with assault on Mrs. Keeler by means of force likely to produce great bodily injury (Pen. Code, 5 245). His motion to set aside the information for lack of probable cause (Pen. Code, § 995) was denied, and he now seeks a writ of prohibition; as will appear, only the murder count is actually in issue. Pending our disposition of the matter, petitioner is free on bail.

I

Penal Code section 187 provides: “Murder is the unlawful killing of a human being, with malice aforethought.” The dispositive question is whether the fetus which petitioner is accused of killing was, on February 23, 1969, a “human being” within the meaning of the statute. If it was not, petitioner cannot be charged with its “murder” and prohibition will lie.

Section 187 was enacted as part of the Penal Code of 1872. Inasmuch as the provision has not been amended since that date, we must determine the intent of the Legislature at the time of its enactment. But section 187 was, in turn, taken verbatim from the first California statute defining murder, part of the Crimes and Punishments Act of 1850. (Stats. 1850, ch. 99, § 19, p. 231.)2 Penal Code section 5 (also enacted in 1872) *625declares: “The provisions of this code, so far as they are substantially the same as existing statutes, must be construed as continuations thereof, and not as new enactments.” We begin, accordingly, by inquiring into the intent of the Legislature in 1850 when it first defined murder as the unlawful and malicious killing of a “human being.”

It will be presumed, of course, that in enacting a statute the Legislature was familiar with the relevant rules of the common law, and, when it couches its enactment in common law language, that its intent was to continue those rules in statutory form. (Baker v. Baker (1859) 13 Cal. 87, 95-96; Morris v. Oney (1963) 217 Cal.App.2d 864, 870 [32 Cal.Rptr. 88].) This is particularly appropriate in considering the work of the first session of our Legislature: its precedents were necessarily drawn from the common law, as modified in certain respects by the Constitution and by legislation of our sister states.* *3

We therefore undertake a brief review of the origins and development of the common law of abortional homicide. (For a more detailed treatment, see Means, The Law of New York Concerning Abortion and the Status of the Foetus, 1664—1968: A Case of Cessation of Constitutionality (1968) 14 N.Y.L.F. 411 [hereinafter cited as Means]; Stern, Abortion: Reform and the Law (1968) 59 J.Crim.L., C. & P.S. 84; Quay, Justifiable Abortion— Medical and Legal Foundations II (1961) 49 Geo.L.J. 395.) From that inquiry it appears that by the year 1850—the date with which we are concerned—an infant could not be the subject of homicide at common law unless it had been born alive.4 Perhaps the most influential statement of the “born alive” rule is that of Coke, in mid-17th century: “If a woman be quick with childe,5 and by a potion or otherwise killeth *626it in her wombe, or if a man beat her, whereby the childe dyeth in her body, and she is delivered of a dead childe, this is a great misprision [i.e., misdemeanor], and no murder; but if the childe be born alive and dyeth of the potion, battery, or other cause, this is murder; for in law it is accounted a reasonable creature, in rerum natura, when it is born alive.” (3 Coke, Institutes *50 (1648).) In short, “By Coke’s time, the common law regarded abortion as murder only if the foetus is (1) quickened, (2) bom alive, (3) lives for a brief interval, and (4) then dies.” (Means, at p. 420.) Whatever intrinsic defects there may have been in Coke’s work (see 3 Stephen, A History of the Criminal Law of England (1883) pp. 52-60), the common law accepted his views as authoritative. In the 18th century, for example, Coke’s requirement that an infant be bom alive in order to be the subject of homicide was reiterated and expanded by both Blackstone6 and Hale.7

Against this background, a series of infanticide prosecutions were brought in the English courts in mid-19th century. In each, a woman or her accomplice was charged with murdering a newborn child, and it was uniformly declared to be the law that a verdict of murder could not be returned unless it was proved the infant had been born alive. Thus in Rex v. Brain (1834) 6 Car. & P. 349, 350, 172 Eng.Reprint 1272, the court instructed the jury that “A child must be actually wholly in the world in a living state to be the subject of a charge of murder; but if it has been wholly born, and is alive, it is not essential that it should have breathed at the time it was killed; as many children are bom alive, and yet do not breathe for some time after their birth. But you must be satisfied that the child was wholly born into the world at the time it was killed, or you ought not to find the prisoner guilty of murder.” (Accord, Rex v. Boulton (1832) 5 Car. & P. 329, 172 Eng.Reprint 997; Rex v. Enoch (1833) 5 Car. & P. 539, 172 Eng.Reprint 1089; Rex v. Crutchley (1836) 7 Car. & P. 814, 173 Eng.Reprint 355; Rex v. Sellis (1836) 7 Car. &P. 850,173 Eng.Reprint *627370; Reg. v. Reeves (1839) 9 Car. & P. 25, 173 Eng.Reprint 724; Reg. v. Wright (1841) 9 Car. & P. 754, 173 Eng.Reprint 1039; Reg. v. Trilloe (1842) Car. & M. 650, 174 Eng.Reprint 674; see also cases collected in Atkinson, Life, Birth, and Livebirth (1904) 20 L.Q.Rev. 134, 139-145.)

Of these decisions, some pointed out that evidence of breathing is not conclusive because that function may begin before the infant is fully born (Boulton, Enoch, Sellis), while others observed that the infant can possess an “independent circulation”—one of the tests used to determine live birth—even though the umbilical cord may not yet be severed (Reeves, Trilloe). But all were in agreement that however live birth was to be proved, unless that event had occurred before the alleged criminal act there could be no conviction of homicide.

By the year 1850 this rule of the common law had long been accepted in the United States. As early as 1797 it was held that proof the child was born alive is necessary to support an indictment for murder (State v. McKee (Pa.) Addison 1), and the same rule was reiterated on the eve of the first session of our Legislature (State v. Cooper (1849) 22 N.J.L. 52). Although the precise issue in Cooper was whether an attempted abortion on a woman whose fetus had not yet “quickened” was a common law crime, the opinion begins by a recital of the common law rules on abortional homicide. In its argument the State took the position that attempted abortion was an offense against the person of the child, and the court replied that “the very point of inquiry is, whether that be at all an offense or not, and whether the child be in esse, so that any crime can be committed against its person. In regard to offences against the person of the child, a distinction is well settled between its condition before and after its birth. Thus, it is not murder to kill a child before it be born, even though it be killed in the very process of delivery.” (Id. at p. 54.) In support of this proposition, the court then set out in full each of the passages of Coke, Blackstone, and Hale quoted hereinabove. (Id. at pp. 54-55.)

While it was thus “well settled” in American case law that the killing of an unborn child was not homicide, a number of state legislatures in the first half of the 19th century undertook to modify the common law in this respect.8 The movement began when New York abandoned the common law of abortion in 1830. The revisers’ notes on that legislation recognized the existing rule,9 but nevertheless proposed a special feticide statute *628which, as enacted, provided that “The wilful killing of an unborn quick child, by any injury to the mother of such child, which would be murder if it resulted in the death of such mother, shall be deemed manslaughter in the first degree.” (N.Y. Rev. Stat. 1829, pt. IV, ch. 1, tit. 2, § 8, quoted in Means, at p. 443.) At the same time the New York Legislature enacted a companion section (§9) which, although punishing a violaton thereof as second degree manslaughter, was in essence an “abortion law” similar to those in force in most states today.10

In the years between 1830 and 1850 at least five other states followed New York and enacted, as companion provisions, (1) a statute declaring feticide to be a crime, punishable as manslaughter, and (2) a statute prohibiting abortion.* 11 In California, however, the pattern was not repeated. Much of the Crimes and Punishments Act of 1850 was based on existing New York statute law; but although a section proscribing abortion was included in the new Act (§ 45), the Legislature declined to adopt any provision defining and punishing a special crime of feticide.

We conclude that in declaring murder to be the unlawful and malicious killing of a “human being” the Legislature of 1850 intended that term to have the settled common law meaning of a person who had been born alive, and did not intend the act of feticide—as distinguished from abortion—to be an offense under the laws of California.

Nothing occurred between the years 1850 and 1872 to suggest that in adopting the new Penal Code on the latter date the Legislature entertained any different intent. The case law of our sister states, for example, remained consonant with the common law. In Abrams v. Foshee (Cole ed. 1856) 3 Iowa 274, 278, the court noted that Iowa’s feticide statute (Iowa Rev. Stat. 1843, § 10) had been repealed by the Iowa Code of 1851; it was contended *629that an unborn child is nevertheless “a human being, within the meaning of [Iowa Code] section 2508, which provides that whoever kills any human being, with malice aforethought, either express or implied, is guilty of murder.” The court observed that “notwithstanding the infant in ventre sa mere, is treated by the law for some purposes, as born, or as a human being, yet we are not aware that it has been so treated, so far as to make the act of its miscarriage murder, unless so declared by statute. . . . When the child is born, however, it becomes a human being, within the meaning of the law; and if it shall then die, by reason of any potions or bruises it received in the womb, it would be murder in those who administered or gave them, with a view of causing the miscarriage.” (Ibid.) Citing Blackstone, Coke, and other authorities, the court concluded that “an infant in ventre sa mere, is not a human being within the meaning of” the statute defining murder. (Id. at p. 279.)12

Any lingering doubt on this subject must be laid to rest by a consideration of the legislative history of the Penal Code of 1872. The Act establishing the California Code Commission (Stats. 1870, ch. 516, § 2, p. 774) required the commissioners to revise all statutes then in force, correct errors and omissions, and “recommend all such enactments as shall, in the judgment of the Commission, be necessary to supply the defects of and give completeness to the existing legislation of the State. . . .” In discharging this duty the statutory schemes of our sister states were carefully examined,13 and we must assume the commissioners had knowl*630edge of the feticide laws noted hereinabove.14 Yet the commissioners proposed no such law for California, and none has been adopted to this day.

That such an omission was not an oversight clearly appears, moreover, from the commissioners’ explanatory notes to Penal Code section 187. After quoting the definitions of murder given by Coke, Blackstone, and Hawkins, the commissioners conclude; “A child within its mother’s womb is not a ‘human being’ within the meaning of that term as used in defining murder. The rule is that it must be born.—Rex vs. Brain, 6 Car. & P., p. 349. That every part of it must have come from the mother before the killing of it will constitute a felonious homicide.—Rex vs. Brain, 6 Car. & P., p. 349; Rex vs. Crutchley, 7 Car. & P., p. 814; Rex vs. Sellis, 7 Car. & P., p. 850; Rex vs. Poulton, 5 Car. & P., p. 329; 2 Bishop’s Cr. Law, Secs. 541, 542.” (Code Commissioners’ Note, Pen. Code of Cal. (1st ed. 1872), p. 81.) The cited cases, of course, are among those we have discussed earlier as representing the settled common law rule that live birth is a prerequisite to a conviction of homicide.

When there is persuasive evidence of a legislative intent contrary to the views expressed in code commissioners’ notes, those views will not be followed in construing the statute. (See, e.g., People v. Valentine (1946) 28 Cal.2d 121, 138, 142-144 [169 P.2d 1] [contrary to commissioners’ note, Legislature omitted important limitation of prior statute in codifying manslaughter provision of Penal Code].) Here, however, the views of the commissioners are in full accord with the history of section 187; and as we have seen, the Legislature made no significant change in that statute when it was codified into the Penal Code. The rule is therefore applicable that “Reports of commissions which have proposed statutes that are subsequently adopted are entitled to substantial weight in construing the statutes. [Citations.] This is particularly true where the statute proposed by the commission is adopted by the Legislature without any change whatsoever and where the commission’s comment is brief, because in such a situation there is ordinarily strong reason to believe that the legislators’ votes were based in large measure upon the explanation of the commission proposing the bill.” (Van Arsdale v. Hollinger (1968) 68 Cal.2d 245, 249-250 [66 Cal.Rptr. 20, 437 P.2d 508].)15

*631It is the policy of this state to construe a penal statute as favorably to the defendant as its language and the circumstances of its application may reasonably permit; just as in the case of a question of fact, the defendant is entitled to the benefit of every reasonable doubt as to the true interpretation of words or the construction of language used in a statute. (Walsh v. Department of Alcoholic Bev. Control (1963) 59 Cal.2d 757, 764-765 [31 Cal.Rptr. 297, 382 P.2d 337], and cases cited.) We hold that in adopting the definition of murder in Penal Code section 187 the Legislature intended to exclude from its reach the act of killing an unborn fetus.

II

The People urge, however, that the sciences of obstetrics and pediatrics have greatly progressed since 1872, to the point where with proper medical care a normally developed fetus prematurely born at 28 weeks or more has an excellent chance of survival, i.e., is “viable”; that the common law requirement of live birth to prove the fetus had become a “human being” who may be the victim of murder is no longer in accord with scientific fact, since an unborn but viable fetus is now fully capable of independent life; and that one who unlawfully and maliciously terminates such a life should therefore be liable to prosecution for murder under section 187. We may grant the premises of this argument; indeed, we neither deny nor denigrate the vast progress of medicine in the century since the enactment of the Penal Code. But we cannot join in the conclusion sought to be deduced: we cannot hold this petitioner to answer for murder by reason of his alleged act of killing an unborn—even though viable—fetus. To such a charge there are two insuperable obstacles, one “jurisdictional” and the other constitutional.

Penal Code section 6 declares in relevant part that “No act or omission” accomplished after the code has taken effect “is criminal or punishable, except as prescribed or authorized by this code, or by some of the statutes which it specifies as continuing in force and as not affected by its provisions, or by some ordinance, municipal, county, or township regulation. . . .” This section embodies a fundamental principle of our tripartite form of government, i.e., that subject to the constitutional prohibition against cruel and unusual punishment, the power to define crimes and fix penalties is vested exclusively in the legislative branch. (People v. Knowles (1950) 35 Cal.2d 175, 181 [217 P.2d 1]; Harbor Comrs. v. Excelsior Redwood Co. (1891) 88 Cal. 491, 493 [26 P. 375]; People v. Hess (1951) 104 Cal.App.2d 642, 685 [234 P.2d 65]; In re Finley (1905) 1 Cal.App. 198, 201 [81 P. 1041].) Stated differently, there are no common law crimes in California. (People v. Redmond (1966) 246 *632Cal.App.2d 852, 862 [55 Cal.Rptr. 195]; People v. Harris (1961) 191 Cal.App.2d 754, 758 [12 Cal.Rptr. 916]; In re Harder (1935) 9 Cal.App.2d 153, 155 [49 P.2d 304].) “In this state the common law is of no effect so far as the specification of what acts or conduct shall constitute a crime is concerned. [Citations.] In order that a public offense be committed, some statute, ordinance or regulation prior in time to the commission of the act, must denounce it; likewise with excuses or justifications— if no statutory excuse or justification apply as to the commission of the particular offense, neither the common law nor the so-called ‘unwritten law’ may legally supply it.” (People v. Whipple (1929) 100 Cal.App. 261, 262 [279 P. 1008].)

Settled rules of construction implement this principle. Although the Penal Code commands us to construe its provisions “according to the fair import of their terms, with a view to effect its objects and to promote justice” (Pen. Code, § 4), it is clear the courts cannot go so far as to create an offense by enlarging a statute, by inserting or deleting words, or by giving the terms used false or unusual meanings. (People v. Baker (1968) 69 Cal.2d 44, 50 [69 Cal.Rptr. 595, 442 P.2d 675].) Penal statutes will not be made to reach beyond their plain intent; they include only those offenses coming clearly within the import of their language. (De Mille v. American Fed. of Radio Artists (1947) 31 Cal.2d 139', 156 [187 P.2d 769, 175 A.L.R. 382].) Indeed, “Constructive crimes— crimes built up by courts with the aid of inference, implication, and strained interpretation-—-are repugnant to the spirit and letter of English and American criminal law.” (Ex parte McNulty (1888) 77 Cal. 164, 168 [19 P. 237].)

Applying these rules to the case at bar, we would undoubtedly act in excess of the judicial power if we were to adopt the People’s proposed construction of section 187. As we have shown, the Legislature has defined the crime of murder in California to apply only to the unlawful and malicious killing of one who has been bom alive. We recognize that the killing of an unborn but viable fetus may be deemed by some to be an offense of similar nature and gravity; but as Chief Justice Marshall warned long ago, “It would be dangerous, indeed, to carry the principle, that a case which is within the reason or mischief of a statute, is within its provisions, so far as to punish a crime not enumerated in the statute, because it is of equal atrocity, or of kindred character, with those which are enumerated.” (United States v. Wiltberger (1820) 18 U.S. (5 Wheat.) 76, 96 [5 L.Ed. 37, 42].) Whether to thus extend liability for murder in California is a determination solely within the province of the Legisla*633ture.16 For a court to simply declare, by judicial fiat, that the time has now come to prosecute under section 187 one who kills an unborn but viable fetus would indeed be to rewrite the statute under the guise of construing it. Nor does a need to fill an asserted “gap” in the law between abortion and homicide—as will appear, no such gap in fact exists—justfy judicial legislation of this nature: to make it “a judicial function ‘to explore such new fields óf crime as they may appear from time to time’ is wholly foreign to the American concept of criminal justice” and “raises very serious questions concerning the principle of separation of powers.” (In re Davis (1966) 242 Cal.App.2d 645, 655-656 & fn. 12 [51 Cal.Rptr. 702].)

The second obstacle to the proposed judicial enlargement of section 187 is the guarantee of due process of law. Assuming arguendo that we have the power to adopt the new construction of this statute as the law of California, such a ruling, by constitutional command, could operate only prospectively, and thus could not in any event reach the conduct of petitioner on February 23, 1969.

The first essential of due process is fair warning of the act which is made punishable as a crime. “That the terms of a penal statute creating a new offense must be sufficiently explicit to inform those who are subject to it what conduct on their part will render them liable to its penalties, is a well-recognized requirement, consonant alike with ordinary notions of fair play and the settled rules of law.” (Connally v. General Constr. Co. (1926) 269 U.S. 385, 391 [70 L.Ed. 322, 328, 46 S.Ct. 126].) “No one *634may be required at peril of life, liberty or property to speculate as to the meaning of penal statutes. All are entitled to be informed as to what the State commands or forbids.” (Lanzetta v. New Jersey (1939) 306 U.S. 451, 453 [83 L.Ed. 888, 890, 59 S.Ct. 618].) The law of California is in full accord. (In re Newbern, supra, 53 Cal.2d 786, 792; In re Davis, supra, 242 Cal.App.2d 645, 650-651.)

This requirement of fair warning is reflected in the constitutional prohibition against the enactment of ex post facto laws (U.S. Const., art. I, §§ 9, 10; Cal. Const., art. I, § 16). When a new penal statute is applied retrospectively to make punishable an act which was not criminal at the time it was performed, the defendant has been given no advance notice consistent with due process. And precisely the same effect occurs when such an act is made punishable under a preexisting statute but by means of an unforeseeable judicial enlargement thereof. (Bouie v. City of Columbia (1964) 378 U.S. 347 [12 L.Ed.2d 894, 84 S.Ct. 1697].)

In Bouie two Negroes took seats in the restaurant section of a South Carolina drugstore; no notices were posted restricting the area to whites only. When the defendants refused to leave upon demand, they were arrested and convicted of violating a criminal trespass statute which prohibited entry on the property of another “after notice” forbidding such conduct. Prior South Carolina decisions had emphasized the necessity of proving such notice to support a conviction under the statute. The South Carolina Supreme Court nevertheless affirmed the convictions, construing the statute to prohibit not only the act of entering after notice not to do so but also the wholly different act of remaining on the property after receiving notice to leave.

The United States Supreme Court reversed the convictions, holding that the South Carolina court’s ruling was “unforeseeable” and when an “unforeseeable state-court construction of a criminal statute is applied retroactively to subject a person to criminal liability for past conduct, the effect is to deprive him of due process of law in the sense of fair warning that his contemplated conduct constitutes a crime.” (378 U.S. at pp. 354-355 [12 L.Ed.2d at p. 900].) Analogizing to the prohibition against retrospective penal legislation, the high court reasoned “Indeed, an unforeseeable judicial enlargement of a criminal statute, applied retroactively, .operates precisely like an ex post facto law, such as Art. I, § 10, of the Constitution forbids. An ex post facto law has been defined by this Court as one ‘that makes an action done before the passing of the law, and which was innocent when done, criminal; and punishes such action,’ or ‘that aggravates a crime, or makes it greater than it was, when committed.’ Calder v. Bull, 3 Dali. 386, 390,1 L.Ed. 648, 650. If a state legislature is barred by the *635Ex Post Facto Clause from passing such a law, it must follow that a State Supreme Court is barred by the Due Process Clause from achieving precisely the same result by judicial construction. Cf. Smith v. Gaboon, 283 U.S. 553, 565, 75 L.Ed. 1264, 1273, 51 S.Ct. 582. The fundamental principle that ‘the required criminal law must have existed when the conduct in issue occurred,’ Hall, General Principles of Criminal Law (2d ed. 1960), at 58-59, must apply to bar retroactive criminal prohibitions emanating from courts as well as from legislatures. If a judicial construction of a criminal statute is ‘unexpected and indefensible by reference to the law which had been expressed prior to the conduct in issue,’ it must not be given retroactive effect. Id., at 61.” (Fn. omitted.) (Id. at pp. 353-354 [12 L.Ed.2d at pp. 899-900].)

The court remarked in conclusion that “Application of this rule is particularly compelling where, as here, the petitioners’ conduct cannot be deemed improper or immoral.” (Id. at p. 362 [12 L.Ed.2d at p. 905].) In the case at bar the conduct with which petitioner is charged is certainly “improper” and “immoral,” and it is not contended he was exercising a constitutionally favored right. But the matter is simply one of degree, and it cannot be denied that the guarantee of due process extends to violent as well as peaceful men. The issue remains, would the judicial enlargement of section 187 now proposed have been foreseeable to this petitioner?

It is true that section 187, on its face, is not as “narrow and precise” as the South Carolina statute involved in Bouie; on the other hand, neither is it as vague as the statutes struck down in Connally and Lanzetta.17 Rather, section 187 bears a plain, common sense meaning, well settled in the common law and fortified by its legislative history in California. In Bouie, moreover, the court stressed that a breach of the peace statute was also in force in South Carolina at the time of the events, and that the defendants were in fact arrested on that ground and prosecuted (but not convicted) for that offense. Here, too, there was another statute on the books which petitioner could well have believed he was violating: Penal Code section 274 defines the crime of abortion, in relevant part, as the act of “Every person who . . . uses or employs any instrument or any other means whatever, with intent thereby to procure the miscarriage” of any woman, and does not come within the exceptions provided by law. The gist of the crime is the performance, with the requisite intent, of any of the acts enumerated in the statute. (People v. Root (1966) 246 Cal.App.2d 600, 604 [55 Cal.Rptr. 89]; People v. Moore (1963) 213 Cal.App.2d 160, 167 [23 Cal.Rptr. 502].) It is therefore no defense to a *636charge of violating section 274 that the act was committed unusually late in the woman’s pregnancy or by a method not commonly employed for that purpose. The prohibition is against “any means which might be used to effect a miscarriage” (People v. Clapp (1944) 67 Cal.App.2d 197, 202 [153 P.2d 758]), and has been applied to instances of beating or other physical violence inflicted upon the person of the woman for this purpose.18 In the present case, it will be remembered, petitioner’s avowed goal was not primarily to kill the fetus while it was inside his wife’s body, but rather to “stomp it out of” her; although one presumably cannot be done without the other, petitioner’s choice of words is significant and strongly implies an “intent thereby to procure the miscarriage” of his wife in violation of section 274.

Turning to the case law, we find no reported decision of the California courts which should have given petitioner notice that the killing of an unborn but viable fetus was prohibited by section 187. Indeed, the contrary clearly appears from People v. Eldridge (1906) 3 Cal.App. 648, 649 [86 P. 832], in which the defendant challenged as uncertain an information which charged him with the murder of “a human being,” to wit, the infant child “born to the said Glover H. Eldridge and said Mabel Eldridge on or about said twentieth day of February, 1905.” It was urged that “such charge might include the killing before birth, and therefore it cannot be determined from the information whether murder or abortion was intended to be charged.” The Court of Appeal rejected the contention, observing that “The only reasonable construction which can be given to the language employed in the information is to say that it charges that a child born to the defendant was by him unlawfully killed and murdered. That it was born is clearly stated; that it could be killed after birth of necessity implies that it was born alive, and we think the charge of murder was set forth with the degree of certainty required.” (Italics added.)

Properly understood, the often cited case of People v. Chavez (1947) 77 Cal.App.2d 621 [176 P.2d 92], does not derogate from this rule. There the defendant was charged with the murder of her newborn child, and convicted of manslaughter. She testified that the baby dropped from her womb into the toilet bowl; that she picked it up two or three minutes later, and cut but did not tie the umbilical cord; that the baby was limp and made no cry; and that after 15 minutes she wrapped it in a newspaper and concealed it, where it was found dead the next day. The autopsy surgeon testified that the baby was a full-term, nine-month child, weighing six and one-half pounds and appearing normal in every respect; that the *637body had very little blood in it, indicating the child had bled to death through the untied umbilical cord; that such a process would have taken about an hour; and that in his opinion “the child was born alive, based on conditions he found and the fact that the lungs contained air and the blood was extravasated or pushed back into the tissues, indicating heart action.” (Id. at p. 623.)

On appeal, the defendant emphasized that a doctor called by the defense had suggested other tests which the autopsy surgeon could have performed to determine the matter of live birth; on this basis, it was contended that the question of whether the infant was born alive “rests entirely on pure speculation.” (Id. at p. 624.) The Court of Appeal found only an insignificant conflict in that regard (id. at p. 627), and focussed its attention instead on testimony of the autopsy surgeon admitting the possibility that the evidence of heart and lung action could have resulted from the child’s breathing “after presentation of the head but before the birth was completed” (id. at p. 624).

The court cited the mid-19th century English infanticide cases mentioned hereinabove, and noted that the decisions had not reached uniformity on whether breathing, heart action, severance of the umbilical cord, or some combination of these or other factors established the status of “human being” for the purposes of the law of homicide. (Id. at pp. 624-625.) The court then adverted to the state of modern medical knowledge, discussed the phenomenon of viability, and held that “a viable child in the process of being born is a human being within the meeting of the homicide statutes, whether or not the process has been fully completed. It should at least be considered a human being where it is a living baby and where in the natural course of events a birth which is already started would naturally be successfully completed.” (Italics added.) (Id. at p. 626.) Since the testimony of the autopsy surgeon left no doubt in that case that a live birth had at least begun, the court found “the evidence is sufficient here to support the implied finding of the jury that this child was born alive and became a human being within the meaning of the homicide statutes.” (Italics added.) (Id. at p. 627.)19

Chavez thus stands for the proposition—to which we adhere— that a viable fetus “in the process of being born” is a human being within the meaning of the homicide statutes. But it stands for no more; in particular it does not hold that a fetus, however viable, which is not “in the process of being born” is nevertheless a “human being” in the law of homicide. On the contrary, the opinion is replete with references to the *638common law requirement that the child be “born alive,” however that term as defined, and must accordingly be deemed to reaffirm that requirement as part of the law of California.20

The Chavez court relied in part on Scott v. McPheeters (1939) 33 Cal.App.2d 629 [92 P.2d 678, 93 P.2d 562], a decision holding that an unborn child is an “existing person,” within the meaning of Civil Code section 29, for the purpose of bringing a postnatal action for prenatal injuries. In People v. Belous, supra, 71 Cal.2d 954, however, a majority of this court distinguished such civil law rules on the ground they either “require a live birth or reflect the interest of the parents.” (Id. at p. 968 & fn. 12.) We need not repeat that analysis here; but two further bases of distinction deserve mention. First, Scott emphasized that the child’s right' of action for prenatal injuries was unknown to the common law and would not exist in California but for statutory authorization.21 By the same token, as we have seen, the fetus’ status as a “human being” within the definition of murder was unknown to the common law and exists only where special feticide statutes have been enacted. Secondly, the law’s protection of the property interests of an unborn child dates not from Scott but from a far earlier time: for example, in Blackstone’s day it was already well settled that “An infant in ventre sa mere, or in the mother’s womb, is supposed in law to be born for many purposes. It is capable of having a guardian assigned to it; and it is enabled to have an estate limited to its use, and to take afterwards by such limitation, as if it were then *639actually born.” (1 Blackstone, Commentaries *130 (1765).) Inasmuch as such rules coexisted for centuries with the common law requirement of live birth to support a conviction of homicide, they cannot reasonably be deemed to have given petitioner notice that the killing of an unborn but viable fetus would now be murder.

Finally, although a defendant is not bound to know the decisional law of other states, the United States Supreme Court in Bouie (378 U.S. at pp. 360-361 [12 L.Ed.2d at pp. 903-904]) referred to reported cases of jurisdictions other than South Carolina in concluding that the South Carolina Supreme Court’s construction of the statute “is no less inconsistent with the law of other States than it is with the prior case law of South Carolina and, of course, with the language of the statute itself.” Here, too, the cases decided in our sister states from Chavez to the present are unanimous in requiring proof that the child was born alive before a charge of homicide can be sustained. (Bennett v. State (Wyo. 1963) 377 P.2d 634, 635-637; People v. Ryan (1956) 91 I11.2d 467 [138 N.E.2d 516, 518-520]; People v. Hayner (1949) 300 N.Y. 171 [90 N.E.2d 23, 24]; Singleton v. State (1948) 33 Ala.App. 536 [35 So.2d 375, 378]; Montgomery v. State (1947) 202 Ga. 678 [44 S.E.2d 242, 243-244]; cf. Watson v. State (1955) 208 Md. 210 [117 A.2d 549, 552].) And the text writers of the same period are no less unanimous on the point. (Perkins on Criminal Law, supra, pp. 29-30; Clark & Marshall, Crimes (6th ed. 1958) § 10.00, pp. 534-536; 1 Wharton, Criminal Law and Procedure (Anderson ed. 1957) § 189; 2 Burdick, Law of Crime (1946) § 445; 40 Am.Jur.2d, Homicide, §§ 9, 434; 40 C.J.S., Homicide, § 2b.)

We conclude that the judicial enlargement of section 187 now urged upon us by the People would not have been forseeable to this petitioner, and hence that its adoption at this time would deny him due process of law.

Let a peremptory writ of prohibition issue restraining respondent court from taking any further proceedings on Count I of the information, charging petitioner with the crime of murder.

McComb, J., Peters, J., Tobriner, J., and Peek, J.,* concurred.

BURKE, Acting C. J.

The majority hold that “Baby Girl” Vogt, who, according to medical testimony, had reached the 35th week of development, *640had a 96 percent chance of survival, and was “definitely” alive and viable at the time of her death, nevertheless was not a “human being” under California’s homicide statutes. In my view, in so holding, the majority ignore significant common law precedents, frustrate the express intent of the Legislature, and defy reason, logic and common sense.

Penal Code section 187 defines murder as “the unlawful killing of a human being, with malice aforethought.” Penal Code section 192 defines manslaughter as “the unlawful killing of a human being, without malice.” The majority pursue the meaning of the term “human being” down the ancient hallways of the common law, citing Coke, Blackstone and Hale to the effect that the slaying of a “quickened” (i.e. stirring in the womb) child constituted “a great misprision,” but not murder. Although, as discussed below, I strongly disagree with the premise that the words of our penal statutes must be construed as of 1648 or 1765, nevertheless, there is much common law precedent which would support the view that a viable fetus such as Baby Girl Vogt is a human being under those statutes.

The majority cast a passing glance at the common law concept of quickening, but fail to explain the significance of that concept: At common law, the quickened fetus was considered to be a human being, a second life separate and apart from its mother. As stated by Blackstone, in the passage immediately preceding that portion quoted in the majority opinion (fn. 6), “Life is the immediate gift of God, a right inherent by nature in every individual; and it begins in contemplation of law as soon as an infant is able to stir in the mother’s womb.” (Italics added; 1 Blackstone, Commentaries, p. 129; see Rex v. Anonymous (1811) 3 Campb. 73, 170 Eng.Reprint 1310, 1311-1312; State v. Cooper, 22 N.J.L. 52, 54-55.)

Modern scholars have confirmed this aspect of common law jurisprudence. As Means observes, “The common law itself prohibited abortion after quickening and hanging a pregnant felon after quickening, because the life of a second human being would thereby be taken, although it did not call the offense murder or manslaughter.” (Italics added; Means, The Law of New York Concerning Abortion and the Status of the Foetus, 1664-1968: A Case of Cessation of Constitutionality (1968) 14 N.Y.L.F. 411, 504.)

This reasoning explains why the killing of a quickened child was considered “a great misprision,” although the killing of an unquickened child was no crime at all at common law (Means, supra, at p. 420). Moreover, although the common law did not apply the labels of “murder” or “manslaughter” to the killing of a quickened fetus, it appears that at common law this “great misprision” was severely punished. As late as 1837, the wilful *641aborting of a woman quick with child was punishable by death in England. (Lord Landsdowne’s Act of 1828 (9 Geo. IV, c. 31; Lord Ellenborough’s Act of 1803 (43 Geo. Ill, c. 58) j1 see Means, supra, at p. 440, fn. 64.)

Thus, at common law, the killing of a quickened child was severely punished, since that child was considered to be a human being. The majority would have us assume that the Legislature in 1850 and 1872 simply overlooked this “great misprision” in codifying and classifying criminal offenses in California, or reduced that offense to the lesser offense of illegal abortion with its relatively lenient penalties (Pen. Code, § 274).2

In my view, we cannot assume that the Legislature intended a person such as defendant, charged with the malicious slaying of a fully viable child, to suffer only the mild penalties imposed upon common abortionists who, ordinarily, procure only the miscarriage of a nonviable fetus or embryo. (See Comment, Model Penal Code, § 207.11, p. 149 (Tent. Draft No. 9, 1959).) To do so would completely ignore the important common law distinction between the quickened and unquickened child.

Of course, I do not suggest that we should interpret the term “human being” in our homicide statutes in terms of the common law concept of quickening. At one time, that concept had a value in differentiating, as accurately as was then scientifically possible, between life and nonlife. The analogous concept of viability is clearly more satisfactory, for it has a well defined and medically determinable meaning denoting the ability of the fetus to live or survive apart from its mother.3

The majority opinion suggests that we are confined to common law concepts, and to the common law definition of murder or manslaughter. However, the Legislature, in Penal Code sections 187 and 192, has defined those offenses for us: homicide is the unlawful killing of a “human being.” Those words need not be frozen in place as of any particular time, but must be fairly and reasonably interpreted by this court to promote justice and to carry out the evident purposes of the Legislature in adopting a homicide statute. Thus, Penal Code section 4, which was enacted in 1872 along with sections 187 and 192, provides: “The rule of the common law, that penal statutes are to be strictly construed, has no application to this code. *642All its provisions are to be construed according to the fair import of their terms, with a view to effect its objects and to promote justice.” (Accord, In re Cregler, 56 Cal.2d 308, 312 [14 Cal.Rptr. 289, 363 P.2d 305]; People v. Valentine, 28 Cal.2d 121, 142 [169 P.2d 1]; In re Haines, 195 Cal. 605, 614, 621 [234 P. 883].)

As the majority opinion recognizes, “ ‘In this state the common law is of no effect so far as the specification of what acts or conduct shall constitute a crime is concerned.’ ” (Ante, p. 632, quoting from People v. Whipple, 100 Cal.App. 261, 262 [279 P. 1008].) Instead, we must construe penal statutes in accordance with the “fair import” of their terms, rather than restrict those statutes to common law principles. As stated in Katz v. Walkinshaw, 141 Cal. 116, 122-123 [70 P. 663, 74 P. 766], “The idea that the doctrine ... is a part of the common law adopted by our statute, and beyond the power of the court to change or modify, is founded upon a misconception of the extent to which the common law is adopted by such statutory provisions, and a failure to observe some of the rules and principles of the common law itself. . . . The true doctrine is, that the common law by its own principles adapts itself to varying conditions, and modifies its own rules so as to serve the ends of justice under the different circumstances, a principle adopted into our code by section 3510 of the Civil Code: ‘When the reason of a rule ceases, so should the rule itself.’ ”

Penal Code section 4, which abolishes the common law principle of the strict construction of penal statutes, embodies the doctrine of Katz v. Walkinshaw, supra, 141 Cal. 116, and permits this court fairly to construe the terms of those statutes to serve the ends of justice. Consequently, nothing should prevent this court from holding that Baby Girl Vogt was a human (“belonging or relating to man; characteristic of man”).4 being (“existence, as opp. to nonexistence; specif, life”)5 under California’s homicide statutes.

We commonly conceive of human existence as a spectrum stretching from birth to death. However, if this court properly might expand the definition of “human being” at one end of that spectrum, we may do so at the other end. Consider the following example: All would agree that “Shooting or otherwise damaging a corpse is not homicide . . . .” (Perkins, Criminal Law (2d ed. 1969) ch. 2, § 1, p. 31.) In other words, a corpse is not considered to be a “human being” and thus cannot be the subject of a “killing” as those terms are used in homicide statutes. However, it is readily apparent that our concepts of what constitutes a “corpse” *643have been and are being continually modified by advances in the field of medicine, including new techniques for life revival, restoration and resuscitation such as artificial respiration, open heart massage, transfusions, transplants and a variety of life-restoring stimulants, drugs and new surgical methods. Would this court ignore these developments and exonerate the killer of an apparently “drowned” child merely because that child would have been pronounced dead in 1648 or 1850? Obviously not. Whether a homicide occurred in that case would be determined by medical testimony regarding the capability of the child to have survived prior to the defendant’s act. And that is precisely the test which this court should adopt in the instant case.

The common law reluctance to characterize the killing of a quickened fetus as a homicide was based solely upon a presumption that the fetus would have been born dead. (People v. Chavez, supra, 77 Cal.App.2d 621, 626; Atknson, Life, Birth and Livebirth, 20 L.Q.Rev. 134.) This presumption seems to have persisted in this country at least as late as 1876. (See State v. Winthrop, 43 Iowa 519, 523.) Based upon the state of the medical art in the 17th, 18th and 19th centuries, that presumption may have been well-founded. However, as we approach the 21st century, it has become apparent that “This presumption is not only contrary to common experience and the ordinary course of nature, but it is contrary to the usual rule with respect to presumptions followed in this state.” (People v. Chavez, supra, at p. 626; see Civ. Code,.§ 3546 [formerly Code Civ. Proc., § 1963, subd. 28]; cf. Health & Saf. Code, § 25953, forbidding therapeutic abortion after the 20th week of pregnancy.)

There are no accurate statistics disclosing fetal death rates in “common law England,” although the foregoing presumption of death indicates a significantly high death experience. On the other hand, in California the fetal death rate6 in 1968 is estimated to be 12 deaths in 1,000, a ratio which would have given Baby Girl Vogt a 98.8 percent chance of survival. (Cal. Statistical Abstract (1969) Table E-3, p. 65.) If, as I have contended, the term “human being” in our homicide statutes is a fluid concept to be defined in accordance with present conditions, then there can be no question that the term should include the fully viable fetus.

The majority suggest that to do so would improperly create some new offense. However, the offense of murder is no new offense. Contrary to the majority opinion, the Legislature has not “defined the crime of murder in California to apply only to the unlawful and malicious killing of one who has been born alive.” {Ante, p. 632.) Instead, the Legislature simply used the broad term “human being” and directed the courts to construe *644that term according to its “fair import” with a view to effect the objects of the homicide statutes and promote justice. (Pen. Code, § 4.) What justice will be promoted, what objects effectuated, by construing “human being” as excluding Baby Girl Vogt and her unfortunate successors? Was defendant’s brutal act of stomping her to death any less an act of homicide than the murder of a newly born baby? No one doubts that the term “human being” would include the elderly or dying persons whose potential for life has nearly lapsed; their proximity to death is deemed immaterial. There is no sound reason for denying the viable fetus, with its unbounded potential for life, the same status.

The majority also suggest that such an interpretation of our homicide statutes would deny defendant “fair warning” that his act was punishable as a crime. (Ante, p. 634.) Aside from the absurdity of the underlying premise that defendant consulted Coke, Blackstone or Hale before kicking Baby Girl Vogt to death, it is clear that defendant had adequate notice that his act could constitute homicide. Due process only precludes prosecution under a new statute insufficiently explicit regarding the specific conduct proscribed, or under a preexisting statute “by means of an unforeseeable judicial enlargement thereof.” (Ante, p. 634.)

Our homicide statutes have been in effect in this state since 1850. The fact that the California courts have not been called upon to determine the precise question before us does not render “unforeseeable” a decision which determines that a viable fetus is a “human being” under those statutes. Can defendant really claim suprise that a 5-pound, 18-inch, 34-week-old, living, viable child is considered to be a human being?

The fact is that the foregoing construction of our homicide statutes easily could have been anticipated from strong dicta in People v. Chavez, supra, 77 Cal.App.2d 621, 625-626 (hg. den. by S.Ct.), wherein the court reviewed common law precedents but disapproved their requirement that the child be born alive and completely separated from its mother. The court in Chavez held that a viable child killed during, but prior to completion of, the birth process, was a human being under the homicide statutes. However, the court did not hold that partial birth was a prerequisite, for the court expressly set forth its holding “Without drawing a line of distinction applicable to all cases . . . .” (77 Cal.App.2d at p. 627.) In dicta, the court discussed the question when an unborn infant becomes a human being under the homicide statutes, as follows: “There is not much change in the child itself between a moment before and a moment after its expulsion from the body of its mother, and normally, while still *645dependent upon its mother, the child for some time before it is born, has not only the possibility but a strong probability of an ability to live an independent life. . . . While before birth or removal it is in a sense dependent upon its mother for life, there is another sense in which it has started an independent existence after it has reached a state of development where it is capable of living and where it will, in the normal course of nature and with ordinary care, continue to live and grow as a separate being. While it may not be possible to draw an exact line applicable to all cases, the rules of law should recognize and make some attempt to follow the natural and scientific facts to which they relate. . . . [I]t would be a mere fiction to hold that a child is not a human being because the process of birth has not been fully completed, when it has reached that state of viability when the destruction of the life of its mother would not end its existence and when, if separated from the mother naturally or by artificial means, it will live and grow in the normal manner.” (77 Cal. App.2d at pp. 625-626.)

Thus the Chavez case explodes the majority’s premise that a viability test for defining “human being” under our homicide statutes was unforeseeable; Chavez approved and advocated this interpretation 23 years ago. (See also Scott v. McPheeters, 33 Cal.App.2d 629, 635 [92 P.2d 678, 93 P.2d 562] [“Who may say that such a viable child is not in fact a human being in actual existence?”].) I would conclude that defendant had sufficient notice that the words “human being” could include a viable fetus. As stated in People v. Victor, 62 Cal.2d 280, 299 [42 Cal.Rptr. 199, 398 P.2d 391], “Admittedly the word [‘imminent’] is to some extent a relative one; but ‘the law is full of instances where a man’s fate depends on his estimating rightly, that is, as the jury subsequently estimates it, some matter of degree.’ [Citation.]”

In summary, I have shown that at common law, the slaying of a quickened fetus was a “great misprision” and was severely punished, since that fetus was considered to be a human being. We should not presume that the Legislature ignored these common law developments and intended to punish the malicious killing of a viable fetus as the lesser offense of illegal abortion. Moreover, apart from the common law approach, our Legislature has expressly directed us to construe the homicide statutes in accordance with the fair import of their terms. There is no good reason why a fully viable fetus should not be considered a “human being” under those statutes. To so construe them would not create any new offense, and would not deny defendant fair warning or due process since the Chavez case anticipated that construction long ago.

*646The trial court’s denial of defendant’s motion to set aside the information was proper, and the peremptory writ of prohibition should be denied.

Sullivan, J., concurred.

The petition of the real party in interest for a rehearing was denied September 10, 1970. Burke, J., and Sullivan, J., were of the opinion that the petition should be granted.

Mr. Justice Stewart

delivered the opinion of the Court.

A California statute makes it a criminal offense for a person to “be addicted to the use of narcotics.” 1 This *661appeal draws into question the constitutionality of that provision of the state law, as construed by the California courts in the present case.

The appellant was convicted after a jury trial in the Municipal Court of Los Angeles. The evidence against him was given by two Los Angeles police officers. Officer Brown testified that he had had occasion to examine the appellant’s arms one evening on a street in Los Angeles some four months before the trial.2 The officer testified that at that time he had observed “scar tissue and discoloration on the inside” of the appellant’s right arm, and “what appeared to be numerous needle marks and a scab which was approximately three inches below the crook of the elbow” on the appellant’s left arm. The officer also testified that the appellant under questioning had admitted to the occasional use of narcotics.

Officer Lindquist testified that he had examined the appellant the following morning in the Central Jail in Los Angeles. The officer stated that at that time he had observed discolorations and scabs on the appellant’s arms, *662and he identified photographs which had been taken of the appellant’s arms shortly after his arrest the night before. Based upon more than ten years of experience as a member of the Narcotic Division of the Los Angeles Police Department, the witness gave his opinion that “these marks and the discoloration were the result of the injection of hypodermic needles into the tissue into the vein that was not sterile.” He stated that the scabs were several days old at the time of his examination, and that the appellant was neither under the influence of narcotics nor suffering withdrawal symptoms at the time he saw him. This witness also testified that the appellant had admitted using narcotics in the past.

The appellant testified in his own behalf, denying the alleged conversations with the police officers and denying that he had ever used narcotics or been addicted to their use. He explained the marks on his arms as resulting from an allergic condition contracted during his military service. His testimony was corroborated by two witnesses.

The trial judge instructed the jury that the statute made it a misdemeanor for a person “either to use narcotics, or to be addicted to the use of narcotics ... .3 That portion of the statute referring to the ‘use’ of narcotics is based upon the ‘act’ of using. That portion of the statute referring to ‘addicted to the use’ of narcotics is based upon a condition or status. They are not identical. ... To be addicted to the use of narcotics is said to be a status or condition and not an act. It is a continuing offense and differs from most other offenses in the fact that [it] is *663chronic rather than acute; that it continues after it is complete and subjects the offender to arrest at any time before he reforms. The existence of such a chronic condition may be ascertained from a single examination, if the characteristic reactions of that condition be found present.”

The judge further instructed the jury that the appellant could be convicted under a general verdict if the jury agreed either that he was of the “status” or had committed the “act” denounced by the statute.4 “All that the People must show is either that the defendant did use a narcotic in Los Angeles County, or that while in the City of Los Angeles he was addicted to the use of narcotics . ...”5

Under these instructions the jury returned a verdict finding the appellant “guilty of the offense charged.” *664An appeal was taken to the Appellate Department of the Los Angeles County Superior Court, “the highest court of a State in which a decision could be had” in this case. 28 U. S. C. § 1257. See Smith v. California, 361 U. S. 147, 149; Edwards v. California, 314 U. S. 160, 171. Although expressing some doubt as to the constitutionality of “the crime of being a narcotic addict,” the reviewing court in an unreported opinion affirmed the judgment of conviction, citing two of its own previous unreported decisions which had upheld the constitutionality of the statute.6 We noted probable jurisdiction of this appeal, 368 U. S. 918, because it squarely presents the issue whether the statute as construed by the California courts in this case is repugnant to the Fourteenth Amendment of the Constitution.

The broad power of a State to regulate the narcotic drugs traffic within its borders is not here in issue. More than forty years ago, in Whipple v. Martinson, 256 U. S. 41, this Court explicitly recognized the validity of that power: “There can be no question of the authority of the State in the exercise of its police power to regulate the administration, sale, prescription and use of dangerous and habit-forming drugs .... The right to exercise this power is so manifest in the interest of the public health and welfare, that it is Unnecessary to enter upon a discussion of it beyond saying that it is too firmly established to be successfully called in question.” 256 U. S., at 45.

Such regulation, it can be assumed, could take a variety of valid forms. A State might impose criminal sanctions, for example, against the unauthorized manufacture, prescription, sale, purchase, or possession of narcotics within its borders. In the interest of discouraging the viola*665tion of such laws, or in the interest of the general health or welfare of its inhabitants, a State might establish a program of compulsory treatment for those addicted to narcotics.7 Such a program of treatment might require periods of involuntary confinement. And penal sanctions might be imposed for failure to comply with established compulsory treatment procedures. Cf. Jacobson v. Massachusetts, 197 U. S. 11. Or a State might choose to attack the evils of narcotics traffic on broader fronts also — through public health education, for example, or by efforts to ameliorate the economic and social conditions under which those evils might be thought to flourish. In short, the range of valid choice which a State might make in this area is undoubtedly a wide one, and the wisdom of any particular choice within the allowable spectrum is not for us to decide. Upon that premise we.turn to the California law in issue here.

It would be possible to construe the statute under which the appellant was convicted as one which is operative only upon proof of the actual use of narcotics within the State’s jurisdiction. But the California courts have not so construed this law. Although there was evidence in the present case that the appellant had used narcotics in Los Angeles, the jury were instructed that they could convict him even if they disbelieved that evidence. The appellant could be convicted, they were told, if they found simply that the appellant’s “status” or “chronic condition” was that of being “addicted to the use of narcotics.” And it is impossible to know from the jury’s verdict that the defendant was not convicted upon precisely such a finding.

*666The instructions of the trial court, implicitly approved on appeal, amounted to “a ruling on a question of state law that is as binding on us as though the precise words had been written” into the statute. Terminiello v. Chicago, 337 U. S. 1, 4. “We can only take the statute as the state courts read it.” Id., at 6. Indeed, in their brief in this Court counsel for the State have emphasized that it is “the proof of addiction by circumstantial evidence ... by the tell-tale track of needle marks and scabs over the veins of his arms, that remains the gist of the section.”

This statute, therefore,, is not one which punishes a person for the use of narcotics, for their purchase, sale or possession, or for antisocial or disorderly behavior resulting from their administration. It is not a law which even purports to provide or require medical treatment. Rather, we deal with a statute which makes the “status” of narcotic addiction a criminal offense, for which the offender may be prosecuted “at any time before he reforms.” California has said that a person can be continuously guilty of this offense, whether or not he has ever used or possessed any narcotics within the State, and whether or not he has been guilty of any antisocial behavior there.

It is unlikely that any State at this moment in history would attempt to make it a criminal offense for a person to be mentally ill, or a leper, or to be afflicted with a venereal disease. A State might determine that the general health and welfare require that the victims of these and other human afflictions be dealt with by compulsory treatment, involving quarantine, confinement, or sequestration. But, in the light of contemporary human knowledge, a law which made a criminal offense of such a disease would doubtless be universally thought to be an infliction of cruel and unusual punishment in violation of the Eighth and Fourteenth Amendments. See Francis v. Resweher, 329 U. S. 459.

*667We cannot but consider the statute before us as of the same category. In this Court counsel for the State recognized that narcotic addiction is an illness.8 Indeed, it is apparently an illness which may be contracted innocently or involuntarily.9 We hold that a state law which imprisons a person thus afflicted as a criminal, even though he has never touched any narcotic drug within the State or been guilty of any irregular behavior there, inflicts a cruel and unusual punishment in violation of the Fourteenth Amendment. To be sure, imprisonment for ninety days is not, in the abstract, a punishment which is either cruel or unusual. But the question cannot be considered in the abstract. Even one day in prison would be a cruel and unusual punishment for the “crime” of having a common cold.

We are not unmindful that the vicious evils of the narcotics traffic have occasioned the grave concern of government. There are, as we have said, countless fronts on *668which those evils may be legitimately attacked. We deal in this case only with an individual provision of a particularized local law as it has so far been interpreted by the California courts.

Reversed.

Mr. Justice Frankfurter took no part in the consideration or decision of this case.

Mr. Justice Douglas,

concurring.

While I join the Court’s opinion, I wish to make more explicit the reasons why I think it is “cruel and unusual” punishment in the sense of the Eighth Amendment to treat as a criminal a person who is a drug addict.

In Sixteenth Century England one prescription for insanity was to beat the subject “until he had regained his reason.” Deutsch, The Mentally Ill in America (1937), p. 13. In America “the violently insane went to the whipping post and into prison dungeons or, as sometimes happened, were burned at the stake or hanged”; and “the pauper insane often roamed the countryside as wild men and from time to time were pilloried, whipped, and jailed.” Action for Mental Health (1961), p. 26.

As stated by Dr. Isaac Ray many years ago:

“Nothing can more strongly illustrate the popular ignorance respecting insanity than the proposition, equally objectionable in its humanity and its logic, that the insane should be punished for criminal acts, in order to deter other insane persons from doing the same thing.” Treatise on the Medical Jurisprudence of Insanity (6th ed. 1871), p. 66.

Today we have our differences over the legal definition of insanity. But however insanity is defined, it is in end effect treated as a disease. While afflicted people *669may be confined either for treatment or for the protection of society, they are not branded as criminals.

Yet terror and punishment linger on as means of dealing with some diseases. As recently stated:

“. . . the idea of basing treatment for disease on purgatorial acts and ordeals is an ancient one in medicine. It may trace back to the Old Testament belief that disease of any kind, whether mental or physical, represented punishment for sin; and thus relief could take the form of a final heroic act of atonement. This superstition appears to have given support to fallacious medical rationales for such procedures as purging, bleeding, induced vomiting, and blistering, as well as an entire chamber of horrors constituting the early treatment of mental illness. The latter included a wide assortment of shock techniques, such as the ‘water cures’ (dousing, ducking, and near-drowning), spinning in a chair, centrifugal swinging, and an early form of electric shock. All, it would appear, were planned as means of driving from the body some evil spirit or toxic vapor.” Action for Mental Health (1961), pp. 27-28.

That approach continues as respects drug addicts. Drug addiction is more prevalent in this country than in any other nation of the western world.1 S. Rep. No. 1440, 84th Cong., 2d Sess., p. 2. It is sometimes referred to as “a contagious disease.” Id., at p. 3. But those living in a world of black and white put the addict in the cate*670gory of those who could, if they would, forsake their evil ways.

The first step toward addiction may be as innocent as a boy’s puff on a cigarette in an alleyway. It may come from medical prescriptions. Addiction may even be present at birth. Earl Ubell recently wrote:

“In Bellevue Hospital’s nurseries, Dr. Saul Krug-man, head of pediatrics, has been discovering babies minutes old who are heroin addicts.

“More than 100 such infants have turned up in the last two years, and they show all the signs of drug withdrawal: irritability, jitters, loss of appetite, vomiting, diarrhea, sometimes convulsions and death.

“ 'Of course, they get the drug while in the womb from their mothers who are addicts,’ Dr. Krugman said yesterday when the situation came to light. ‘We control the symptoms with Thorazine, a tranquilizing drug.

“ ‘You should see some of these children. They have a high-pitched cry. They appear hungry but they won’t eat when offered food. They move around so much in the crib that their noses and toes become red and excoriated.’

“Dr. Lewis Thomas, professor of medicine at New York University-Bellevue, brought up the problem of the babies Monday night at a symposium on narcotics addiction sponsored by the New York County Medical Society. He saw in the way the babies respond to treatment a clue to the low rate of cure of addiction.

“ ‘Unlike the adult addict who gets over his symptoms of withdrawal in a matter of days, in most cases,’ Dr. Thomas explained later, ‘the infant has to be treated for weeks and months. The baby continues to show physical signs of the action of the drugs.

*671“ ‘Perhaps in adults the drugs continue to have physical effects for a much longer time after withdrawal than we have been accustomed to recognize. That would mean that these people have a physical need for the drug for a long period, and this may be the clue to recidivism much more than the social or psychological pressures we’ve been talking about.’ ” N. Y. Herald Tribune, Apr. 25, 1962, p. 25, cols. 3-4.

The addict is under compulsions not capable of management without outside help. As stated by the Council on Mental Health:

“Physical dependence is defined as the development of an altered physiological state which is brought about by the repeated administration of the drug and which necessitates continued administration of the drug to prevent the appearance of the characteristic illness which is termed an abstinence syndrome. When an addict says that he has a habit, he means that he is physically dependent on a drug. When he says that one drug is habit-forming and another is not, he means that the first drug is one on which physical dependence can be developed and that the second is a drug on which physical dependence cannot be developed. Physical dependence is a real physiological disturbance. It is associated with the development of hyperexcitability in reflexes mediated through multineurone arcs. It can be induced in animals, it has been shown to occur in the paralyzed hind limbs of addicted chronic spinal dogs, and also has been produced in dogs whose cerebral cortex has been removed.” Report on Narcotic Addiction, 165 A. M. A. J. 1707, 1713.

Some say the addict has a disease. See Hesse, Narcotics and Drug Addiction (1946), p. 40 et seq.

*672Others say addiction is not a disease but “a symptom of a mental or psychiatric disorder.” H. R. Rep. No. 2388, 84th Cong., 2d Sess., p. 8. And see Present Status of Narcotic Addiction, 138 A. M. A. J. 1019, 1026; Narcotic Addiction, Report to Attorney General Brown by Citizens Advisory Committee to the Attorney General on Crime Prevention (1954), p. 12; Finestone, Narcotics and Criminality, 22 Law & Con temp. Prob. 69, 83-85 (1957).

The extreme symptoms of addiction have been described as follows:

"To be a confirmed drug addict is to be one of the walking dead .... The teeth have rotted out; the appetite is lost and the stomach and intestines don’t function properly. The gall bladder becomes inflamed; eyes and skin turn a billious yellow. In some cases membranes of the nose turn a flaming red; the partition separating the nostrils is eaten away — breathing is difficult. Oxygen in the blood decreases; bronchitis and tuberculosis develop. Good traits of character disappear and bad ones emerge. Sex organs become affected. Veins collapse and livid purplish scars remain. Boils and abscesses plague the skin; gnawing pain racks the body. Nerves snap; vicious twitching develops. Imaginary and fantastic fears blight the mind and sometimes complete insanity results. Often times, too, death comes — much too early in life .... Such is the torment of being a drug addict; such is the plague of being one of the walking dead.” N. Y. L. J., June 8, 1960, p. 4, col. 2.

Some States punish addiction, though most do not. See S. Doc. No. 120, 84th Cong., 2d Sess., pp. 41, 42. Nor does the Uniform Narcotic Drug Act, first approved in 1932 and now in effect in most of the States. Great Britain, beginning in 1920 placed “addiction and the *673treatment of addicts squarely and exclusively into the hands of the medical profession.” Lindesmith, The British System of Narcotics Control, 22 Law & Contemp. Prob. 138 (1957). In England the doctor “has almost complete professional autonomy in reaching decisions about the treatment of addicts.” Schur, British Narcotics Policies, 51 J. Crim. L. & Criminology 619, 621 (1961). Under British law “addicts are patients, not criminals.” Ibid. Addicts have not disappeared in England but they have decreased in number (id., at 622) and there is now little “addict-crime” there. Id., at 623.

The fact that England treats the addict as a sick person, while a few of our States, including California, treat him as a criminal, does not, of course, establish the unconstitutionality of California’s penal law. But we do know that there is “a hard core” of “chronic and incurable drug addicts who, in reality, have lost their power of self-control.” S. Rep. No. 2033, 84th Cong., 2d Sess., p. 8. There has been a controversy over the type of treatment— whether enforced hospitalization or ambulatory care is better. H. R. Rep. No. 2388, 84th Cong., 2d Sess., pp. 66-68. But there is little disagreement with the statement of Charles Winick: “The hold of drugs on persons addicted to them is so great that it would be almost appropriate to reverse the old adage and say that opium derivatives represent the religion of the people who use them.” Narcotics Addiction and its Treatment, 22 Law & Contemp. Prob. 9 (1957). The abstinence symptoms and their treatment are well known. Id., at 10-11. Cure is difficult because of the complex of forces that make for addiction. Id., at 18-23. “After the withdrawal period, vocational activities, recreation, and some kind of psychotherapy have a major role in the treatment program, which ideally lasts from four to six months.” Id., at 23-24. Dr. Marie Nyswander tells us that normally a drug addict *674must be hospitalized in order to be cured. The Drug Addict as a Patient (1966), p. 138.

The impact that an addict has on a community causes alarm and often leads to punitive measures. Those measures are justified when they relate to acts of transgression. But I do not see how under our system being an addict can be punished as a crime. If addicts can be punished for their addiction, then the insane can also be punished for their insanity. Each has a disease and each must be treated as a sick person.2 As Charles Winick has said:

"There can be no single program for the elimination of an illness as complex as drug addiction, which *675carries so much emotional freight in the community. Cooperative interdisciplinary research and action, more local community participation, training the various healing professions in the techniques of dealing with addicts, regional treatment facilities, demonstration centers, and a thorough and vigorous post-treatment rehabilitation program would certainly appear to be among the minimum requirements for any attempt to come to terms with this problem. The addict should be viewed as a sick person, with a chronic disease which requires almost emergency action.” 22 Law & Contemp. Prob. 9, 33 (1957).

The Council on Mental Health reports that criminal sentences for addicts interferes “with the possible treatment and rehabilitation of addicts and therefore should be abolished.” 165 A. M. A. J. 1968, 1972.

The command of the Eighth Amendment, banning “cruel and unusual punishments,” stems from the Bill of Rights of 1688. See Francis v. Resweber, 329 U. S. 459, 463. And it is applicable to the States by reason of the Due Process Clause of the Fourteenth Amendment. Ibid.

The historic punishments that were cruel and unusual included “burning at the stake, crucifixion, breaking on the wheel” (In re Kemmler, 136 U. S. 436, 446), quartering, the rack and thumbscrew (see Chambers v. Florida, 309 U. S. 227, 237), and in some circumstances even solitary confinement (see Medley, 134 U. S. 160, 167-168). *676The question presented in the earlier cases concerned the degree of severity with which a particular offense was punished or the element of cruelty present.3 A punishment out of all proportion to the offense may bring it within the ban against “cruel and unusual punishments.” See O’Neil v. Vermont, 144 U. S. 323, 331. So may the cruelty of the method of punishment, as, for example, disemboweling a person alive. See Wilkerson v. Utah, 99 U. S. 130, 135. But the principle that would deny power to exact capital punishment for a petty crime would also deny power to punish a person by fine or imprisonment for being sick.

The Eighth Amendment expresses the revulsion of civilized man against barbarous acts — the “cry of horror” against man’s inhumanity to his fellow man. See O’Neil v. Vermont, supra, at 340 (dissenting opinion); Francis v. Resweber, supra, at 473 (dissenting opinion).

By the time of Coke, enlightenment was coming as respects the insane. Coke said that the execution of a madman “should be a miserable spectacle, both against law, and of extreame inhumanity and cruelty, and can be no example to others.” 6 Coke’s Third Inst. (4th ed. 1797), p. 6. Blackstone endorsed this view of Coke. 4 Commentaries (Lewis ed. 1897), p. 25.

We should show the same discernment respecting drug addiction. The addict is a sick person. He may, of course, be confined for treatment or for the protection of society.4 Cruel and unusual punishment results not from confinement, but from convicting the addict of a crime. The purpose of § 11721 is not to cure, but to penalize. *677Were the purpose to cure, there would be no need for a mandatory jail term of not less than 90 days. Contrary to my Brother Clark, I think the means must stand constitutional scrutiny, as well as the end to be achieved. A prosecution for addiction, with its resulting stigma and irreparable damage to the good name of the accused, cannot be justified as a means of protecting society, where a civil commitment would do as well. Indeed, in § 5350 of the Welfare and Institutions Code, California has expressly provided for civil proceedings for the commitment of habitual addicts. Section 11721 is, in reality, a direct attempt to punish those the State cannot commit civilly.5 This prosecution has no relationship to the curing *678of an illness. Indeed, it cannot, for the prosecution is aimed at penalizing an illness, rather than at providing medical care for it. We would forget the teachings of the Eighth Amendment if we allowed sickness to be made a crime and permitted sick people to be punished for being sick. This age of enlightenment cannot tolerate such barbarous action.

Mr. Justice Harlan,

concurring.

I am not prepared to hold that on the present state of medical knowledge it is completely irrational and hence unconstitutional for a State to conclude that narcotics addiction is something other than an illness nor that it amounts to cruel and unusual punishment for the State to subject narcotics addicts to its criminal law. Insofar as addiction may be identified with the use or possession of narcotics within the State (or, I would suppose, without the State), in violation of local statutes prohibiting such acts, it may surely be reached by the State’s criminal law. But in this case the trial court’s instructions permitted the jury to find the appellant guilty on no more proof than that he was present in California while he was addicted to narcotics.* Since addiction alone cannot *679reasonably be thought to amount to more than a compelling propensity to use narcotics, the effect of this instruction was to authorize criminal punishment for a bare desire to commit a criminal act.

If the California statute reaches this type of conduct, and for present purposes we must accept the trial court’s construction as binding, Terminiello v. Chicago, 337 U. S. 1, 4, it is an arbitrary imposition which exceeds the power that a State may exercise in enacting its criminal law. Accordingly, I agree that the application of the California statute was unconstitutional in this case and join the judgment of reversal.

Mr. Justice Clark,

dissenting.

The Court finds § 11721 of California’s Health and Safety Code, making it an offense to “be addicted to the use of narcotics,” violative of due process as “a cruel and unusual punishment.” I cannot agree.

The statute must first be placed in perspective. California has a comprehensive and enlightened program for the control of narcotism based on the overriding policy of prevention and cure. It is the product of an extensive investigation made in the mid-Fifties by a committee of distinguished scientists, doctors, law enforcement officers and laymen appointed by the then Attorney General, now Governor, of California. The committee filed a detailed study entitled “Report on Narcotic Addiction” which was given considerable attention. No recommendation was made therein for the repeal of § 11721, and the State Legislature in its discretion continued the policy of that section.

Apart from prohibiting specific acts such as the purchase, possession and sale of narcotics, California has taken certain legislative steps in regard to the status of being a narcotic addict — a condition commonly recognized as a threat to the State and to the individual. The *680Code deals with this problem in realistic stages. At its incipiency narcotic addiction is handled under § 11721 of the Health and Safety Code which is at issue here. It provides that a person found to be addicted to the use of narcotics shall serve a term in the county jail of not less than 90 days nor more than one year, with the minimum 90-day confinement applying in all cases without exception. Provision is made for parole with periodic tests to detect readdiction.

The trial court defined “addicted to narcotics” as used in § 11721 in the following charge to the jury:

“The word ‘addicted' means, strongly disposed to some taste or practice or habituated, especially to drugs. In order to inquire as to whether a person is addicted to the use of narcotics is in effect an inquiry as to his habit in that regard. Does he use them habitually. To use them often or daily is, according to the ordinary acceptance of those words, to use them habitually.”

There was no suggestion that the term “narcotic addict” as here used included a person who acted without volition or who had lost the power of self-control. Although the section is penal in appearance — perhaps a carry-over from a less sophisticated approach — its present provisions are quite similar to those for civil commitment and treatment of addicts who have lost the power of self-control, and its present purpose is reflected in a statement which closely follows § 11721: “The rehabilitation of narcotic addicts and the prevention of continued addiction to narcotics is a matter of statewide concern.” California Health and Safety Code § 11728.

Where narcotic addiction has progressed beyond the incipient, volitional stage, California provides for commitment of three months to two years in a state hospital. *681California Welfare and Institutions Code § 5355. For the purposes of this provision, a narcotic addict is defined as

“any person who habitually takes or otherwise uses to the extent of having lost the -power of self-control any opium, morphine, cocaine, or other narcotic drug as defined in Article 1 of Chapter 1 of Division 10 of the Health and Safety Code.” California Welfare and Institutions Code § 5350. (Emphasis supplied.)

This proceeding is clearly civil in nature with a purpose of rehabilitation and cure. Significantly, if it is found that a person committed under § 5355 will not receive substantial benefit from further hospital treatment and is not dangerous to society, he may be discharged — but only after a minimum confinement of three months. § 5355.1.

Thus, the “criminal” provision applies to the incipient narcotic addict who retains self-control, requiring confinement of three months to one year and parole with frequent tests to detect renewed use of drugs. Its overriding purpose is to cure the less seriously addicted person by preventing further use. On the other hand, the “civil” commitment provision deals with addicts who have lost the power of self-control, requiring hospitalization up to two years. Each deals with a different type of addict but with a common purpose. This is most apparent when the sections overlap: if after civil commitment of an addict it is found that hospital treatment will not be helpful, the addict is confined for a minimum period of three months in the same manner as is the volitional addict under the “criminal” provision.

In the instant case the proceedings against the petitioner were brought under the volitional-addict section. There was testimony that he had been using drugs only four months with three to four relatively mild doses a *682week. At arrest and trial he appeared normal. His testimony was clear and concise, being simply that he had never used drugs. The scabs and pocks on his arms and body were caused, he said, by “overseas shots” administered during army service preparatory to foreign assignment. He was very articulate in his testimony but the jury did not believe him, apparently because he had told the clinical expert while being examined after arrest that he had been using drugs, as I have stated above. The officer who arrested him also testified to like statements and to scabs — some 10 or 15 days old — showing narcotic injections. There was no evidence in the record of withdrawal symptoms. Obviously he could not have been committed under § 5355 as one who had completely “lost the power of self-control.” The jury was instructed that narcotic “addiction” as used in § 11721 meant strongly disposed to a taste or practice or habit of its use, indicated by the use of narcotics often or daily. A general verdict was returned against petitioner, and he was ordered confined for 90 days to be followed by a two-year parole during which he was required to take periodic Nalline tests.

The majority strikes down the conviction primarily on the grounds that petitioner was denied due process by the imposition of criminal penalties for nothing more than being in a status. This viewpoint is premised upon the theme that § 11721 is a “criminal” provision authorizing a punishment, for the majority admits that “a State might establish a program of compulsory treatment for those addicted to narcotics” which “might require periods of involuntary confinement.” I submit that California has done exactly that. The majority’s error is in instructing the California Legislature that hospitalization is the only treatment for narcotics addiction — that anything less is a punishment denying due process. California has found otherwise after a study which I suggest was more extensive than that conducted by the Court. *683Even in California’s program for hospital commitment of nonvolitional narcotic addicts — which the majority approves — it is recognized that some addicts will not respond to or do not need hospital treatment. As to these persons its provisions are identical to those of § 11721— confinement for a period of not less than 90 days. Section 11721 provides this confinement as treatment for the volitional addicts to whom its provisions apply, in addition to parole with frequent tests to detect and prevent further use of drugs. The fact that § 11721 might be labeled “criminal” seems irrelevant,* not only to the majority’s own “treatment” test but to thp “concept of ordered liberty” to which the States must attain under the Fourteenth Amendment. The test is the overall purpose and effect of a State’s act, and I submit that California’s program relative to narcotic addicts — including both the “criminal” and “civil” provisions — is inherently one of treatment and lies well within the power of a State.

However, the case in support of the judgment below need not rest solely on this reading of California law. For even if the overall statutory scheme is ignored and a purpose and effect of punishment is attached to § 11721, that provision still does not violate the Fourteenth Amendment. The majority acknowledges, as it must, that a State can punish persons who purchase, possess or use narcotics. Although none of these acts are harmful to society in themselves, the State constitutionally may attempt to deter and prevent them through punishment because of the grave threat of future harmful conduct which they pose. Narcotics addiction — including the incipient, volitional, addiction to which this provision speaks — is no different. California courts have taken judicial notice that “the inordinate use of a narcotic drug tends *684to create an irresistible craving and forms a habit for its continued use until one becomes an addict, and he respects no convention or obligation and will lie, steal, or use any other base means to gratify his passion for the drug, being lost to all considerations of duty or social position.” People v. Jaurequi, 142 Cal. App. 2d 555, 561, 298 P. 2d 896, 900 (1956). Can this Court deny the legislative and judicial judgment of California that incipient, volitional narcotic addiction poses a threat of serious crime similar to the threat inherent in the purchase or possession of narcotics? And if such a threat is inherent in addiction, can this Court say that California is powerless to deter it by punishment?

It is no answer to suggest that we are dealing with an involuntary status and thus penal sanctions will be ineffective and unfair. The section at issue applies only to persons who use narcotics often or even daily but not to the point of losing self-control. When dealing with involuntary addicts California moves only through § 5355 of its Welfare Institutions Code which clearly is not penal. Even if it could be argued that § 11721 may not be limited to volitional addicts, the petitioner in the instant case undeniably retained the power of self-control and thus to him the statute would be constitutional. Moreover, “status” offenses have long been known and recognized in the criminal law. 4 Blackstone, Commentaries (Jones ed. 1916), 170. A ready example is drunkenness, which plainly is as involuntary after addiction to alcohol as is the taking of drugs.

Nor is the conjecture relevant that petitioner may have acquired his habit under lawful circumstances. There was no suggestion by him to this effect at trial, and surely the State need not rebut all possible lawful sources of addiction as part of its prima facie case.

The argument that the statute constitutes a cruel and unusual punishment is governed by the discussion above. *685Properly construed, the statute provides a treatment rather than a punishment. But even if interpreted as penal, the sanction of incarceration for 3 to 12 months is not unreasonable when applied to a person who has voluntarily placed himself in a condition posing a serious threat to the State. Under either theory, its provisions for 3 to 12 months’ confinement can hardly be deemed unreasonable when compared to the provisions for 3 to 24 months’ confinement under § 5355 which the majority approves.

I would affirm the judgment.

Mr. Justice White,

dissenting.

If appellant’s conviction rested upon sheer status, condition or illness or if he was convicted for being an addict who had lost his power of self-control, I would have other thoughts about' this case. But this record presents neither situation. And I believe the Court has departed from its wise rule of not deciding constitutional questions except where necessary and from its equally sound practice of construing state statutes, where possible, in a manner saving their constitutionality.1

*6861 am not at all ready to place the use of narcotics beyond the reach of the States’ criminal laws. I do not consider appellant’s conviction to be a punishment for having an illness or for simply being in some status or condition, but rather a conviction for the regular, repeated or habitual use of narcotics immediately prior to his arrest and in violation of the California law. As defined by the trial court,2 addiction is the regular use of narcotics and can be proved only by evidence of such use. To find addiction in this case the jury had to believe that appellant had frequently used narcotics in the recent past.3 California is entitled to have its statute and the record so read, particularly where the State’s only purpose in allowing prosecutions for addiction was to supersede its own venue requirements applicable to prosecutions for the use of narcotics and in effect to allow convictions for use *687where there is no precise evidence of the county where the use took place.4

Nor do I find any indications in this record that California would apply § 11721 to the case of the helpless addict. I agree with my Brother Clark that there was no evidence at all that appellant had lost the power to control his acts. There was no evidence of any use within 3 days prior to appellant’s arrest. The most recent marks might have been 3 days old or they might have been 10 *688days old. The appellant admitted before trial that he had last used narcotics 8 days before his arrest. At the trial he denied having taken narcotics at all. The uncon-troverted evidence was that appellant was not under the influence of narcotics at the time of his arrest nor did he have withdrawal symptoms. He was an incipient addict, a redeemable user, and the State chose to send him to jail for 90 days rather than to attempt to confine him by civil proceedings under another statute which requires a finding that the addict has lost the power of self-control. In my opinion, on this record, it was within the power of the State of California to confine him by criminal proceed- ■ ings for the use of narcotics or for regular use amounting to habitual use.5

The Court clearly does not rest its decision upon the narrow ground that the jury was not expressly instructed not to convict if it believed appellant’s use of narcotics was beyond his control. The Court recognizes no degrees of addiction. The Fourteenth Amendment is today held to bar any prosecution for addiction regardless of the degree or frequency of use, and the Court’s opinion bristles with indications of further consequences. If it is “cruel and unusual punishment” to convict appellant for addiction, it is difficult to understand why it would be any less offensive to the Fourteenth Amendment to convict him for use on the same evidence of use which proved he was an addict. It is significant that in purporting to reaffirm the power of the States to deal with the narcotics traffic, the Court does not include among the obvious powers of the State the power to punish for the use of narcotics. I cannot think that the omission was inadvertent.

*689The Court has not merely tidied up California’s law by removing some irritating vestige of an outmoded approach to the control of narcotics. At the very least, it has effectively removed California’s power to deal effectively with the recurring case under the statute where there is ample evidence of use but no evidence of the precise location of use. Beyond this it has cast serious doubt upon the power of any State to forbid the use of narcotics under threat of criminal punishment. I cannot believe that the Court would forbid the application of the criminal laws to the use of narcotics under any circumstances. But the States, as well as the Federal Government, are now on notice. They will have to await a final answer in another case.

Finally, I deem this application of “cruel and unusual punishment” so novel that I suspect the Court was hard put to find a way to ascribe to the Framers of the Constitution the result reached today rather than to its own notions of ordered liberty. If this case involved economic regulation, the present Court’s allergy to substantive due process would surely save the statute and prevent the Court from imposing its own philosophical predilections upon state legislatures or Congress. I fail to see why the Court deems it more appropriate to write into the Constitution its own abstract notions of how best to handle the narcotics problem, for it obviously cannot match either the States or Congress in expert understanding..

I respectfully dissent.

*14Justice O’Connor

announced the judgment of the Court and delivered an opinion, in which The Chief Justice and Justice Kennedy join.

In this case, we decide whether the Eighth Amendment prohibits the State of California from sentencing a repeat felon to a prison term of 25 years to life under the State’s “Three Strikes and You’re Out” law.

r — <

A

California’s three strikes law reflects a shift in the State’s sentencing policies toward incapacitating and deterring repeat offenders who threaten the public safety. The law was designed “to ensure longer prison sentences and greater punishment for those who commit a felony and have been previously convicted of serious and/or violent felony offenses.” Cal. Penal Code Ann. § 667(b) (West 1999). On March 3, 1993, California Assemblymen Bill Jones and Jim Costa introduced Assembly Bill 971, the legislative version of what would later become the three strikes law. The Assembly Committee on Public Safety defeated the bill only weeks later. Public outrage over the defeat sparked a voter initiative to add Proposition 184, based loosely on the bill, to the ballot in the November 1994 general election.

On October 1, 1993, while Proposition 184 was circulating, 12-year-old Polly Klaas was kidnaped from her home in Petaluma, California. Her admitted killer, Richard Allen Davis, had a long criminal history that included two prior kidnaping convictions. Davis had served only half of his *15most recent sentence (16 years for kidnaping, assault, and burglary). Had Davis served his entire sentence, he would still have been in prison on the day that Polly Klaas was kidnaped.

Polly Klaas’ murder galvanized support for the three strikes initiative. Within days, Proposition 184 was on its way to becoming the fastest qualifying initiative in California history. On January 3, 1994, the sponsors of Assembly Bill 971 resubmitted an amended version of the bill that conformed to Proposition 184. On January 31,1994, Assembly Bill 971 passed the Assembly by a 63 to 9 margin. The Senate passed it by a 29 to 7 margin on March 3, 1994. Governor Pete Wilson signed the bill into law on March 7, 1994. California voters approved Proposition 184 by a margin of 72 to 28 percent on November 8, 1994.

California thus became the second State to enact a three strikes law. In November 1993, the voters of Washington State approved their own three strikes law, Initiative 593, by a margin of 3 to 1. U. S. Dept. of Justice, National Institute of Justice, J. Clark, J. Austin, & D. Henry, “Three Strikes and You’re Out”: A Review of State Legislation 1 (Sept. 1997) (hereinafter Review of State Legislation). Between 1993 and 1995, 24 States and the Federal Government enacted three strikes laws. Ibid. Though the three strikes laws vary from State to State, they share a common goal of protecting the public safety by providing lengthy prison terms for habitual felons.

B

California’s current three strikes law consists of two virtually identical statutory schemes “designed to increase the prison terms of repeat felons.” People v. Superior Court of San Diego Cty. ex rel. Romero, 13 Cal. 4th 497, 504, 917 P. 2d 628, 630 (1996) (Romero). When a defendant is convicted of a felony, and he has previously been convicted of one or more prior felonies defined as “serious” or “violent” in Cal. Penal Code Ann. §§ 667.5 and 1192.7 (West Supp. 2002), sentencing *16is conducted pursuant to the three strikes law. Prior convictions must be alleged in the charging document, and the defendant has a right to a jury determination that the prosecution has proved the prior convictions beyond a reasonable doubt. § 1025; § 1158 (West 1985).

If the defendant has one prior “serious” or “violent” felony conviction, he must be sentenced to “twice the term otherwise provided as punishment for the current felony conviction.” § 667(e)(1) (West 1999); § 1170.12(c)(1) (West Supp. 2002). If the defendant has two or more prior “serious” or “violent” felony convictions, he must receive “an indeterminate term of life imprisonment.” § 667(e)(2)(A) (West 1999); § 1170.12(c)(2)(A) (West Supp. 2002). Defendants sentenced to life under the three strikes law become eligible for parole on a date calculated by reference to a “minimum term,” which is the greater of (a) three times the term otherwise provided for the current conviction, (b) 25 years, or (c) the term determined by the court pursuant to §1170 for the underlying conviction, including any enhancements. §§667(e)(2)(A)(i)-(iii) (West 1999); §§ 1170.12(c)(2)(A)(i)-(iii) (West Supp. 2002).

Under California law, certain offenses may be classified as either felonies or misdemeanors. These crimes are known as “wobblers.” Some crimes that would otherwise be misdemeanors become “wobblers” because of the defendant’s prior record. For example, petty theft, a misdemeanor, becomes a “wobbler” when the defendant has previously served a prison term for committing specified theft-related crimes. §490 (West 1999); §666 (West Supp. 2002). Other crimes, such as grand theft, are “wobblers” regardless of the defendant’s prior record. See § 489(b) (West 1999). Both types of “wobblers” are triggering offenses under the three strikes law only when they are treated as felonies. Under California law, a “wobbler” is presumptively a felony and “remains a felony except when the discretion is actually exercised” to make the crime a misdemeanor. People v. Wil*17liams, 27 Cal. 2d 220, 229, 163 P. 2d 692, 696 (1945) (emphasis deleted and internal quotation marks omitted).

In California, prosecutors may exercise their discretion to charge a “wobbler” as either a felony or a misdemeanor. Likewise, California trial courts have discretion to reduce a “wobbler” charged as a felony to a misdemeanor either before preliminary examination or at sentencing to avoid imposing a three strikes sentence. Cal. Penal Code Ann. §§ 17(b)(5), 17(b)(1) (West 1999); People v. Superior Court of Los Angeles Cty. ex rel. Alvarez, 14 Cal. 4th 968, 978, 928 P. 2d 1171, 1177-1178 (1997). In exercising this discretion, the court may consider “those factors that direct similar sentencing decisions,” such as “the nature and circumstances of the offense, the defendant’s appreciation of and attitude toward the offense,... [and] the general objectives of sentencing.” Ibid, (internal quotation marks and citations omitted).

California trial courts can also vacate allegations of prior “serious” or “violent” felony convictions, either on motion by the prosecution or sua sponte. Romero, supra, at 529-530, 917 P. 2d, at 647-648. In ruling whether to vacate allegations of prior felony convictions, courts consider whether, “in light of the nature and circumstances of [the defendant’s] present felonies and prior serious and/or violent felony convictions, and the particulars of his background, character, and prospects, the defendant may be deemed outside the [three strikes’] scheme’s spirit, in whole or in part.” People v. Williams, 17 Cal. 4th 148, 161, 948 P. 2d 429, 437 (1998). Thus, trial courts may avoid imposing a three strikes sentence in two ways: first, by reducing “wobblers” to misdemeanors (which do not qualify as triggering offenses), and second, by vacating allegations of prior “serious” or “violent” felony convictions.

C

On parole from a 9-year prison term, petitioner Gary Ewing walked into the pro shop of the El Segundo Golf *18Course in Los Angeles County on March 12, 2000. He walked out with three golf clubs, priced at $399 apiece, concealed in his pants leg. A shop employee, whose suspicions were aroused when he observed Ewing limp out of the pro shop, telephoned the police. The police apprehended Ewing in the parking lot.

Ewing is no stranger to the criminal justice system. In 1984, at the age of 22, he pleaded guilty to theft. The court sentenced him to six months in jail (suspended), three years’ probation, and a $300 fine. In 1988, he was convicted of felony grand theft auto and sentenced to one year in jail and three years’ probation. After Ewing completed probation, however, the sentencing court reduced the crime to a misdemeanor, permitted Ewing to withdraw his guilty plea, and dismissed the case. In 1990, he was convicted of petty theft with a prior and sentenced to 60 days in the county jail and three years’ probation. In 1992, Ewing was convicted of battery and sentenced to 30 days in the county jail and two years’ summary probation. One month later, he was convicted of theft and sentenced to 10 days in the county jail and 12 months’ probation. In January 1993, Ewing was convicted of burglary and sentenced to 60 days in the county jail and one year’s summary probation. In February 1993, he was convicted of possessing drug paraphernalia and sentenced to six months in the county jail and three years’ probation. In July 1993, he was convicted of appropriating lost property and sentenced to 10 days in the county jail and two years’ summary probation. In September 1993, he was convicted of unlawfully possessing a firearm and trespassing and sentenced to 30 days in the county jail and one year’s probation.

In October and November 1993, Ewing committed three burglaries and one robbery at a Long Beach, California, apartment complex over a 5-week period. He awakened one of his victims, asleep on her living room sofa, as he tried to disconnect her video cassette recorder from the television in *19that room. When she screamed, Ewing ran out the front door. On another occasion, Ewing accosted a victim in the mailroom of the apartment complex. Ewing claimed to have a gun and ordered the victim to hand over his wallet. When the victim resisted, Ewing produced a knife and forced the victim back to the apartment itself. While Ewing rifled through the bedroom, the victim fled the apartment screaming for help. Ewing absconded with the victim’s money and credit cards.

On December 9,1993, Ewing was arrested on the premises of the apartment complex for trespassing and lying to a police officer. The knife used in the robbery and a glass cocaine pipe were later found in the back seat of the patrol car used to transport Ewing to the police station. A jury convicted Ewing of first-degree robbery and three counts of residential burglary. Sentenced to nine years and eight months in prison, Ewing was paroled in 1999.

Only 10 months later, Ewing stole the golf clubs at issue in this case. He was charged with, and ultimately convicted of, one count of felony grand theft of personal property in excess of $400. See Cal. Penal Code Ann. § 484 (West Supp. 2002); §489 (West 1999). As required by the three strikes law, the prosecutor formally alleged, and the trial court later found, that Ewing had been convicted previously of four serious or violent felonies for the three burglaries and the robbery in the Long Beach apartment complex. See § 667(g) (West 1999); § 1170.12(e) (West Supp. 2002).

At the sentencing hearing, Ewing asked the court to reduce the conviction for grand theft, a “wobbler” under California law, to a misdemeanor so as to avoid a three strikes sentence. See §§ 17(b), 667(d)(1) (West 1999); § 1170.12(b)(1) (West Supp. 2002). Ewing also asked the trial court to exercise its discretion to dismiss the allegations of some or all of his prior serious or violent felony convictions, again for purposes of avoiding a three strikes sentence. See Romero, 13 Cal. 4th, at 529-531, 917 P. 2d, at 647-648. Before sen-*20fencing Ewing, the trial court took note of his entire criminal history, including the fact that he was on parole when he committed his latest offense. The court also heard arguments from defense counsel and a plea from Ewing himself.

In the end, the trial judge determined that the grand theft should remain a felony. The court also ruled that the four prior strikes for the three burglaries and the robbery in Long Beach should stand. As a newly convicted felon with two or more “serious” or “violent” felony convictions in his past, Ewing was sentenced under the three strikes law to 25 years to life.

The California Court of Appeal affirmed in an unpublished opinion. No. B143745 (Apr. 25, 2001). Relying on our decision in Rummel v. Estelle, 445 U. S. 263 (1980), the court rejected Ewing’s claim that his sentence was grossly disproportionate under the Eighth Amendment. Enhanced sentences under recidivist statutes like the three strikes law, the court reasoned, serve the “legitimate goal” of deterring and incapacitating repeat offenders. The Supreme Court of California denied Ewing’s petition for review, and we granted certiorari, 535 U. S. 969 (2002). We now affirm.

II

A

The Eighth Amendment, which forbids cruel and unusual punishments, contains a “narrow proportionality principle” that “applies to noncapital sentences.” Harmelin v. Michigan, 501 U. S. 957, 996-997 (1991) (Kennedy, J., concurring in part and concurring in judgment); cf. Weems v. United States, 217 U. S. 349, 371 (1910); Robinson v. California, 370 U. S. 660, 667 (1962) (applying the Eighth Amendment to the States via the Fourteenth Amendment). We have most recently addressed the proportionality principle as applied to terms of years in a series of cases beginning with Rummel v. Estelle, supra.

*21In Rummel, we held that it did not violate the Eighth Amendment for a State to sentence a three-time offender to life in prison with the possibility of parole. Id., at 284-285. Like Ewing, Rummel was sentenced to a lengthy prison term under a recidivism statute. Rummel’s two prior offenses were a 1964 felony for “fraudulent use of a credit card to obtain $80 worth of goods or services,” and a 1969 felony conviction for “passing a forged check in the amount of $28.36.” Id., at 265. His triggering offense was a conviction for felony theft — “obtaining $120.75 by false pretenses.” Id., at 266.

This Court ruled that “[hjaving twice imprisoned him for felonies, Texas was entitled to place upon Rummel the onus of one who is simply unable to bring his conduct within the social norms prescribed by the criminal law of the State.” Id., at 284. The recidivism statute “is nothing more than a societal decision that when such a person commits yet another felony, he should be subjected to the admittedly serious penalty of incarceration for life, subject only to the State’s judgment as to whether to grant him parole.” Id., at 278. We noted that this Court “has on occasion stated that the Eighth Amendment prohibits imposition of a sentence that is grossly disproportionate to the severity of the crime.” Id., at 271. But “[ojutside the context of capital punishment, successful challenges to the proportionality of particular sentences have been exceedingly rare.” Id., at 272. Although we stated that the proportionality principle “would . . . come into play in the extreme example ... if a legislature made overtime parking a felony punishable by life imprisonment,” id., at 274, n. 11, we held that “the mandatory life sentence imposed upon this petitioner does not constitute cruel and unusual punishment under the Eighth and Fourteenth Amendments,” id., at 285.

In Hutto v. Davis, 454 U. S. 370 (1982) (per curiam), the defendant was sentenced to two consecutive terms of 20 years in prison for possession with intent to distribute nine *22ounces of marijuana and distribution of marijuana. We held that such a sentence was constitutional: “In short, Rummel stands for the proposition that federal courts should be reluctant to review legislatively mandated terms of imprisonment, and that successful challenges to the proportionality of particular sentences should be exceedingly rare.” Id., at 374 (citations and internal quotation marks omitted).

Three years after Rummel, in Solem v. Helm, 463 U. S. 277, 279 (1983), we held that the Eighth Amendment prohibited “a life sentence without possibility of parole for a seventh nonviolent felony.” The triggering offense in Solem was “uttering a ‘no account’ check for $100.” Id., at 281. We specifically stated that the Eighth Amendment’s ban on cruel and unusual punishments “prohibits .. . sentences that are disproportionate to the crime committed,” and that the “constitutional principle of proportionality has been recognized explicitly in this Court for almost a century.” Id., at 284, 286. The Solem Court then explained that three factors may be relevant to a determination of whether a sentence is so disproportionate that it violates the Eighth Amendment: “(i) the gravity of the offense and the harshness of the penalty; (ii) the sentences imposed on other criminals in the same jurisdiction; and (iii) the sentences imposed for commission of the same crime in other jurisdictions.” Id., at 292.

Applying these factors in Solem, we struck down the defendant’s sentence of life without parole. We specifically noted the contrast between that sentence and the sentence in Rummel, pursuant to which the defendant was eligible for parole. 463 U. S., at 297; see also id., at 300 (“[T]he South Dakota commutation system is fundamentally different from the parole system that was before us in Rummel”). Indeed, we explicitly declined to overrule Rummel: “[Q]ur conclusion today is not inconsistent with Rummel v. Estelle.” 463 U. S., at 303, n. 32; see also id., at 288, n. 13 (“[O]ur decision *23is entirely consistent with this Court’s prior cases — including Rummel v. Estelle”).

Eight years after Solem, we grappled with the proportionality issue again in Harmelin. Harmelin was not a recidivism case, but rather involved a first-time offender convicted of possessing 672 grams of cocaine. He was sentenced to life in prison without possibility of parole. A majority of the Court rejected Harmelin’s claim that his sentence was so grossly disproportionate that it violated the Eighth Amendment. The Court, however, could not agree on why his proportionality argument failed. Justice Scalia, joined by The Chief Justice, wrote that the proportionality principle was “an aspect of our death penalty jurisprudence, rather than a generalizable aspect of Eighth Amendment law.” 501 U. S. at 994. He would thus have declined to apply gross disproportionality principles except in reviewing capital sentences. Ibid.

Justice Kennedy,

joined by two other Members of the Court, concurred in part and concurred in the judgment. Justice Kennedy specifically recognized that “[t]he Eighth Amendment proportionality principle also applies to noncap-ital sentences.” Id., at 997. He then identified four principles of proportionality review — “the primacy of the legislature, the variety of legitimate penological schemes, the nature of our federal system, and the requirement that proportionality review be guided by objective factors” — that “inform the final one: The Eighth Amendment does not require strict proportionality between crime and sentence. Rather, it forbids only extreme sentences that are ‘grossly disproportionate’ to the crime.” Id., at 1001 (citing Solem, supra, at 288). Justice Kennedy’s concurrence also stated that Solem “did not mandate” comparative analysis “within and between jurisdictions.” 501 U. S., at 1004-1005.

The proportionality principles in our cases distilled in Justice Kennedy’s concurrence guide our application of the *24Eighth Amendment in the new context that we are called upon to consider.

B

For many years, most States have had laws providing for enhanced sentencing of repeat offenders. See, e. g., U. S. Dept. of Justice, Bureau of Justice Assistance, National Assessment of Structured Sentencing (1996). Yet between 1993 and 1995, three strikes laws effected a sea change in criminal sentencing throughout the Nation.1 These laws responded to widespread public concerns about crime by targeting the class of offenders who pose the greatest threat to public safety: career criminals. As one of the chief architects of California’s three strikes law has explained: “Three Strikes was intended to go beyond simply making sentences tougher. It was intended to be a focused effort to create a sentencing policy that would use the judicial system to reduce serious and violent crime.” Ardaiz, California’s Three Strikes Law: History, Expectations, Consequences, 32 McGeorge L. Rev. 1, 12 (2000) (hereinafter Ardaiz).

Throughout the States, legislatures enacting three strikes laws made a deliberate policy choice that individuals who have repeatedly engaged in serious or violent criminal behavior, and whose conduct has not been deterred by more conventional approaches to punishment, must be isolated from society in order to protect the public safety. Though three strikes laws may be relatively new, our tradition of deferring to state legislatures in making and implementing such important policy decisions is longstanding. Weems, 217 U. S., at 379; Gore v. United States, 357 U. S. 386, 393 *25(1958); Payne v. Tennessee, 501 U. S. 808, 824 (1991); Rummel, 445 U. S., at 274; Solem, 463 U. S., at 290; Harmelin, 501 U. S., at 998 (Kennedy, J., concurring in part and concurring in judgment).

Our traditional deference to legislative policy choices finds a corollary in the principle that the Constitution “does not mandate adoption of any one penological theory.” Id., at 999 (Kennedy, J., concurring in part and concurring in judgment). A sentence can have a variety of justifications, such as incapacitation, deterrence, retribution, or rehabilitation. See 1 W. LaFave & A. Scott, Substantive Criminal Law § 1.5, pp. 30-36 (1986) (explaining theories of punishment). Some or all of these justifications may play a role in a State’s sentencing scheme. Selecting the sentencing rationales is generally a policy choice to be made by state legislatures, not federal courts.

When the California Legislature enacted the three strikes law, it made a judgment that protecting the public safety requires incapacitating criminals who have already been convicted of at least one serious or violent crime. Nothing in the Eighth Amendment prohibits California from making that choice. To the contrary, our cases establish that “States have a valid interest in deterring and segregating habitual criminals.” Parke v. Raley, 506 U. S. 20, 27 (1992); Oyler v. Boles, 368 U. S. 448, 451 (1962) (“[T]he constitutionality of the practice of inflicting severer criminal penalties . upon habitual offenders is no longer open to serious challenge”). Recidivism has long been recognized as a legitimate basis for increased punishment. See Almendarez-Torres v. United States, 523 U. S. 224, 230 (1998) (recidivism “is as typical a sentencing factor as one might imagine”); Witte v. United States, 515 U. S. 389, 400 (1995) (“In repeatedly upholding such recidivism statutes, we have rejected double jeopardy challenges because the enhanced punishment imposed for the later offense ... [is] ‘a stiffened penalty for the latest crime, which is considered to be an aggravated *26offense because a repetitive one’ ” (quoting Gryger v. Burke, 334 U. S. 728, 732 (1948))).

California’s justification is no pretext. Recidivism is a serious public safety concern in California and throughout the Nation. According to a recent report, approximately 67 percent of former inmates released from state prisons were charged with at least one “serious” new crime within three years of their release. See U. S. Dept. of Justice, Bureau of Justice Statistics, P. Langan & D. Levin, Special Report: Recidivism of Prisoners Released in 1994, p. 1 (June 2002). In particular, released property offenders like Ewing had higher recidivism rates than those released after committing violent, drug, or public-order offenses. Id., at 8. Approximately 73 percent of the property offenders released in 1994 were arrested again within three years, compared to approximately 61 percent of the violent offenders, 62 percent of the public-order offenders, and 66 percent of the drug offenders. Ibid.

In 1996, when the Sacramento Bee studied 233 three strikes offenders in California, it found that they had an aggregate of 1,165 prior felony convictions, an average of 5 apiece. See Furillo, Three Strikes — The Verdict: Most Offenders Have Long Criminal Histories, Sacramento Bee, Mar. 31, 1996, p. A1. The prior convictions included 322 robberies and 262 burglaries. Ibid. About 84 percent of the 233 three strikes offenders had been convicted of at least one violent crime. Ibid. In all, they were responsible for 17 homicides, 7 attempted slayings, and 91 sexual assaults and child molestations. Ibid. The Sacramento Bee concluded, based on its investigation, that “[i]n the vast majority of the cases, regardless of the third strike, the [three strikes] law is snaring [the] long-term habitual offenders with multiple felony convictions . . ..” Ibid.

The State’s interest in deterring crime also lends some support to the three strikes law. We have long viewed both incapacitation and deterrence as rationales for recidivism *27statutes: “[A] recidivist statute[’s] . . . primary goals are to deter repeat offenders and, at some point in the life of one who repeatedly commits criminal offenses serious enough to be punished as felonies, to segregate that person from the rest of society for an extended period of time.” Rummel, supra, at 284. Four years after the passage of California’s three strikes law, the recidivism rate of parolees returned to prison for the commission of a new crime dropped by nearly 25 percent. California Dept. of Justice, Office of the Attorney General, “Three Strikes and You’re Out” — Its Impact on the California Criminal Justice System After Four Years, p. 10 (1998). Even more dramatically:

“An unintended but positive consequence of ‘Three Strikes’ has been the impact on parolees leaving the state. More California parolees are now leaving the state than parolees from other jurisdictions entering California. This striking turnaround started in 1994. It was the first time more parolees left the state than entered since 1976. This trend has continued and in 1997 more than 1,000 net parolees left California.” Ibid.

See also Janiskee & Erler, Crime, Punishment, and Romero: An Analysis of the Case Against California’s Three Strikes Law, 39 Duquesne L. Rev. 43, 45-46 (2000) (“Prosecutors in Los Angeles routinely report that ‘felons tell them they are moving out of the state because they fear getting a second or third strike for a nonviolent offense’ ” (quoting Sanchez, A Movement Builds Against “Three Strikes” Law, Washington Post, Feb. 18, 2000, p. A3)).

To be sure, California’s three strikes law has sparked controversy. Critics have doubted the law’s wisdom, cost-efficiency, and effectiveness in reaching its goals. See, e. g., Zimring, Hawkins, & Kamin, Punishment and Democracy: Three Strikes and You’re Out in California (2001); Vitiello, Three Strikes: Can We Return to Rationality? 87 J. Crim. *28L. & C. 395, 423 (1997). This criticism is appropriately directed at the legislature, which has primary responsibility for making the difficult policy choices that underlie any criminal sentencing scheme. We do not sit as a “superlegisla-ture” to second-guess these policy choices. It is enough that the State of California has a reasonable basis for believing that dramatically enhanced sentences for habitual felons “advance[s] the goals of [its] criminal justice system in any substantial way.” See Solem, 463 U. S., at 297, n. 22.

III

Against this backdrop, we consider Ewing’s claim that his three strikes sentence of 25 years to life is unconstitutionally disproportionate to his offense of “shoplifting three golf clubs.” Brief for Petitioner 6. We first address the gravity of the offense compared to the harshness of the penalty. At the threshold, we note that Ewing incorrectly frames the issue. The gravity of his offense was not merely “shoplifting three golf clubs.” Rather, Ewing was convicted of felony grand theft for- stealing nearly $1,200 worth of merchandise after previously having been convicted of at least two “violent” or “serious” felonies. Even standing alone, Ewing’s theft should not be taken lightly. His crime was certainly not “one of the most passive felonies a person could commit.” Solem, supra, at 296 (internal quotation marks omitted). To the contrary, the Supreme Court of California has noted the “seriousness” of grand theft in the context of proportionality review. See In re Lynch, 8 Cal. 3d 410, 432, n. 20, 503 P. 2d 921, 936, n. 20 (1972). Theft of $1,200 in property is a felony under federal law, 18 U. S. C. § 641, and in the vast majority of States. See App. B to Brief for Petitioner 21a.

That grand theft is a “wobbler” under California law is of no moment. Though California courts have discretion to reduce a felony grand theft charge to a misdemeanor, it remains a felony for all purposes “unless and until the trial *29court imposes a misdemeanor sentence.” In re Anderson, 69 Cal. 2d 618, 626, 447 P. 2d 117, 126 (1968) (Tobriner, J., concurring); see generally 1 B. Witkin & N. Epstein, California Criminal Law §73 (3d ed. 2000). “The purpose of the trial judge’s sentencing discretion” to downgrade certain felonies is to “impose a misdemeanor sentence in those cases in which the rehabilitation of the convicted defendant either does not require, or would be adversely affected by, incarceration in a state prison as a felon.” Anderson, supra, at 664-665, 447 P. 2d, at 152 (Tobriner, J., concurring). Under California law, the reduction is not based on the notion that a “wobbler” is “conceptually a misdemeanor.” Necochea v. Superior Court, 23 Cal. App. 3d 1012, 1016, 100 Cal. Rptr. 693, 695 (1972). Rather, it is “intended to extend misde-meanant treatment to a potential felon.” Ibid. In Ewing’s case, however, the trial judge justifiably exercised her discretion not to extend such lenient treatment given Ewing’s long criminal history.

In weighing the gravity of Ewing’s offense, we must place on the scales not only his current felony, but also his long history of felony recidivism. Any other approach would fail to accord proper deference to the policy judgments that find expression in the legislature’s choice of sanctions. In imposing a three strikes sentence, the State’s interest is not merely punishing the offense of conviction, or the “triggering” offense: “[I]t is in addition the interest... in dealing in a harsher manner with those who by repeated criminal acts have shown that they are simply incapable of conforming to the norms of society as established by its criminal law.” Rummel, 445 U. S., at 276; Solem, supra, at 296. To give full effect to the State’s choice of this legitimate penological goal, our proportionality review of Ewing’s sentence must take that goal into account.

Ewing’s sentence is justified by the State’s public-safety interest in incapacitating and deterring recidivist felons, and *30amply supported by his own long, serious criminal record.2 Ewing has been convicted of numerous misdemeanor and felony offenses, served nine separate terms of incarceration, and committed most of his crimes while on probation or parole. His prior “strikes” were serious felonies including robbery and three residential burglaries. To be sure, Ewing’s sentence is a long one. But it reflects a rational legislative judgment, entitled to deference, that offenders who have committed serious or violent felonies and who continue to commit felonies must be incapacitated. The State of California “was entitled to place upon [Ewing] the onus of one who is simply unable to bring his conduct within the social norms prescribed by the criminal law of the State.” Rummel, supra, at 284. Ewing’s is not “the rare case in which a threshold comparison of the crime committed and the sentence imposed leads to an inference of gross disproportion-ality.” Harmelin, 501 U. S., at 1005 (Kennedy, J., concurring in part and concurring in judgment).

We hold that Ewing’s sentence of 25 years to life in prison, imposed for the offense of felony grand theft under the three strikes law, is not grossly disproportionate and therefore does not violate the Eighth Amendment’s prohibition on *31cruel and unusual punishments. The judgment of the California Court of Appeal is affirmed.

It is so ordered.

Justice Scalia,

concurring in the judgment.

In my opinion in Harmelin v. Michigan, 501 U. S. 957, 985 (1991), I concluded that the Eighth Amendment’s prohibition of “cruel and unusual punishments” was aimed at excluding only certain modes of punishment, and was hot a “guarantee against disproportionate sentences.” Out of respect for the principle of stare decisis, I might nonetheless accept the contrary holding of Solem v. Helm, 463 U. S. 277 (1983) — that the Eighth Amendment contains a narrow proportionality principle — if I felt I could intelligently apply it. This case demonstrates why I cannot.

Proportionality — the notion that the punishment should fit the crime — is inherently a concept tied to the penological goal of retribution. “[I]t becomes difficult even to speak intelligently of ‘proportionality,’ once deterrence and rehabilitation are given significant weight,” Harmelin, supra, at 989 — not to mention giving weight to the purpose of California’s three strikes law: incapacitation. In the present case, the game is up once the plurality has acknowledged that “the Constitution does not mandate adoption of any one penological theory,” and that a “sentence can have a variety of justifications, such as incapacitation, deterrence, retribution, or rehabilitation.” Ante, at 25 (internal quotation marks omitted). That acknowledgment having been made, it no longer suffices merely to assess “the gravity of the offense compared to the harshness of the penalty,” ante, at 28; that classic description of the proportionality principle (alone and in itself quite resistant to policy-free, legal analysis) now becomes merely the “first” step of the inquiry, ibid. Having completed that step (by a discussion which, in all fairness, does not convincingly establish that 25-years-to-life is a “proportionate” punishment for stealing three golf clubs), the *32plurality must then add an analysis to show that “Ewing’s sentence is justified by the State’s public-safety interest in incapacitating and deterring recidivist felons.” Ante, at 29.

Which indeed it is — though why that has anything to do with the principle of proportionality is a mystery. Perhaps the plurality should revise its terminology, so that what it reads into the Eighth Amendment is not the unstated proposition that all punishment should be reasonably proportionate to the gravity of the offense, but rather the unstated proposition that all punishment should reasonably pursue the multiple purposes of the criminal law. That formulation would make it clearer than ever, of course, that the plurality is not applying law but evaluating policy.

Because I agree that petitioner’s sentence does not violate the Eighth Amendment’s prohibition against cruel and unusual punishments, I concur in the judgment.

Justice Thomas,

concurring in the judgment.

I agree with Justice Scalia’s view that the proportionality test announced in Solem v. Helm, 463 U. S. 277 (1983), is incapable of judicial application. Even were Solem’,s test perfectly clear, however, I would not feel compelled by stare decisis to apply it. In my view, the Cruel and Unusual Punishments Clause of the Eighth Amendment contains no proportionality principle. See Harmelin v. Michigan, 501 U. S. 957, 966-985 (1991) (opinion of Scalia, J.).

Because the plurality concludes that petitioner’s sentence does not violate the Eighth Amendment’s prohibition on cruel and unusual punishments, I concur in the judgment.

Justice Stevens,

with whom Justice Souter, Justice Ginsburg, and Justice Breyer join, dissenting.

Justice Breyer has cogently explained why the sentence imposed in this case is both cruel and unusual.1 The eoncur-*33rences prompt this separate writing to emphasize that proportionality review is not only capable of judicial application but also required by the Eighth Amendment.

“The Eighth Amendment succinctly prohibits ‘excessive’ sanctions.” Atkins v. Virginia, 536 U. S. 304, 311 (2002); see also U. S. Const., Arndt. 8 (“Excessive bail shall not be required, nor excessive fines imposed, nor cruel and unusual punishments inflicted”). Faithful to the Amendment’s text, this Court has held that the Constitution directs judges to apply their best judgment in determining the proportionality of fines, see, e. g., United States v. Bajakajian, 524 U. S. 321, 334-336 (1998), bail, see, e. g., Stack v. Boyle, 342 U. S. 1, 5 (1951), and other forms of punishment, including the imposition of a death sentence, see, e. g., Coker v. Georgia, 433 U. S. 584, 592 (1977). It “would be anomalous indeed” to suggest that the Eighth Amendment makes proportionality review applicable in the context of bail and fines but not in the context of other forms of punishment, such as imprisonment. Solem v. Helm, 463 U. S. 277, 289 (1983). Rather, by broadly prohibiting excessive sanctions, the Eighth Amendment directs judges to exercise their wise judgment in assessing the proportionality of all forms of punishment.

The absence of a black-letter rule does not disable judges from exercising their discretion in construing the outer limits on sentencing authority that the Eighth Amendment imposes. After all, judges are “constantly called upon to draw . . . lines in a variety of contexts,” id., at 294, and to exercise their judgment to give meaning to the Constitution’s broadly phrased protections. For example, the Due Process Clause directs judges to employ proportionality re*34view in assessing the constitutionality of punitive damages awards on a case-by-case basis. See, e. g., BMW of North America, Inc. v. Gore, 517 U. S. 559, 562 (1996). Also, although the Sixth Amendment guarantees criminal defendants the right to a speedy trial, the courts often are asked to determine on a case-by-case basis whether a particular delay is constitutionally permissible or not. See, e. g., Doggett v. United States, 505 U. S. 647 (1992).2

Throughout most of the Nation’s history — before guideline sentencing became so prevalent — federal and state trial judges imposed specific sentences pursuant to grants of authority that gave them uncabined discretion within broad ranges. See K. Stith & J. Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts 9 (1998) (hereinafter Stith & Cabranes) (“From the beginning of the Republic, federal judges were entrusted with wide sentencing discretion”); see also Mistretta v. United States, 488 U. S. 361, 364 (1989). It was not unheard of for a statute to authorize a sentence ranging from one year to life, for example. See, e. g., State v. Perley, 86 Me. 427, 30 A. 74, 75 (1894) (citing Maine statute that made robbery punishable by imprisonment for life or any term of years); In re Southard, 298 Mich. 75, 77, 298 N. W. 457 (1941) (“The offense of ‘robbery armed’ is punishable by imprisonment for life or any term *35of years”). In exercising their discretion, sentencing judges wisely employed a proportionality principle that took into account all of the justifications for punishment — namely, deterrence, incapacitation, retribution, and rehabilitation. See Stith & Cabranes 14. Likewise, I think it clear that the Eighth Amendment’s prohibition of “cruel and unusual punishments” expresses a broad and basic proportionality principle that takes into account all of the justifications for penal sanctions. It is this broad proportionality principle that would preclude reliance on any of the justifications for punishment to support, for example, a life sentence for overtime parking. See Rummel v. Estelle, 445 U. S. 263, 274, n. 11 (1980).

Accordingly, I respectfully dissent.

Justice Breyer,

with whom Justice Stevens, Justice Souter, and Justice Ginsburg join, dissenting.

The constitutional question is whether the “three strikes” sentence imposed by California upon repeat-offender Gary Ewing is “grossly disproportionate” to his crime. Ante, at 14, 30-31 (plurality opinion). The sentence amounts to a real prison term of at least 25 years. The sentence-triggering criminal conduct consists of the theft of three golf clubs priced at a total of $1,197. See ante, at 18. The offender has a criminal history that includes four felony convictions arising out of three separate burglaries (one armed). Ante, at 18-19. In Solem v. Helm, 463 U. S. 277 (1983), the Court found grossly disproportionate a somewhat longer sentence imposed on a recidivist offender for triggering criminal conduct that was somewhat less severe. In my view, the differences are not determinative, and the Court should reach the same ultimate conclusion here.

I

This Court’s precedent sets forth a framework for analyzing Ewing’s Eighth Amendment claim. The Eighth Amendment forbids, as “cruel and unusual punishments,” prison *36terms (including terms of years) that are “grossly disproportionate.” Solem, supra, at 303; see Lockyer v. Andrade, post, at 71. In applying the “gross disproportionality” principle, courts must keep in mind that “legislative policy” will primarily determine the appropriateness of a punishment’s “severity,” and hence defer to such legislative policy judgments. Gore v. United States, 357 U. S. 386, 393 (1958); see Harmelin v. Michigan, 501 U. S. 957, 998 (1991) (Kennedy, J., concurring in part and concurring in judgment); Solem, supra, at 289-290; Rummel v. Estelle, 445 U. S. 263, 274-276 (1980); Weems v. United States, 217 U. S. 349, 373 (1910). If courts properly respect those judgments, they will find that the sentence fails the test only in rare instances. Solem, supra, at 290, n. 16; Harmelin, supra, at 1004 (Kennedy, J., concurring in part and concurring in judgment); Rummel, supra, at 272 (“[Successful challenges to the proportionality of particular sentences have been exceedingly rare”). And they will only “‘rarely’” find it necessary to “‘engage in extended analysis’” before rejecting a claim that a sentence is “grossly disproportionate.” Harmelin, supra, at 1004 (Kennedy, J., concurring in part and concurring in judgment) (quoting Solem, supra, at 290, n. 16).

The plurality applies Justice Kennedy’s analytical framework in Harmelin, supra, at 1004-1005 (opinion concurring in part and concurring in judgment). Ante, at 23-24. And, for present purposes, I will consider Ewing’s Eighth Amendment claim on those terms. But see ante, at 32-33, n. 1 (Stevens, J., dissenting). To implement this approach, courts faced with a “gross disproportionality” claim must first make “a threshold comparison of the crime committed and the sentence imposed.” Harmelin, supra, at 1005 (Kennedy, J., concurring in part and concurring in judgment). If a claim crosses that threshold — itself a rare occurrence — then the court should compare the sentence at issue to other sentences “imposed on other criminals” in the same, or in other, jurisdictions. Solem, supra, at 290-291; *37Harmelin, 501 U. S., at 1005 (Kennedy, J., concurring in part and concurring in judgment). The comparative analysis will “validate” or invalidate “an initial judgment that a sentence is grossly disproportionate to a crime.” Ibid.

I recognize the warnings implicit in the Court’s frequent repetition of words such as “rare.” Nonetheless I believe that the case before us is a “rare” case — one in which a court can say with reasonable confidence that the punishment is “grossly disproportionate” to the crime.

II

Ewing’s claim crosses the gross disproportionality “threshold.” First, precedent makes clear that Ewing’s sentence raises a serious disproportionality question. Ewing is a recidivist. Hence the two cases most directly in point are those in which the Court considered the constitutionality of recidivist sentencing: Rummel and Solem. Ewing’s claim falls between these two cases. It is stronger than the claim presented in Rummel, where the Court upheld a recidivist’s sentence as constitutional. It is weaker than the claim presented in Solem, where the Court struck down a recidivist sentence as unconstitutional.

Three kinds of sentence-related characteristics define the relevant comparative spectrum: (a) the length of the prison term in real time, i. e., the time that the offender is likely actually to spend in prison; (b) the sentence-triggering criminal conduct, i. e., the offender’s actual behavior or other offense-related circumstances; and (c) the offender’s criminal history. See Rummel, supra, at 265-266, 269, 276, 278, 280-281 (using these factors); Solem, supra, at 290-303 (same). Cf. United States Sentencing Commission, Guidelines Manual ch. 1, pt. A, intro., n. 5 (Nov. 1987) (USSG) (empirical study of “summary reports of some 40,000 convictions [and] a sample of 10,000 augmented presentence reports” leads to sentences based primarily upon (a) offense characteristics and (b) offender’s criminal record); see id., p. s. 3.

*38In Rummel, the Court held constitutional (a) a sentence of life imprisonment with parole available within 10 to 12 years, (b) for the offense of obtaining $120 by false pretenses, (c) committed by an offender with two prior felony convictions (involving small amounts of money). 445 U. S. 263; ante, at 21. In Solem, the Court held unconstitutional (a) a sentence of life imprisonment without parole, (b) for the crime of writing a $100 check on a nonexistent bank account, (c) committed by an offender with six prior felony convictions (including three for burglary). 463 U. S. 277; ante, at 22-23. Which of the three pertinent comparative factors made the constitutional difference?

The third factor, prior record, cannot explain the difference. The offender’s prior record was worse in Solem, where the Court found the sentence too long, than in Rum-mel, where the Court upheld the sentence. The second factor, offense conduct, cannot explain the difference. The nature of the triggering offense — viewed in terms of the actual monetary loss — in the two cases was about the same. The one critical factor that explains the difference in the outcome is the length of the likely prison term measured in real time. In Rummel, where the Court upheld the sentence, the state sentencing statute authorized parole for the offender, Rum-mel, after 10 or 12 years. 445 U. S., at 280; id., at 293 (Powell, J., dissenting). In Solem, where the Court struck down the sentence, the sentence required the offender, Helm, to spend the rest of his life in prison.

Now consider the present case. The third factor, offender characteristics — i. e., prior record — does not differ significantly here from that in Solem. Ewing’s prior record consists of four prior felony convictions (involving three burglaries, one with a knife) contrasted with Helm’s six prior felony convictions (including three burglaries, though none with weapons). The second factor, offense behavior, is worse than that in Solem, but only to a degree. It would be difficult to say that the actual behavior itself here (shop*39lifting) differs significantly from that at issue in Solem (passing a bad check) or in Rummel (obtaining money through false pretenses). Rather the difference lies in the value of the goods obtained. That difference, measured in terms of the most relevant feature (loss to the victim, i. e., wholesale value) and adjusted for the irrelevant feature of inflation, comes down (in 1979 values) to about $379 here compared with $100 in Solem, or (in 1973 values) to $232 here compared with $120.75 in Rummel. See USSG §2B1.1, comment., n. 2(A)(i) (Nov. 2002) (loss to victim properly measures value of goods unlawfully taken); U. S. Dept. of Labor, Bureau of Labor Statistics, Inflation and Consumer Spending, Inflation Calculator (Jan. 23, 2003), http://www.bls.gov (hereinafter Inflation Calculator). Alternatively, if one measures the inflation-adjusted value difference in terms of the golf clubs’ sticker price, it comes down to $505 here compared to $100 in Solem, or $309 here compared to $120.75 in Rummel. See Inflation Calculator.

The difference in length of the real prison term — the first, and critical, factor in Solem and Rummel — is considerably more important. Ewing’s sentence here amounts, in real terms, to at least 25 years without parole or good-time credits. That sentence is considerably shorter than Helm’s sentence in Solem, which amounted, in real terms, to life in prison. Nonetheless Ewing’s real prison term is more than twice as long as the term at issue in Rummel, which amounted, in real terms, to at least 10 or 12 years. And, Ewing’s sentence, unlike Rummel’s (but like Helm’s sentence in Solem), is long enough to consume the productive remainder of almost any offender’s life. (It means that Ewing himself, seriously ill when sentenced at age 38, will likely die in prison.)

The upshot is that the length of the real prison term — the factor that explains the Solem/Rummel difference in outcome — places Ewing closer to Solem than to Rummel, though the greater value of the golf clubs that Ewing stole *40moves Ewing’s case back slightly in Rummel’s direction. Overall, the comparison places Ewing’s sentence well within the twilight zone between Solem and Rummel — a zone where the argument for unconstitutionality is substantial, where the cases themselves cannot determine the constitutional outcome.

Second, Ewing’s sentence on its face imposes one of the most severe punishments available upon a recidivist who subsequently engaged in one of the less serious forms of criminal conduct. See infra, at 44-45. I do not deny the seriousness of shoplifting, which an amicus curiae tells us costs retailers in the range of $30 billion annually. Brief for California District Attorneys Association as Amicus Curiae 27. But consider that conduct in terms of the factors that this Court mentioned in Solem — the “harm caused or threatened to the victim or society,” the “absolute magnitude of the crime,” and the offender’s “culpability.” 463 U. S., at 292-293. In respect to all three criteria, the sentence-triggering behavior here ranks well toward the bottom of the criminal conduct scale.

The Solicitor General has urged us to consider three other criteria: the “frequency” of the crime’s commission, the “ease or difficulty of detection,” and “the degree to which the crime may be deterred by differing amounts of punishment.” Brief for United States as Amicus Curiae 24-25. When considered in terms of these criteria — or at least the latter two — the triggering conduct also ranks toward the bottom of the scale. Unlike, say, drug crimes, shoplifting often takes place in stores open to other customers whose presence, along with that of store employees or cameras, can help to detect the crime. Nor is there evidence presented here that the law enforcement community believes lengthy prison terms necessary adequately to deter shoplifting. To the contrary, well-publicized instances of shoplifting suggest that the offense is often punished without any prison sentence at all. On the other hand, shoplifting is a frequently com*41mitted crime; but “frequency,” standing alone, cannot make a critical difference. Otherwise traffic offenses would warrant even more serious punishment.

This ease, of course, involves shoplifting engaged in by a recidivist. One might argue that any crime committed by a recidivist is a serious crime potentially warranting a 25-year sentence. But this Court rejected that view in Solem, and in Harmelin, with the recognition that “no penalty is per se constitutional.” Solem, supra, at 290; Harmelin, 501 U. S., at 1001 (Kennedy, J., concurring in part and concurring in judgment). Our cases make clear that, in cases involving recidivist offenders, we must focus upon “the [offense] that triggers the life sentence,” with recidivism playing a “relevant,” but not necessarily determinative, role. Solem, supra, at 296, n. 21; see Witte v. United States, 515 U. S. 389, 402, 403 (1995) (the recidivist defendant is “punished only for the offense of conviction,” which “‘is considered to be an aggravated offense because a repetitive one’” (quoting Gryger v. Burke, 334 U. S. 728, 732 (1948))). And here, as I have said, that offense is among the less serious, while the punishment is among the most serious. Cf. Rummel, 445 U. S., at 288 (Powell, J., dissenting) (overtime parking violation cannot trigger a life sentence even for a serious recidivist).

Third, some objective evidence suggests that many experienced judges would consider Ewing’s sentence disproportionately harsh. The United States Sentencing Commission (having based the federal Sentencing Guidelines primarily upon its review of how judges had actually sentenced offenders) does not include shoplifting (or similar theft-related offenses) among the crimes that might trigger especially long sentences for recidivists, see USSG §4B1.1 (Nov. 2002) (Guideline for sentencing “career offenders”); id., ch. 1, pt. A, intro., n. 5 (sentences based in part upon Commission’s review of “summary reports of some 40,000 convictions [and] a sample of 10,000 augmented presentence reports”); see also *42infra, at 45, nor did Congress include such offenses among triggering crimes when it sought sentences “at or near the statutory maximum” for certain recidivists, S. Rep. No. 98-225, p. 175 (1983); 28 U. S. C. § 994(h) (requiring sentence “at or near the maximum” where triggering crime is crime of “violence” or drug related); 18 U. S. C. § 3559(c) (grand theft not among triggering or “strike” offenses under federal “three strikes” law); see infra, at 45-46. But see 28 U. S. C. § 994(i)(l) (requiring “a substantial term of imprisonment” for those who have “a history of two or more prior ... felony convictions”).

Taken together, these three circumstances make clear that Ewing’s “gross disproportionality” argument is a strong one. That being so, his claim must pass the “threshold” test. If it did not, what would be the function of the test? A threshold test must permit arguably unconstitutional sentences, not only actually unconstitutional sentences, to pass the threshold — at least where the arguments for unconstitutionality are unusually strong ones. A threshold test that blocked every ultimately invalid constitutional claim — even strong ones — would not be a threshold test but a determinative test. And, it would be a determinative test that failed to take account of highly pertinent sentencing information, namely, comparison with other sentences, Solem, supra, at 291-292, 298-300. Sentencing comparisons are particularly important because they provide proportionality review with objective content. By way of contrast, a threshold test makes the assessment of constitutionality highly subjective. And, of course, so to transform that threshold test would violate this Court’s earlier precedent. See 463 U. S., at 290, 291-292; Harmelin, supra, at 1000, 1005 (Kennedy, J., concurring in part and concurring in judgment).

> — l ► — <

Believing Ewing’s argument a strong one, sufficient to pass the threshold, I turn to the comparative analysis. A *43comparison of Ewing’s sentence with other sentences requires answers to two questions. First, how would other jurisdictions (or California at other times, i. e,, without the three strikes penalty) punish the same offense conduct? Second, upon what other conduct would other jurisdictions (or California) impose the same prison term? Moreover, since hypothetical punishment is beside the point, the relevant prison time, for comparative purposes, is real prison time, i. e., the time that an offender must actually serve.

Sentencing statutes often shed little light upon real prison time. That is because sentencing laws normally set maximum sentences, giving the sentencing judge discretion to choose an actual sentence within a broad range, and because many States provide good-time credits and parole, often permitting release after, say, one-third of the sentence has been served, see, e.g., Alaska Stat. § 33.20.010(a) (2000); Conn. Gen. Stat. § 18-7a (1998). Thus, the statutory maximum is rarely the sentence imposed, and the sentence imposed is rarely the sentence that is served. For the most part, the parties’ briefs discuss sentencing statutes. Nonetheless, that discussion, along with other readily available information, validates my initial belief that Ewing’s sentence, comparatively speaking, is extreme.

As to California itself, we know the following: First, between the end of World War II and 1994 (when California enacted the three strikes law, ante, at 15), no one like Ewing could have served more than 10 years in prison. We know that for certain because the maximum sentence for Ewing’s crime of conviction, grand theft, was for most of that period 10 years. Cal. Penal Code Ann. §§ 484, 489 (West 1970); see Cal. Dept. of Corrections, Offender Information Services, Administrative Services Division, Historical Data for Time Served by Male Felons Paroled from Institutions: 1945 Through 1981, p. 11 (1982) (Table 10) (hereinafter Historical Data for Time Served by California Felons), Lodging of Petitioner. From 1976 to 1994 (and currently, absent application *44of the three strikes penalty), a Ewing-type offender would have received a maximum sentence of four years. Cal. Penal Code Ann. §489 (West 1999), § 667.5(b) (West Supp. 2002). And we know that California’s “habitual offender” laws did not apply to grand theft. §§ 644(a), (b) (West 1970) (repealed 1977). We also know that the time that any offender actually served was likely far less than 10 years. This is because statistical data show that the median time actually served for grand theft (other than auto theft) was about two years, and 90 percent of all those convicted of that crime served less than three or four years. Historical Data for Time Served by California Felons 11 (Table 10).

Second, statistics suggest that recidivists of all sorts convicted during that same time period in California served a small fraction of Ewing’s real-time sentence. On average, recidivists served three to four additional (recidivist-related) years in prison, with 90 percent serving less than an additional real seven to eight years. Id., at 22 (Table 21).

Third, we know that California has reserved, and still reserves, Ewing-type prison time, i. e., at least 25 real years in prison, for criminals convicted of crimes far worse than was Ewing’s. Statistics for the years 1945 to 1981, for example, indicate that typical.(nonrecidivist) male first-degree murderers served between 10 and 15 real years in prison, with 90 percent of all such murderers serving less than 20 real years. Id., at 3 (Table 2). Moreover, California, which has moved toward a real-time sentencing system (where the statutory punishment approximates the time served), still punishes far less harshly those who have engaged in far more serious conduct. It imposes, for example, upon nonrecidi-vists guilty of arson causing great bodily injury a maximum sentence of nine years in prison, Cal. Penal Code Ann. § 451(a) (West 1999) (prison term of 5, 7, or 9 years for arson that causes great bodily injury); it imposes upon those guilty of voluntary manslaughter a maximum sentence of 11 years, § 193 (prison term of 3, 6, or 11 years for voluntary man*45slaughter). It reserves the sentence that it here imposes upon (former-burglar-now-golf-club-thief) Ewing for non-recidivist, first-degree murderers. See § 190(a) (West Supp. 2003) (sentence of 25 years to life for first-degree murder).

As to other jurisdictions, we know the following: The United States, bound by the federal Sentencing Guidelines, would impose upon a recidivist, such as Ewing, a sentence that, in any ordinary case, would not exceed 18 months in prison. USSG §2Bl.l(a) (Nov. 1999) (assuming a base offense level of 6, a criminal history of VI, and no mitigating or aggravating adjustments); id., ch. 5, pt. A, Sentencing Table. The Guidelines, based in part upon a study of some 40,000 actual federal sentences, see supra, at 37, 41, reserve a Ewing-type sentence for Ewing-type recidivists who currently commit such crimes as murder, §2A1.2; air piracy, §2A5.1; robbery (involving the discharge of a firearm, serious bodily injury, and about $1 million), §2B3.1; drug offenses involving more than, for example, 20 pounds of heroin, §2D1.1; aggravated theft of more than $100 million, § 2B1.1; and other .similar offenses. The Guidelines reserve 10 years of real prison time (with good time) — less than 40 percent of Ewing’s sentence — for Ewing-type recidivists who go on to commit, for instance, voluntary manslaughter, §2A1.3; aggravated assault with a firearm (causing serious bodily injury and motivated by money), §2A2.2; kidnaping, §2A4.1; residential burglary involving more than $5 million, §2B2.1; drug offenses involving at least one pound of cocaine, § 2D 1.1; and other similar offenses. Ewing also would not have been subject to the federal “three strikes” law, 18 U. S. C. § 3559(c), for which grand theft is not a triggering offense.

With three exceptions, see infra, at 46-47, we do not have before us information about actual time served by Ewing-type offenders in other States. We do know, however, that the law would make it legally impossible for a Ewing-type offender to serve more than 10 years in prison in 33 jurisdictions, as well as the federal courts, see Appendix, *46Part A, infra, more than 15 years in 4 other States, see Appendix, Part B, infra, and more than 20 years in 4 additional States, see Appendix, Part C, infra. In nine other States, the law might make it legally possible to impose a sentence of 25 years or more, see Appendix, Part D, infra — though that fact by itself, of course, does not mean that judges have actually done so. But see infra this page. I say “might” because the law in five of the nine last mentioned States restricts the sentencing judge’s ability to impose a term so long that, with parole, it would amount to at least 25 years of actual imprisonment. See Appendix, Part D, infra.

We also know that California, the United States, and other States supporting California in this case, despite every incentive to find someone else like Ewing who will have to serve, or who has actually served, a real prison term anywhere approaching that imposed upon Ewing, have come up with precisely three examples. Brief for United States as Amicus Curiae 28-29, n. 13. The Government points to Ex parte Howington, 622 So. 2d 896 (Ala. 1993), where an Alabama court sentenced an offender with three prior burglary convictions and two prior grand theft convictions to “life” for the theft of a tractor-trailer. The Government also points to State v. Heftel, 513 N. W. 2d 397 (S. D. 1994), where a South Dakota court sentenced an offender with seven prior felony convictions to 50 years’ imprisonment for theft. And the Government cites Sims v. State, 107 Nev. 438, 814 P. 2d 63 (1991), where a Nevada court sentenced a defendant with three prior felony convictions (including armed robbery) and nine misdemeanor convictions to life without parole for the theft of a purse and wallet containing $476.

The first of these cases, Howington, is beside the point, for the offender was eligible for parole after 10 years (as in Rummel), not 25 years (as here). Ala. Code §15-22-28(e) (West 1982). The second case, Heftel, is factually on point, but it is not legally on point, for the South Dakota courts did not consider the constitutionality of the sentence. 513 N. W. *472d, at 401. The third case, Sims, is on point both factually and legally, for the Nevada Supreme Court (by a vote of 3 to 2) found the sentence constitutional. I concede that example — a single instance of a similar sentence imposed outside the context of California’s three strikes law, out of a prison population now approaching two million individuals. U. S. Dept. of Justice, Office of Justice Programs, Bureau of Justice Statistics, Prison Statistics (Jan. 8, 2003), http://www.ojp.usdoj.gov/bjs/prisons.htm (available in Clerk of Court’s case file).

The upshot is that comparison of other sentencing practices, both in other jurisdictions and in California at other times (or in respect to other crimes), validates what an initial threshold examination suggested. Given the information available, given the state and federal parties’ ability to provide additional contrary data, and given their failure to do so, we can assume for constitutional purposes that the following statement is true: Outside the California three strikes context, Ewing’s recidivist sentence is virtually unique in its harshness for his offense of conviction, and by a considerable degree.

IV

This is not the end of the matter. California sentenced Ewing pursuant to its “three strikes” law. That law represents a deliberate effort to provide stricter punishments for recidivists. Cal. Penal Code Ann. § 667(b) (West 1999) (“It is the intent of the Legislature ... to ensure longer prison sentences and greater punishment for those who commit a felony and have been previously convicted of serious and/or violent felony offenses”); ante, at 24. And, it is important to consider whether special criminal justice concerns related to California’s three strikes policy might justify including Ewing’s theft within the class of triggering criminal conduct (thereby imposing a severe punishment), even if Ewing’s sentence would otherwise seem disproportionately harsh. *48Cf. Harmelin, 501 U. S., at 998-999, 1001 (noting “the primacy of the legislature” in making sentencing policy).

I can find no such special criminal justice concerns that might justify this sentence. The most obvious potential justification for bringing Ewing’s theft within the ambit of the statute is administrative. California must draw some kind of workable line between conduct that will trigger, and con-' duct that will not trigger, a “three strikes” sentence. “But the fact that a line has to be drawn somewhere does not justify its being drawn anywhere.” Pearce v. Commissioner, 315 U. S. 543, 558 (1942) (Frankfurter, J., dissenting). The statute’s administrative objective would seem to be one of separating more serious, from less serious, triggering criminal conduct. Yet the statute does not do that job particularly well.

The administrative line that the statute draws separates “felonies” from “misdemeanors.” See Brief for Respondent 6 (“The California statute relies, fundamentally, on traditional classifications of certain crimes as felonies”). Those words suggest a graduated difference in degree. But an examination of how California applies these labels in practice to criminal conduct suggests that the offenses do not necessarily reflect those differences. See United States v. Watson, 423 U. S. 411, 438-441 (1976) (Marshall, J., dissenting) (felony/misdemeanor distinction often reflects history, not logic); Rummel, 445 U. S., at 284 (“The most casual review of the various criminal justice systems now in force in the 50 States of the Union shows that the line dividing felony theft from petty larceny, a line usually based on the value of the property taken, varies markedly from one State to another”). Indeed, California uses those words in a way unrelated to the seriousness of offense conduct in a set of criminal statutes called “ ‘wobblers,’ ” see ante, at 16, one of which is at issue in this case.

Most “wobbler” statutes classify the same criminal conduct either as a felony or as a misdemeanor, depending upon *49the actual punishment imposed, Cal. Penal Code Ann. §§ 17(a), (b) (West 1999); ante, at 16-17, which in turn depends primarily upon whether “the rehabilitation of the convicted defendant” either does or does not “require” (or would or would not “be adversely affected by”) “incarceration in a state prison as a felon.” In re Anderson, 69 Cal. 2d 613, 664-665, 447 P. 2d 117, 152 (1968) (Tobriner, J., concurring in part and dissenting in part); ante, at 29. In such cases, the felony/misdemeanor classification turns primarily upon the nature of the offender, not the comparative seriousness of the offender’s conduct.

A subset of “wobbler” statutes, including the “petty theft with a prior” statute, Cal. Penal Code Ann. § 666 (West Supp. 2002), defining the crime in the companion case, Lockyer v. Andrade, post, p. 63, authorizes the treatment of otherwise misdemeanor conduct, see Cal. Penal Code Ann. § 490 (West 1999), as a felony only when the offender has previously committed a property crime. Again, the distinction turns upon characteristics of the offender, not the specific offense conduct at issue.

The result of importing this kind of distinction into California’s three strikes statute is a series of anomalies. One anomaly concerns the seriousness of the triggering behavior. “Wobbler” statutes cover a wide variety of criminal behavior, ranging from assault with a deadly weapon, §245, vehicular manslaughter, § 193(c)(1), and money laundering, § 186.10(a), to the defacement of property with graffiti, § 594(b)(2)(A) (West Supp. 2002), or stealing more than $100 worth of chickens, nuts, or avocados, § 487(b)(1)(A) (West Supp. 2003); § 489 (West 1999). Some of this behavior is obviously less serious, even if engaged in twice, than other criminal conduct that California statutes classify as pure misdemeanors, such as reckless driving, Cal. Veh. Code Ann. §23103 (West Supp. 2003); § 23104(a) (West 2000) (reckless driving causing bodily injury), the use of force or threat of force to interfere with another’s civil rights, Cal. Penal Code *50Ann. §422.6 (West 1999), selling poisoned alcohol, §347b, child neglect, § 270, and manufacturing or selling false government documents with the intent to conceal true citizenship, § 112(a) (West Supp. 2002).

Another anomaly concerns temporal order. An offender whose triggering crime is his third crime likely will not fall within the ambit of the three strikes statute provided that (a) his first crime was chicken theft worth more than $100, and (b) he subsequently graduated to more serious crimes, say, crimes of violence. That is because such chicken theft, when a first offense, will likely be considered a misdemeanor. A similar offender likely will fall within the scope of the three strikes statute, however, if such chicken theft was his third crime. That is because such chicken theft, as a third offense, will likely be treated as a felony.

A further anomaly concerns the offender’s criminal record. California’s “wobbler” “petty theft with a prior” statute, at issue in Lockyer v. Andrade, post, p. 63, classifies a petty theft as a “felony” if, but only if, the offender has a prior record that includes at least one conviction for certain theft-related offenses. Cal. Penal Code Ann. §666 (West Supp. 2002). Thus a violent criminal who has committed two violent offenses and then steals $200 will not fall within the ambit of the three strikes statute, for his prior record reveals no similar property crimes. A similar offender will fall within the scope of the three strikes statute, however, if that offender, instead of having committed two previous violent crimes, has committed one previous violent crime and one previous petty theft. (Ewing’s conduct would have brought him within the realm of the petty theft statute prior to 1976 but for inflation.)

At the same time, it is difficult to find any strong need to define the lower boundary as the State has done. The three strikes statute itself, when defining prior “strikes,” simply lists the kinds of serious criminal conduct that falls within the definition of a “strike.” § 667.5(c) (listing “violent” felon*51ies); § 1192.7(e) (West Supp. 2003) (listing “serious” felonies). There is no obvious reason why the statute could not enumerate, consistent with its purposes, the relevant triggering crimes. Given that possibility and given the anomalies that result from California’s chosen approach, I do not see how California can justify on administrative grounds a sentence as seriously disproportionate as Ewing’s. See Parts II and III, supra.

Neither do I see any other way in which inclusion of Ewing’s conduct (as a “triggering crime”) would further a significant criminal justice objective. One might argue that those who commit several property crimes should receive long terms of imprisonment in order to “incapacitate” them, i. e., to prevent them from committing further crimes in the future. But that is not the object of this particular three strikes statute. Rather, as the plurality says, California seeks “‘to reduce serious and violent crime.’” Ante, at 24 (quoting Ardaiz, California’s Three Strikes Law: History, Expectations, Consequences, 32 McGeorge L. Rev. 1 (2000) (emphasis added)). The statute’s definitions of both kinds of crime include crimes against the person, crimes that create danger of physical harm, and drug crimes. See, e. g., Cal. Penal Code Ann. § 667.5(c)(1) (West Supp. 2002), § 1192.7(c)(1) (West Supp. 2003) (murder or voluntary manslaughter); §667.5(c)(21) (West Supp. 2002), § 1192.7(c)(18) (West Supp. 2003) (first-degree burglary); § 1192.7(c)(24) (selling or giving or offering to sell or give heroin or cocaine to a minor). They do not include even serious crimes against property, such as obtaining large amounts of money, say, through theft, embezzlement, or fraud. Given the omission of vast categories of property crimes — including grand theft (unarmed)— from the “strike” definition, one cannot argue, on property-crime-related incapacitation grounds, for inclusion of Ewing’s crime among the triggers.

Nor do the remaining criminal law objectives seem relevant. No one argues for Ewing’s inclusion within the ambit *52of the three strikes statute on grounds of “retribution.” Cf. Vitiello, Three Strikes: Can We Return to Rationality? 87 J. Crim. L. & C. 395, 427 (1997) (California’s three strikes law, like other “[hjabitual offender statutes[, is] not retributive” because the term of imprisonment is “imposed without regard to the culpability of the offender or [the] degree of social harm caused by the offender’s behavior,” and “has little to do with the gravity of the offens[e]”). For reasons previously discussed, in terms of “deterrence,” Ewing’s 25-year term amounts to overkill. See Parts II and III, supra. And “rehabilitation” is obviously beside the point. The upshot is that, in my view, the State cannot find in its three strikes law a special criminal justice need sufficient to rescue a sentence that other relevant considerations indicate is unconstitutional.

V

Justice Scalia and Justice Thomas argue that we should not review for gross disproportionality a sentence to a term of years. Ante, at 81 (Scalia, J., concurring in judgment); ante, at 32 (Thomas, J., concurring in judgment). Otherwise, we make it too difficult for legislators and sentencing judges to. determine just when their sentencing laws and practices pass constitutional muster.

I concede that a bright-line rule would give legislators and sentencing judges more guidance. But application of the Eighth Amendment to a sentence of a term of years requires a case-by-ease approach. And, in my view, like that of the plurality, meaningful enforcement of the Eighth Amendment demands that application — even if only at sentencing’s outer bounds.

A case-by-case approach can nonetheless offer guidance through example. Ewing’s sentence is, at a minimum, 2 to 3 times the length of sentences that other jurisdictions would impose in similar circumstances. That sentence itself is sufficiently long to require a typical offender to spend virtually all the remainder of his active life in prison. These and the *53other factors that I have discussed, along with the questions that I have asked along the way, should help to identify “gross disproportionality” in a fairly objective way — at the outer bounds of sentencing.

In stun, even if I accept for present purposes the plurality’s analytical framework, Ewing’s sentence (life imprisonment with a minimum term of 25 years) is grossly disproportionate to the triggering offense conduct — stealing three golf clubs — Ewing’s recidivism notwithstanding.

For these reasons, I dissent.

APPENDIX TO OPINION OF BREYER, J.

A

Thirty-three jurisdictions, as well as the federal courts, have laws that would make it impossible to sentence a Ewing-type offender to more than 10 years in prison:1

Federal: 12 to 18 months. USSG §2B1.1 (Nov. 1999); id., ch. 5, pt. A, Sentencing Table.

Alaska: three to five years; presumptive term of three years. Alaska Stat. §§ 11.46.130(a)(1), (c), 12.55.125(e) (2000).

Arizona: four to six years; presumptive sentence of five years. Ariz. Rev. Stat. Ann. §§ 13-604(C), 13-1802(E) (West 2001).

Connecticut: 1 to 10 years. Conn. Gen. Stat. §§53a-35a(6), 53a-40(j), 53a-124(a)(2) (2001).

Delaware: not more than two years. Del. Code Ann., Tit. 11, § 840(d) (Supp. 2000); § 4205(b)(7) (1995). Recidivist offender penalty not applicable. See §4214; Buckingham v. State, 482 A. 2d 327 (Del. 1984).

District of Columbia: not more than 10 years. D. C. Code Ann. §22-3212(a) (West 2001). Recidivist offender penalty *54not applicable. See § 22-1804a(c)(2) (West 2001) (amended 2001).

Florida: not more than 10 years. Fla. Stat. Ann. §§775.084(l)(a), (4)(a)(3) (West 2000) (amended 2002); § 812.014(c)(1) (West 2000).

Georgia: 10 years. Ga. Code Ann. § 16-8-12(a)(1) (1996); § 17-10-7(a) (Supp. 1996).

Hawaii: 20 months. Haw. Rev. Stat. §§ 708-831(1)(b), 706-606.5(1)(a)(iv), (7)(a) (Supp. 2001).

Idaho: 1 to 14 years. Idaho Code §§ 18-2403, 18-2407(b)(1), 18-2408(2)(a) (1948-1997). Recidivist/habitual' offender penalty of five years to life in prison, §19-2514, likely not applicable. Idaho has a general rule that “ ‘convictions entered the same day or charged in the same information should count as a single conviction for purposes of establishing habitual offender status.’ ” State v. Harrington, 133 Idaho 563, 565, 990 P. 2d 144, 146 (App. 1999) (quoting State v. Brandt, 110 Idaho 341, 344, 715 P. 2d 1011, 1014 (App. 1986)). However, “the nature of the convictions in any given situation must be examined to make certain that [this] general rule is appropriate.” Ibid. In this case, Ewing’s prior felony convictions stemmed from acts committed at the same apartment complex, and three of the four felonies were committed within a day of each other; the fourth offense was committed five weeks earlier. See App. 6; Tr. 45-46 (Information, Case No. NA018343-01 (Cal. Super. Ct.) (available in Clerk of Court’s ease file)). A review of Idaho ease law suggests that this case is factually distinguishable from cases in which the Idaho courts have declined to adhere to the general rule. See, e. g., Brandt, supra, at 348, 344, 715 P. 2d, at 1013, 1014 (three separately charged property offenses involving three separate homes and different victims committed “during a two-month period”); State v. Mace, 133 Idaho 903, 907, 994 P. 2d 1066, 1070 (App. 2000) (unrelated crimes (grand theft and DU I) committed on different dates in different counties); State v. Smith, 116 Idaho 553, 560, 777 *55P. 2d 1226, 1233 (App. 1989) (separate and distinguishable crimes committed on different victims in different counties).

Illinois: two to five years. Ill. Comp. Stat., ch. 730, §5/ 5-8-1(a)(6) (Supp. 2001); ch. 720, § 5/16-1(b)(4). Recidivist offender penalty not applicable. § 5/33B-1(a) (2000).

Indiana: 18 months (with not more than 18 months added for aggravating circumstances). Ind. Code § 35-43-4-2(a) (1993); §35-50-2-7(a). Recidivist offender penalty not applicable. See § 35-50-2-8 (amended 2001).

Iowa: three to five years. Iowa Code Ann. §§714.2(2), 902.9(5) (West Supp. 2002); §902.8 (West 1994).

Kansas: 9 to 11 months. Kan. Stat. Ann. §§21-3701(b)(2), 21-4704(a) (1995). Recidivist offender penalty not applicable. See §21-4504(e)(3).

Kentucky: 5 to 10 years. Ky. Rev. Stat. Ann. § 514.030(2) (Lexis Supp. 2002); §§ 532.060(2)(c), (d), 532.080(2), (5) (Lexis 1999).

Maine: less than one year. Me. Rev. Stat. Ann., Tit. 17-A, §353 (West 1983); §362(4)(B) (West Supp. 2000) (amended 2001); §1252(2)(D) (West 1983 and Supp. 2002). Recidivist offender penalty not applicable. See § 1252(4-A) (West Supp. 2000) (amended 2001).

Massachusetts: not more than five years. Mass. Gen. Laws, ch. 266, §30(1) (West 2000). Recidivist offender penalty not applicable. See ch. 279, §25 (West 1998); Commonwealth v. Hall, 397 Mass. 466, 468, 492 N. E. 2d 84, 85 (1986).

Minnesota: not more than five years. Minn. Stat. § 609.52, subd. 3(3)(a) (2002). Recidivist offender penalty not applicable. See §609.1095, subd. 2.

Mississippi: not more than five years. Miss. Code Ann. § 97-17-41(1)(a) (Lexis 1973-2000). Recidivist offender penalty not applicable. See §99-19-81.

Nebraska: not more than five years. Neb. Rev. Stat. § 28-105(1) (2000 Cum. Supp.); §28-518(2) (1995). Recidivist offender penalty not applicable. See §29-2221(1).

*56New Jersey: Extended term of between 5 to 10 years (instead of three to five years, N. J. Stat. Ann. § 2C:43-6 (1995)), § 2C:43-7(a)(4) (Supp. 2002), whether offense is treated as theft, § 2C:20-2(b)(2)(a), or shoplifting, §§2C:20-ll(b), (c)(2), because, even if Ewing’s felonies are regarded as one predicate crime, Ewing has been separately convicted and sentenced for at least one other crime for which at least a 6-month sentence was authorized, §2C:44-3(a); §2C:44-4(c) (1995).

New Mexico: 30 months. N. M. Stat. Ann. §30-16-20(B)(3) (1994); §31-18-15(A)(6) (2000); §31-18-17(B) (2000) (amended 2002).

New York: three to four years. N. Y. Penal Law § 70.06(3)(e) (West 1998); §155.30 (West 1999).

North Carolina: 4 to 25 months (with exact sentencing range dependent on details of offender’s criminal history). N. C. Gen. Stat. §§ 15A-1340.14, 15A-1340.17(c), (d), 14-72(a) (2001). Recidivist offender penalty not applicable. See §§ 14-7.1, 14-7.6.

North Dakota: not more than 10 years. N. D. Cent. Code § 12.1-23-05(2)(a) (1997); §§ 12.1-32-09(1), (2)(c) (1997) (amended 2001).

Ohio: 6 to 12 months. Ohio Rev. Code Ann. §§ 2913.02(B)(2), 2929.14(A)(5) (West Supp. 2002). No general recidivist statute.

Oregon: not more than five years. Ore. Rev. Stat. §161.605 (1997); Ore. Rev. Stat. Ann. §§ 164.055(l)(a), (3) (Supp. 1998). No general recidivist statute.

Pennsylvania: not more than five years (if no more than one prior theft was “retail theft”); otherwise, not more than seven years. Pa. Stat. Ann., Tit. 18, §§ 1103(3), 1104(1) (Pur-don 1998); §§ 3903(b), 3929(b)(l)(iii)-(iv) (Purdon Supp. 2002); § 3921 (Purdon 1983). Recidivist offender penalty not applicable. See 42 Pa. Cons. Stat. § 9714(a)(1) (1998).

*57Rhode Island: not more than 10 years. R. I. Gen. Laws § 11-41-5(a) (2002). Recidivist offender penalty not applicable. See § 12-19-21(a).

South Carolina: not more than five years. S. C. Code Ann. §§ 16-13-30, 16-13-110(B)(2) (West 2001 Cum. Supp.). Recidivist offender penalty not applicable. See § 17-25-45.

Tennessee: four to eight years. Tenn. Code Ann. §§39-14-105(3), 40-35-106(a)(1), (c), 40-35-112(b)(4) (1997).

Utah: not more than five years. Utah Code Ann. § 76-3-203(3) (1999) (amended 2000); § 76-6-412(l)(b)(i) (1999). Recidivist offender penalty not applicable. See §76-3-203.5 (Supp. 2002).

Washington: not more than 14 months (with exact sentencing range dependent on details of offender score), Wash. Rev. Code §§ 9A.56.040(l)(a), (2) (2000); §§9.94A.510(1), 9.94A.515, 9.94A.525 (2003 Supp. Pamphlet); maximum sentence of five years, §§9A.56.040(l)(a), (2), 9A.20.021(l)(c) (2000). Recidivist offender penalty not applicable. See §§9.94A.030(27), (31) (2000); § 9.94A.570 (2003 Supp. Pamphlet).

Wyoming: not more than 10 years. Wyo. Stat. Ann. § 6-3-404(a)(i) (Michie 2001). Recidivist offender penalty not applicable. See § 6-10-201(a).

B

In four other States, a Ewing-type offender could not have received a sentence of more than 15 years in prison:

Colorado: 4 to 12 years for “extraordinary aggravating circumstances” (e. g., defendant on parole for another felony at the time of commission of the triggering offense). Colo. Rev. Stat. §§ 18-l-105(l)(a)(V)(A), 18-l-105(9)(a)(II), 18-4-401(2)(c) (2002). Recidivist offender penalty not applicable. See §§ 16-13-101(f)(1.5), (2) (2001).

Maryland: not more than 15 years. Md. Ann. Code, Art. 27, § 342(f)(1) (1996) (repealed 2002). Recidivist offender penalty not applicable. See § 643B.

*58New Hampshire: not more than 15 years. N. H. Stat. Ann. §§637:ll(I)(a), 651:2(II)(a) (West Supp. 2002). Recidivist offender penalty not applicable. See §651:6(I)(c).

Wisconsin: not more than 11 years (at the time of Ewing’s offense). Wis. Stat. Ann. § 939.50(3)(e) (West Supp. 2002); §§939.62(l)(b), (2), 943.20(3)(b) (West 1996) (amended 2001). Wisconsin subsequently amended the relevant statutes so that a Ewing-type offender would only be eligible for a sentence of up to three years. See §§ 939.51(3)(a), 943.20(3)(a), 939.62(l)(a) (West Supp. 2003). And effective February 1, 2003, such an offender is eligible for a sentence of only up to two years. See §§ 939.51(3)(a), 943.20(3)(a), 939.62(l)(a).

C

In four additional States, a Ewing-type offender could not have been sentenced to more than’20 years in prison:

Arkansas: 3 to 20 years. Ark. Code Ann. § 5-36-103(b)(2)(A) (1997); §§ 5-4-501(a)(2)(D), (e)(1) (1997) (amended 2001). Eligible for parole after serving one-third of the sentence. § 5-4-501 (1997); § 16-93-608 (1987).

Missouri: not more than 20 years. Mo. Rev. Stat. §558.016(7X3) (2000); § 570.030(3)(1) (2000) (amended 2002). Eligible for parole after 15 years at the latest. §558.011(4)(l)(c).

Texas: 2 to 20 years. Tex. Penal Code Ann. §§ 12.33(a), 12.35(c)(2)(A) (1994); §§ 12.42(a)(3), 31.03(e)(4)(D) (Supp. 2003). Eligible for parole after serving one-fourth of sentence. Tex. Govt. Code Ann. §508.145(f) (Supp. 2003).

Virginia: statutory range of 1 to 20 years (or less than 12 months at the discretion of the jury or court following bench trial), Va. Code Ann. § 18.2-95 (Supp. 2002), but discretionary sentencing guideline ranges established by the Virginia Sentencing Commission, §§ 17.1-805, 19.2-298.01 (2000), with a maximum of 6 years, 3 months, to 15 years, 7 months, see Virginia Criminal Sentencing Commission, Virginia Sentencing Guidelines Manual, Larceny — Section C Recommenda*59tion Table (6th ed. 2002) (with petitioner likely falling within the discretionary guideline range of 2 years, 1 month, to 5 years, 3 months, see Brief for Petitioner 33, n. 25). Recidivist offender penalty not applicable. See § 19.2-297.1 (2000).

D

In nine other States, the law might make it legally possible to impose a sentence of 25 years or more upon a Ewing-type offender. But in five of those nine States,2 the offender would be parole-eligible before 25 years:

Alabama: “life or any term of not less than 20 years.” Ala. Code § 13A-5-9(c)(2) (Lexis Supp. 2002); §§ 13A-8-3(a), (c) (1994). Eligible for parole after the lesser of one-third of the sentence or 10 years. § 15-22-28(e) (1995).

Louisiana: Louisiana courts could have imposed a sentence of life without the possibility of parole at the time of Ewing’s offense. La. Stat. Ann. §§ 14:67.10(B)(1), 14:2(4), (13)(y) (West Supp. 2003); §§ 15:529.1(A)(l)(b)(ii) and (c)(iMii) (West 1992) (amended 2001). Petitioner argues that, despite the statutory authority to impose such a sentence, Louisiana courts would have carefully scrutinized his life sentence, as they had in other cases involving recidivists charged with a nonviolent crime. Brief for Petitioner 35-36, n. 29; see Brief for Families Against Mandatory Minimums as Amicus Curiae 24-25, and n. 21; State v. Hayes, 98-1526, p. 4 (La. App. 6/25/99), 739 So. 2d 301, 303-304 (holding that a life sentence was impermissibly excessive for a defendant convicted of theft of over $1,000, who had a prior robbery conviction). But see Brief for Respondent 45-46, n. 12 (contesting petitioner’s argument). Louisiana has amended its recidivist statute to require that the triggering offense be a violent felony, and that the offender have at least two prior violent felony convictions to be eligible for a life sentence. La. Stat. *60Ann. § 15:529.l(A)(l)(b)(ii) (West Supp. 2003). Under current law, a Ewing-type offender would face a sentence of 6% to 20 years. §§ 14:67.10(B)(1), 15:529.1(A)(b)(i).

Michigan: “imprisonment for life or for a lesser term,” Mich. Comp. Laws Ann. §769.12(l)(a) (West 2000) (instead of “not more than 15 years,” §769.12(l)(b), as petitioner contends, see Brief for Petitioner 34, n. 26; Brief for Families Against Mandatory Mínimums as Amicus Curiae 16-17, n. 15, 22-23, n. 20), because the triggering offense is “punishable upon a first conviction by imprisonment for a maximum term of 5 years or more,” § 769.12(l)(a) (West 2000). The larceny for which Ewing was convicted was, under Michigan law, “a felony punishable by imprisonment for not more than 5 years.” § 750.356(3)(a) (West Supp. 2002). Eligible for parole following minimum term set by sentencing judge. §769.12(4) (West 2000).

Montana: 5 to 100 years. Mont. Code Ann. §45-6-301(7)(b) (1999); §§46-18-501, 46-18-502(1) (2001). A Ewing-type offender would not have been subject to a minimum term of 10 years in prison (as the State suggests, Brief for Respondent 44) because Ewing does not meet the requirements of §46-18-502(2) (must be a “persistent felony offender,” as defined in §46-18-501, at the time of the offender’s previous felony conviction). See Reply Brief for Petitioner 18, n. 14. Eligible for parole after one-fourth of the term. §46-23-201(2).

Nevada: “life without the possibility of parole,” or “life with the possibility of parole [after serving] 10 years,” or “a definite term of 25 years, with eligibility for parole [after serving] 10 years.” Nev. Rev. Stat. §§207.010(l)(b)(l)-{3) (1995).

Oklahoma: not less than 20 years (at the time of Ewing’s offense). Okla. Stat., Tit. 21, § 51.1(B) (West Supp. 2000) (amended in 2001 to four years to life, § 51.1(C) (West 2001)); § 1704 (West 1991) (amended 2001). Eligible for parole after serving one-third of sentence. Tit. 57, § 332.7(B) (West *612001). Thus, assuming a sentence to a term of years of up to 100 years (as in Montana, see supra, at 60), parole eligibility could arise as late as after 33 years.

South Dakota: maximum penalty of life imprisonment, with no minimum term. S. D. Codified Laws §22-7-8 (1998); §22-30A-17(l) (Supp. 2002). Eligible for parole after serving one-half of sentence. §24-15-5(3) (1998). Thus, assuming a sentence to a term of years of up to 100 years (as in Montana, see supra, at 60), parole eligibility could arise as late as after 50 years.

Vermont: “up to and including life,” Vt. Stat. Ann., Tit. 13, § 11 (1998), or not more than 10 years, §2501; State v. Angelucci, 137 Vt. 272, 289-290, 405 A. 2d 33, 42 (1979) (court has discretion to sentence habitual offender to the sentence that is specified for grand larceny alone). Eligible for parole after six months. Tit. 28, § 501 (2000) (amended 2001).

West Virginia: Petitioner contends that he would only have been subject to a misdemeanor sentence of not more than 60 days for shoplifting, W. Va. Code §§61-3A-1, 61-3A-3(a)(2) (2000); Brief for Petitioner 31, n. 19, 33-34, n. 25. However, a Ewing-type offender could have been charged with grand larceny, see State ex rel. Chadwell v. Duncil, 196 W. Va. 643, 647-648, 474 S. E. 2d 573, 577-578 (1996) (prosecutor has discretion to charge defendant with either shoplifting or grand larceny), a felony punishable by imprisonment in the state penitentiary for 1 to 10 years (or, at the discretion of the trial court, not more than 1 year in jail). §61-3-13(a). Under West Virginia’s habitual offender statute, a felon “twice before convicted ... of a crime punishable by confinement in a penitentiary.. . . shall be sentenced to . . . life [imprisonment],” § 61 — 11—18(c), with parole eligibility after 15 years, § 62-12-13(c). Amicus curiae on behalf of petitioner notes that, in light of existing state-law precedents, West Virginia courts “would not countenance a sentence of life without the possibility of parole for 25 years for shoplifting golf clubs.” Brief for Families Against Mandatory Mínimums as Amicus *62Curiae 25-26 (citing State v. Barker, 186 W. Va. 73, 74-75, 410 S. E. 2d 712, 713-714 (1991) (per curiam); and State v. Deal, 178 W. Va. 142, 146-147, 358 S. E. 2d 226, 230-231 (1987)). But see Brief for Respondent 45, n. 11 (contesting that argument).

*52Justice Kennedy

delivered the opinion of the Court.

The issue before the Court is whether the Constitution permits a juvenile offender to be sentenced to life in prison *53without parole for a nonhomicide crime. The sentence was imposed by the State of Florida. Petitioner challenges the sentence under the Eighth Amendment’s Gruel and Unusual Punishments Clause, made applicable to the States by the Due Process Clause of the Fourteenth Amendment. Robinson v. California, 370 U. S. 660 (1962).

I

Petitioner is Terrance Jamar Graham. He was born on January 6, 1987. Graham’s parents were addicted to crack cocaine, and their drug use persisted in his early years. Graham was diagnosed with attention deficit hyperactivity disorder in elementary school. He began drinking alcohol and using tobacco at age 9 and smoked marijuana at age 13.

In July 2003, when Graham was age 16, he and three other school-age youths attempted to rob a barbeque restaurant in Jacksonville, Florida. One youth, who worked at the restaurant, left the back door unlocked just before closing time. Graham and another youth, wearing masks, entered through the unlocked door. Graham’s masked accomplice twice struck the restaurant manager in the back of the head with a metal bar. When the manager started yelling at the assailant and Graham, the two youths ran out and escaped in a car driven by the third accomplice. The restaurant manager required stitches for his head injury. No money was taken.

Graham was arrested for the robbery attempt. Under Florida law, it is within a prosecutor’s discretion whether to charge 16- and 17-year-olds as adults or juveniles for most felony crimes. Fla. Stat. § 985.227(l)(b) (2003) (subsequently renumbered at § 985.557(l)(b) (2007)). Graham’s prosecutor elected to charge Graham as an adult. The charges against Graham were armed burglary with assault or battery, a first-degree felony carrying a maximum penalty of life imprisonment without the possibility of parole, §§ 810.02(l)(b), (2)(a) (2003); and attempted armed robbery, a second-degree *54felony carrying a maximum penalty of 15 years' imprisonment, §§ 812.13(2)(b), 777.04(1), (4)(a), 775.082(3)(c).

On December 18, 2003, Graham pleaded guilty to both charges under a plea agreement. Graham wrote a letter to the trial court. After reciting “this is my first and last time getting in trouble,” he continued, “I’ve decided to turn my life around.” App. 379-380. Graham said, “I made a promise to God and myself that if I get a second chance, I’m going to do whatever it takes to get to the [National Football League].” Id., at 380.

The trial court accepted the plea agreement. The court withheld adjudication of guilt as to both charges and sentenced Graham to concurrent 3-year terms of probation. Graham was required to spend the first 12 months of his probation in the county jail, but he received credit for the time he had served awaiting trial, and was released on June 25, 2004.

Less than six months later, on the night of December 2, 2004, Graham again was arrested. The State’s case was as follows: Earlier that evening, Graham participated in a home invasion robbery. His two accomplices were Meigo Bailey and Kirkland Lawrence, both 20-year-old men. According to the State, at 7 p.m. that night, Graham, Bailey, and Lawrence knocked on the door of the home where Carlos Rodriguez lived. Graham, followed by Bailey and Lawrence, forcibly entered the home and held a pistol to Rodriguez's chest. For the next 30 minutes, the three held Rodriguez and another man, a friend of Rodriguez, at gunpoint while they ransacked the home searching for money. Before leaving, Graham and his accomplices barricaded Rodriguez and his Mend inside a closet.

The State further alleged that Graham, Bailey, and Lawrence, later the same evening, attempted a second robbery, during which Bailey was shot. Graham, who had borrowed his father’s car, drove Bailey and Lawrence to the hospital and left them there. As Graham drove away, a police ser*55geant signaled him to stop. Graham continued at a high speed but crashed into a telephone pole. He tried to flee on foot but was apprehended. Three handguns were found in his car.

When detectives interviewed Graham, he denied involvement in the crimes. He said he encountered Bailey and Lawrence only after Bailey had been shot. One of the detectives told Graham that the victims of the home invasion had identified him. He asked Graham, “Aside from the two robberies tonight how many more were you involved in?” Graham responded, “Two to three before tonight.” Id., at 160. The night that Graham allegedly committed the robbery, he was 34 days short of his 18th birthday.

On December 13, 2004, Graham’s probation officer filed with the trial court an affidavit asserting that Graham had violated the conditions of his probation by possessing a firearm, committing crimes, and associating with persons engaged in criminal activity. The trial court held hearings on Graham’s violations about a year later, in December 2005 and January 2006. The judge who presided was not the same judge who had accepted Graham’s guilty plea to the earlier offenses.

Graham maintained that he had no involvement in the home invasion robbery; but, even after the court underscored that the admission could expose him to a life sentence on the earlier charges, he admitted violating probation conditions by fleeing. The State presented evidence related to the home invasion, including testimony from the victims. The trial court noted that Graham, in admitting his attempt to avoid arrest, had acknowledged violating his probation. The court further found that Graham had violated his probation by committing a home invasion robbery, by possessing a firearm, and by associating with persons engaged in criminal activity.

The trial court held a sentencing hearing. Under Florida law the minimum sentence Graham could receive absent a *56downward departure by the judge was 5 years’ imprisonment. The maximum was life imprisonment. Graham’s attorney requested the minimum nondeparture sentence of 5 years. A presentence report prepared by the Florida Department of Corrections recommended that Graham receive an even lower sentence — at most 4 years’ imprisonment. The State recommended that Graham receive 30 years on the armed burglary count and 15 years on the attempted armed robbery count.

After hearing Graham’s testimony, the trial court explained the sentence it was about to pronounce:

“Mr. Graham, as I look back on your case, yours is really candidly a sad situation. You had, as far as I can tell, you have quite a family structure. You had a lot of people who wanted to try and help you get your life turned around including the court system, and you had a judge who took the step to try and give you direction through his probation order to give you a chance to get back onto track. And at the time you seemed through your letters that that is exactly what you wanted to do. And I don’t know why it is that you threw your life away. I don’t know why.

“But you did, and that is what is so sad about this today is that you have actually been given a chance to get through this, the original charge, which were very serious charges to begin with. .. . The attempted robbery with a weapon was a very serious charge.

“[I]n a very short period of time you were back before the Court on a violation of this probation, and then here you are two years later standing before me, literally the — facing a life sentence as to — up to life as to count 1 and up to 15 years as to count 2.

“And I don’t understand why you would be given such a great opportunity to do something with your life and *57why you would throw it away. The only thing that I can rationalize is that you decided that this is how you were going to lead your life and that there is nothing that we can do for you. And as the state pointed out, that this is an escalating pattern of criminal conduct on your part and that we can’t help you any further. We can’t do anything to deter you. This is the way you are going to lead your life, and I don’t know why you are going to. You’ve made that decision. I have no idea. But, evidently, that is what you decided to do.

“So then it becomes a focus, if I can’t do anything to help you, if I can’t do anything to get you back on the right path, then I have to start focusing on the community and trying to protect the community from your actions. And, unfortunately, that is where we are today is I don’t see where I can do anything to help you any further. You’ve evidently decided this is the direction you’re going to take in life, and it’s unfortunate that you made that choice.

“I have reviewed the statute. I don’t see where any further juvenile sanctions would be appropriate. I don’t see where any youthful offender sanctions would be appropriate. Given your escalating pattern of criminal conduct, it is apparent to the Court that you have decided that this is the way you are going to live your life and that the only thing I can do now is to try and protect the community from your actions.” Id., at 392-394.

The trial court found Graham guilty of the earlier armed burglary and attempted armed robbery charges. It sentenced him to the maximum sentence authorized by law on each charge: life imprisonment for the armed burglary and 15 years for the attempted armed robbery. Because Florida has abolished its parole system, see Fla. Stat. § 921.002(l)(e) (2003), a life sentence gives a defendant no possibility of release unless he is granted executive clemency.

*58Graham filed a motion in the trial court challenging his sentence under the Eighth Amendment. The motion was deemed denied after the trial court failed to rule on it within 60 days. The First District Court of Appeal of Florida affirmed, concluding that Graham’s sentence was not grossly disproportionate to his crimes. 982 So. 2d 43 (2008). The court took note of the seriousness of Graham’s offenses and their violent nature, as well as the fact that they “were not committed by a pre-teen, but a seventeen-year-old who was ultimately sentenced at the age of nineteen.” Id., at 52. The court concluded further that Graham was incapable of rehabilitation. Although Graham “was given an unheard of probationary sentence for a life felony, . . . wrote a letter expressing his remorse and promising to refrain from the commission of further crime, and . . . had a strong family structure to support him,” the court noted, he “rejected his second chance and chose to continue committing crimes at an escalating pace.” Ibid. The Florida Supreme Court denied review. 990 So. 2d 1058 (2008) (table).

We granted certiorari. 556 U. S. 1220 (2009).

II

The Eighth Amendment states: “Excessive bail shall not be required, nor excessive fines imposed, nor cruel and unusual punishments inflicted.” To determine whether a punishment is cruel and unusual, courts must look beyond historical conceptions to “ ‘the evolving standards of decency that mark the progress of a maturing society.’” Estelle v. Gamble, 429 U. S. 97, 102 (1976) (quoting Trop v. Dulles, 356 U. S. 86, 101 (1958) (plurality opinion)). “This is because ‘[t]he standard of extreme cruelty is not merely descriptive, but necessarily embodies a moral judgment. The standard itself remains the same, but its applicability must change as the basic mores of society change.’ ” Kennedy v. Louisiana, 554 U. S. 407, 419 (2008) (quoting Furman v. Georgia, 408 U. S. 238, 382 (1972) (Burger, C. J., dissenting)).

*59The Cruel and Unusual Punishments Clause prohibits the imposition of inherently barbaric punishments under all circumstances. See, e. g., Hope v. Pelzer, 536 U. S. 730 (2002). “[P]unishments of torture,” for example, “are forbidden.” Wilkerson v. Utah, 99 U. S. 130, 136 (1879). These cases underscore the essential principle that, under the Eighth Amendment, the State must respect the human attributes even of those who have committed serious crimes.

For the most part, however, the Court’s precedents consider punishments challenged not as inherently barbaric but as disproportionate to the crime. The concept of proportionality is central to the Eighth Amendment. Embodied in the Constitution’s ban on cruel and unusual punishments is the “precept of justice that punishment for crime should be graduated and proportioned to [the] offense.” Weems v. United States, 217 U. S. 349, 367 (1910).

The Court’s cases addressing the proportionality of sentences fall within two general classifications. The first involves challenges to the length of term-of-years sentences given all the circumstances in a particular case. The second comprises cases in which the Court implements the proportionality standard by certain categorical restrictions on the death penalty.

In the first classification the Court considers all of the circumstances of the case to determine whether the sentence is unconstitutionally excessive. Under this approach, the Court has held unconstitutional a life without parole sentence for the defendant’s seventh nonviolent felony, the crime of passing a worthless check. Solem v. Helm, 463 U. S. 277 (1983). In other cases, however, it has been difficult for the challenger to establish a lack of proportionality. A leading case is Harmelin v. Michigan, 501 U. S. 957 (1991), in which the offender was sentenced under state law to life without parole for possessing a large quantity of cocaine. A closely divided Court upheld the sentence. The controlling opinion concluded that the Eighth Amendment contains a “narrow *60proportionality principle,” that “does not require strict proportionality between crime and sentence” but rather “forbids only extreme sentences that are ‘grossly disproportionate’ to the crime.” Id., at 997, 1000-1001 (Kennedy, J., concurring in part and concurring in judgment). Again closely divided, the Court rejected a challenge to a sentence of 25 years to life for the theft of a few golf clubs under California’s so-called three-strikes recidivist sentencing scheme. Ewing v. California, 538 U. S. 11 (2003); see also Lockyer v. Andrade, 538 U. S. 63 (2003). The Court has also upheld a sentence of life with the possibility of parole for a defendant’s third nonviolent felony, the crime of obtaining money by false pretenses, Rummel v. Estelle, 445 U. S. 263 (1980), and a sentence of 40 years for possession of marijuana with intent to distribute and distribution of marijuana, Hutto v. Davis, 454 U. S. 370 (1982) (per curiam).

The controlling opinion in Harmelin explained its approach for determining whether a sentence for a term of years is grossly disproportionate for a particular defendant’s crime. A court must begin by comparing the gravity of the offense and the severity of the sentence. 501 U. S., at 1005 (opinion of Kennedy, J.). “[I]n the rare case in which [this] threshold comparison . . . leads to an inference of gross dis-proportionality” the court should then compare the defendant’s sentence with the sentences received by other offenders in the same jurisdiction and with the sentences imposed for the same crime in other jurisdictions. Ibid. If this comparative analysis “validate^] an initial judgment that [the] sentence is grossly disproportionate,” the sentence is cruel and unusual. Ibid.

The second classification of cases has used categorical rules to define Eighth Amendment standards. The previous cases in this classification involved the death penalty. The classification in turn consists of two subsets, one considering the nature of the offense, the other considering the characteristics of the offender. With respect to the nature of the *61offense, the Court has concluded that capital punishment is impermissible for nonhomicide crimes against individuals. Kennedy, 551 U. S., at 437-438; see also Enmund v. Florida, 458 U. S. 782 (1982); Coker v. Georgia, 433 U. S. 584 (1977). In cases turning on the characteristics of the offender, the Court has adopted categorical rules prohibiting the death penalty for defendants who committed their crimes before the age of 18, Roper v. Simmons, 543 U. S. 551 (2005), or whose intellectual functioning is in a low range, Atkins v. Virginia, 536 U. S. 304 (2002). See also Thompson v. Oklahoma, 487 U. S. 815 (1988).

In the cases adopting categorical rules the Court has taken the following approach. The Court first considers “objective indicia of society’s standards, as expressed in legislative enactments and state practice” to determine whether there is a national consensus against the sentencing practice at issue. Roper, supra, at 563. Next, guided by “the standards elaborated by controlling precedents and by the Court’s own understanding and interpretation of the Eighth Amendment’s text, history, meaning, and purpose,” Kennedy, 554 U. S., at 421, the Court must determine in the exercise of its own independent judgment whether the punishment in question violates the Constitution. Roper, supra, at 564.

The present case involves an issue the Court has not considered previously: a categorical challenge to a term-of-years sentence. The approach in cases such as Harmelin and Ewing is suited for considering a gross proportionality challenge to a particular defendant’s sentence, but here a sentencing practice itself is in question. This case implicates a particular type of sentence as it applies to an entire class of offenders who have committed a range of crimes. As a result, a threshold comparison between the severity of the penalty and the gravity of the crime does not advance the analysis. Here, in addressing the question presented, the appropriate analysis is the one used in cases that involved *62the categorical approach, specifically Atkins, Roper, and Kennedy.

Ill

A

The analysis begins with objective indicia of national consensus. “[T]he ‘clearest and most reliable objective evidence of contemporary values is the legislation enacted by the country's legislatures.”' Atkins, supra, at 312 (quoting Penry v. Lynaugh, 492 U. S. 302, 331 (1989)). Six jurisdictions do not allow life without parole sentences for any juvenile offenders. See Appendix, infra, Part III. Seven jurisdictions permit life without parole for juvenile offenders, but only for homicide crimes. Id., Part II. Thirty-seven States as well as the District of Columbia permit sentences of life without parole for a juvenile nonhomicide offender in some circumstances. Id., Part I. Federal law also allows for the possibility of life without parole for offenders as young as 13. See, e.g., 18 U.S.C. §§2241 (2006 ed. and Supp. II), 5032 (2006 ed.). Relying on this metric, the State and its amici argue that there is no national consensus against the sentencing practice at issue.

This argument is incomplete and unavailing. “There are measures of consensus other than legislation.” Kennedy, supra, at 433. Actual sentencing practices are an important part of the Court's inquiry into consensus. See Enmund, supra, at 794-796; Thompson, supra, at 831-832 (plurality opinion); Atkins, supra, at 316; Roper, supra, at 564-565; Kennedy, supra, at 433-434. Here, an examination of actual sentencing practices in jurisdictions where the sentence in question is permitted by statute discloses a consensus against its use. Although these statutory schemes contain no explicit prohibition on sentences of life without parole for juvenile nonhomieide offenders, those sentences are most infrequent. According to a recent study, nationwide there are only 109 juvenile offenders serving sentences of life without *63parole for nonhomicide offenses. See P. Annino, D. Rasmussen, & C. Rice, Juvenile Life without Parole for Non-Homicide Offenses: Florida Compared to Nation 2 (Sept. 14, 2009) (hereinafter Annino).

The State contends that this study’s tally is inaccurate because it does not count juvenile offenders who were convicted of both a homicide and a nonhomicide offense, even when the offender received a life without parole sentence for the nonhomicide. See Brief for Respondent 34; Tr. of Oral Arg. in Sullivan v. Florida, O. T. 2009, No. 08-7621, pp. 28-31. This distinction is unpersuasive. Juvenile offenders who committed both homicide and nonhomicide crimes present a different situation for a sentencing judge than juvenile offenders who committed no homicide. It is difficult to say that a defendant who receives a life sentence on a nonhomicide offense but who was at the same time convicted of homicide is not in some sense being punished in part for the homicide when the judge makes the sentencing determination. The instant case concerns only those juvenile offenders sentenced to life without parole solely for a nonhomicide offense.

Florida farther criticizes this study because the authors were unable to obtain complete information on some States and because the study was not peer reviewed. See Brief for Respondent 40. The State does not, however, provide any data of its own. Although in the first instance it is for the litigants to provide data to aid the Court, we have been able to supplement the study’s findings. The study’s authors were not able to obtain a definitive tally for Nevada, Utah, or Virginia. See Annino 11-13. Our research shows that Nevada has five juvenile nonhomicide offenders serving life without parole sentences, Utah has none, and Virginia has eight. See Letter from Alejandra Livingston, Offender Management Division, Nevada Dept, of Corrections, to Supreme Court Library (Mar. 26, 2010) (available in Clerk of Court’s ease file); Letter from Steve Gehrke, Utah Dept, of *64Corrections, to Supreme Court Library (Mar. 29, 2010) (same); Letter from Dr. Tama S. Celi, Virginia Dept, of Corrections, to Supreme Court Library (Mar. 30, 2010) (same). Finally, since the study was completed, a defendant in Oklahoma has apparently been sentenced to life without parole for a rape and stabbing he committed at the age of 16. See Stogsdill, Delaware County Teen Sentenced in Rape, Assault Case, Tulsa World, May 5, 2010, p. A12.

Thus, adding the individuals counted by the study to those we have been able to locate independently, there are 123 juvenile nonhomicide offenders serving life without parole sentences. A significant majority of those, 77 in total, are serving sentences imposed in Florida. Annino 2. The other 46 are imprisoned in just 10 States — California, Delaware, Iowa, Louisiana, Mississippi, Nebraska, Nevada, Oklahoma, South Carolina, and Virginia. Id., at 14; supra, at 63 and this page; Letter from Thomas P. Hoey, Dept, of Corrections, Government of the District of Columbia, to Supreme Court Library (Mar. 31, 2010) (available in Clerk of Court’s case file); Letter from Judith Simon Garrett, U. S. Dept, of Justice, Federal Bureau of Prisons (BOP), to Supreme Court Library (Apr. 9,2010) (same). Thus, only 11 jurisdictions nationwide in fact impose life without parole sentences on juvenile non-homicide offenders — and most of those do so quite rarely— while 26 States, the District of Columbia, and the Federal Government do not impose them despite statutory authorization.*

*65The numbers cited above reflect all current convicts in a jurisdiction’s penal system, regardless of when they were convicted. It becomes all the more clear how rare these sentences are, even within the jurisdictions that do sometimes impose them, when one considers that a juvenile sentenced to life without parole is likely to live in prison for decades. Thus, these statistics likely reflect nearly all juvenile nonhomicide offenders who have received a life without parole sentence stretching back many years. It is not certain that this opinion has identified every juvenile nonhomi-cide offender nationwide serving a life without parole sentence, for the statistics are not precise. The available data, nonetheless, are sufficient to demonstrate how rarely these sentences are imposed even if there are isolated cases that have not been included in the presentations of the parties or the analysis of the Court.

It must be acknowledged that in terms of absolute numbers juvenile life without parole sentences for nonhomicides are more common than the sentencing practices at issue in some of this Court’s other Eighth Amendment cases. See, e. g., Enmund, 458 U. S., at 794 (only six executions of non-triggerman felony murderers between 1954 and 1982), Atkins, 536 U. S., at 316 (only five executions of mentally retarded defendants in 13-year period). This contrast can be instructive, however, if attention is first given to the base number of certain types of offenses. For example, in the year 2007 (the most recent year for which statistics are available), a total of 13,480 persons, adult and juvenile, were arrested for homicide crimes. That same year, 57,600 juveniles were arrested for aggravated assault; 3,580 for forcible rape; 34,500 for robbery; 81,900 for burglary; 195,700 for drug offenses; and 7,200 for arson. See Dept, of Justice, Office of Juvenile Justice and Delinquency Prevention, Statistical Briefing Book, online at http://ojjdp.ncjrs.org/ojstatbb/ (as visited May 14, 2010, and available in Clerk of Court’s ease file). Although it is not certain how many of these numerous juvenile offenders were eligible for life without parole *66sentences, the comparison suggests that in proportion to the opportunities for its imposition, life without parole sentences for juveniles convicted of nonhomicide crimes is as rare as other sentencing practices found to be cruel and unusual.

The evidence of consensus is not undermined by the fact that many jurisdictions do not prohibit life without parole for juvenile nonhomieide offenders. The Court confronted a similar situation in Thompson, where a plurality concluded that the death penalty for offenders younger than 16 was unconstitutional. A number of States then allowed the juvenile death penalty if one considered the statutory scheme. As is the case here, those States authorized the transfer of some juvenile offenders to adult court; and at that point there was no statutory differentiation between adults and juveniles with respect to authorized penalties. The plurality concluded that the transfer laws show “that the States consider 15-year-olds to be old enough to be tried in criminal court for serious crimes (or too old to be dealt with effectively in juvenile court), but tells us nothing about the judgment these States have made regarding the appropriate punishment for such youthful offenders.” 487 U. S., at 826, n. 24. Justice O’Connor, concurring in the judgment, took a similar view. Id., at 850 (“When a legislature provides for some 15-year-olds to be processed through the adult criminal justice system, and capital punishment is available for adults in that jurisdiction, the death penalty becomes at least theoretically applicable to such defendants. . . . [HJowever, it does not necessarily follow that the legislatures in those jurisdictions have deliberately concluded that it would be appropriate”).

The same reasoning obtains here. Many States have chosen to move away from juvenile court systems and to allow juveniles to be transferred to, or charged directly in, adult court under certain circumstances. Once in adult court, a juvenile offender may receive the same sentence as would be given to an adult offender, including a life without parole *67sentence. But the fact that transfer and direct charging laws make life without parole possible for some juvenile non-homicide offenders does not justify a judgment that many States intended to subject such offenders to life without parole sentences.

For example, under Florida law a child of any age can be prosecuted as an adult for certain crimes and can be sentenced to life without parole. The State acknowledged at oral argument that even a 5-year-old, theoretically, could receive such a sentence under the letter of the law. See Tr. of Oral Arg. 36-87. All would concede this to be unrealistic, but the example underscores that the statutory eligibility of a juvenile offender for life without parole does not indicate that the penalty has been endorsed through deliberate, express, and full legislative consideration. Similarly, the many States that allow life without parole for juvenile nonhomicide offenders but do not impose the punishment should not be treated as if they have expressed the view that the sentence is appropriate. The sentencing practice now under consideration is exceedingly rare. And “it is fair to say that a national consensus has developed against it.” Atkins, supra, at 316.

B

Community consensus, while “entitled to great weight,” is not itself determinative of whether a punishment is cruel and unusual. Kennedy, 554 U. S., at 434 (internal quotation marks omitted). In accordance with the constitutional design, “the task of interpreting the Eighth Amendment remains our responsibility.” Roper, 543 U. S., at 575. The judicial exercise of independent judgment requires consideration of the culpability of the offenders at issue in light of their crimes and characteristics, along with the severity of the punishment in question. Id., at 568; Kennedy, supra, at 438; cf. Solem, 463 U. S., at 292. In this inquiry the Court also considers whether the challenged sentencing practice serves legitimate penological goals. Kennedy, supra, at *68441-446; Roper, supra, at 571-572; Atkins, 536 U. S., at 318-320.

Roper established that because juveniles have lessened culpability they are less deserving of the most severe punishments. 543 U. S., at 569. As compared to adults, juveniles have a “‘lack of maturity and an underdeveloped sense of responsibility’ ”; they “are more vulnerable or susceptible to negative influences and outside pressures, including peer pressure”; and their characters are “not as well formed.” Id., at 569-570. These salient characteristics mean that “[i]t is difficult even for expert psychologists to differentiate between the juvenile offender whose crime reflects unfortunate yet transient immaturity, and the rare juvenile offender whose crime reflects irreparable corruption.” Id., at 573. Accordingly, “juvenile offenders cannot with reliability be classified among the worst offenders.” Id., at 569. A juvenile is not absolved of responsibility for his actions, but his transgression “is not as morally reprehensible as that of an adult.” Thompson, supra, at 835 (plurality opinion).

No recent data provide reason to reconsider the Court’s observations in Roper about the nature of juveniles. As petitioner’s amici point out, developments in psychology and brain science continue to show fundamental differences between juvenile and adult minds. For example, parts of the brain involved in behavior control continue to mature through late adolescence. See Brief for American Medical Association et al. 16-24; Brief for American Psychological Association et al. 22-27. Juveniles are more capable of change than are adults, and their actions are less likely to be evidence of “irretrievably depraved character” than are the actions of adults. Roper, 543 U. S., at 570. It remains true that “[f]rom a moral standpoint it would be misguided to equate the failings of a minor with those of an adult, for a greater possibility exists that a minor’s character deficiencies will be reformed.” Ibid. These matters relate to the status of the offenders in question; and it is relevant to consider *69next the nature of the offenses to which this harsh penalty-might apply.

The Court has recognized that defendants who do not kill, intend to kill, or foresee that life will be taken are categorically less deserving of the most serious forms of punishment than are murderers. Kennedy, supra; Enmund, 458 U. S. 782; Tison v. Arizona, 481 U. S. 137 (1987); Coker, 433 U. S. 584. There is a line “between homicide and other serious violent offenses against the individual.” Kennedy, 554 U. S., at 438. Serious nonhomicide crimes “may be devastating in their harm . . . but ‘in terms of moral depravity and of the injury to the person and to the public/ . . . they cannot be compared to murder in their ‘severity and irrevocability.’” Ibid, (quoting Coker, 433 U. S., at 598 (plurality opinion)). This is because “[l]ife is over for the victim of the murderer,” but for the victim of even a very serious nonhomicide crime, “life ... is not over and normally is not beyond repair.” Ibid, (plurality opinion). Although an offense like robbery or rape is “a serious crime deserving serious punishment,” Enmund, supra, at 797, those crimes differ from homicide crimes in a moral sense.

It follows that, when compared to an adult murderer, a juvenile offender who did not kill or intend to kill has a twice diminished moral culpability. The age of the offender and the nature of the crime each bear on the analysis.

As for the punishment, life without parole is “the second most severe penalty permitted by law.” Harmelin, 501 U. S., at 1001 (Kennedy, J., concurring in part and concurring in judgment). It is true that a death sentence is “unique in its severity and irrevocability,” Gregg v. Georgia, 428 U. S. 153, 187 (1976) (joint opinion of Stewart, Powell, and Stevens, JJ.); yet life without parole sentences share some characteristics with death sentences that are shared by no other sentences. The State does not execute the offender sentenced to life without parole, but the sentence alters the offender’s life by a forfeiture that is irrevocable. It deprives *70the convict of the most basic liberties without giving hope of restoration, except perhaps by executive clemency — the remote possibility of which does not mitigate the harshness of the sentence. Solem, 463 U. S., at 300-301. As one court observed in overturning a life without parole sentence for a juvenile defendant, this sentence “means denial of hope; it means that good behavior and character improvement are immaterial; it means that whatever the future might hold in store for the mind and spirit of [the convict], he will remain in prison for the rest of his days.” Naovarath v. State, 105 Nev. 525, 526, 779 P. 2d 944 (1989).

The Court has recognized the severity of sentences that deny convicts the possibility of parole. In Rummel, 445 U. S. 263, the Court rejected an Eighth Amendment challenge to a life sentence for a defendant’s third nonviolent felony but stressed that the sentence gave the defendant the possibility of parole. Noting that “parole is an established variation on imprisonment of convicted criminals,” it was evident that an analysis of the petitioner’s sentence “could hardly ignore the possibility that he will not actually be imprisoned for the rest of his life.” Id., at 280-281 (internal quotation marks omitted). And in Solem, the only previous case striking down a sentence for a term of years as grossly disproportionate, the defendant’s sentence was deemed “far more severe than the life sentence we considered in Rum-mel,” because it did not give the defendant the possibility of parole. 463 U. S., at 297.

Life without parole is an especially harsh punishment for a juvenile. Under this sentence a juvenile offender will on average serve more years and a greater percentage of his life in prison than an adult offender. A 16-year-old and a 75-year-old each sentenced to life without parole receive the same punishment in name only. See Roper, supra, at 572; cf. Harmelin, supra, at 996 (“In some cases... there will be negligible difference between life without parole and other sentences of imprisonment — for example,... a lengthy term *71sentence without eligibility for parole, given to a 65-year-old man”). This reality cannot be ignored.

The penological justifications for the sentencing practice are also relevant to the analysis. Kennedy, supra, at 420; Roper, 543 U. S., at 571-572; Atkins, 536 U. S., at 318-320. Criminal punishment can have different goals, and choosing among them is within a legislature’s discretion. See Har-melin, supra, at 999 (opinion of Kennedy, J.) (“[T]he Eighth Amendment does not mandate adoption of any one penological theory”). It does not follow, however, that the purposes and effects of penal sanctions are irrelevant to the determination of Eighth Amendment restrictions. A sentence lacking any legitimate penological justification is by its nature disproportionate to the offense. With respect to life without parole for juvenile nonhomicide offenders, none of the goals of penal sanctions that have been recognized as legitimate— retribution, deterrence, incapacitation, and rehabilitation, see Ewing, 538 U. S., at 25 (plurality opinion) — provides an adequate justification.

Retribution is a legitimate reason to punish, but it cannot support the sentence at issue here. Society is entitled to impose severe sanctions on a juvenile nonhomicide offender to express its condemnation of the crime and to seek restoration of the moral imbalance caused by the offense. But “[t]he heart of the retribution rationale is that a criminal sentence must be directly related to the personal culpability of the criminal offender.” Tison, supra, at 149. And as Roper observed, “[wjhether viewed as an attempt to express the community’s moral outrage or as an attempt to right the balance for the wrong to the victim, the case for retribution is not as strong with a minor as with an adult.” 543 U. S., at 571. The case becomes even weaker with respect to a juvenile who did not commit homicide. Roper found that “ [rjetribution is not proportional if the law’s most severe penalty is imposed” on the juvenile murderer. Ibid,. The considerations underlying that holding support as well the con-*72elusion that retribution does not justify imposing the second most severe penalty on the less culpable juvenile nonhomi-cide offender.

Deterrence does not suffice to justify the sentence either. Roper noted that “the same characteristics that render juveniles less culpable than adults suggest... that juveniles will be less susceptible to deterrence.” Ibid. Because juveniles' “lack of maturity and an underdeveloped sense of responsibility . . . often result in impetuous and ill-considered actions and decisions,” Johnson v. Texas, 509 U. S. 350, 367 (1993), they are less likely to take a possible punishment into consideration when making decisions. This is particularly so when that punishment is rarely imposed. That the sentence deters in a few cases is perhaps plausible, but “[t]his argument does not overcome other objections.” Kennedy, 554 U. S., at 441. Even if the punishment has some connection to a valid penological goal, it must be shown that the punishment is not grossly disproportionate in light of the justification offered. Here, in light of juvenile nonhomicide offenders' diminished moral responsibility, any limited deterrent effect provided by life without parole is not enough to justify the sentence.

Incapacitation, a third legitimate reason for imprisonment, does not justify the life without parole sentence in question here. Recidivism is a serious risk to public safety, and so incapacitation is an important goal. See Ewing, supra, at 26 (plurality opinion) (statistics show 67 percent of former inmates released from state prisons are charged with at least one serious new crime within three years). But while incapacitation may be a legitimate penological goal sufficient to justify life without parole in other contexts, it is inadequate to justify that punishment for juveniles who did not commit homicide. To justify life without parole on the assumption that the juvenile offender forever will be a danger to society requires the sentencer to make a judgment that the juvenile is incorrigible. The characteristics of juveniles make that *73judgment questionable. “It is difficult even for expert psychologists to differentiate between the juvenile offender whose crime reflects unfortunate yet transient immaturity, and the rare juvenile offender whose crime reflects irreparable corruption.” Roper, supra, at 573. As one court concluded in a challenge to a life without parole sentence for a 14-year-old, “incorrigibility is inconsistent with youth.” Workman v. Commonwealth, 429 S. W. 2d 374, 378 (Ky. App. 1968).

Here one cannot dispute that this defendant posed an immediate risk, for he had committed, we can assume, serious crimes early in his term of supervised release and despite his own assurances of reform. Graham deserved to be separated from society for some time in order to prevent what the trial court described as an “escalating pattern of criminal conduct,” App. 394, but it does not follow that he would be a risk to society for the rest of his life. Even if the State's judgment that Graham was incorrigible were later corroborated by prison misbehavior or failure to mature, the sentence was still disproportionate because that judgment was made at the outset. A life without parole sentence improperly denies the juvenile offender a chance to demonstrate growth and maturity. Incapacitation cannot override all other considerations, lest the Eighth Amendment’s rule against disproportionate sentences be a nullity.

Finally there is rehabilitation, a penological goal that forms the basis of parole systems. See Solem, 463 U. S., at 300; Mistretta v. United States, 488 U. S. 361, 363 (1989). The concept of rehabilitation is imprecise; and its utility and proper implementation are the subject of a substantial, dynamic field of inquiry and dialogue. See, e. g., Cullen & Gen-dreau, Assessing Correctional Rehabilitation: Policy, Practice, and Prospects, 3 Criminal Justice 2000, pp. 119-133 (2000) (describing scholarly debates regarding the effectiveness of rehabilitation over the last several decades). It is *74for legislatures to determine what rehabilitative techniques are appropriate and effective.

A sentence of life imprisonment without parole, however, cannot be justified by the goal of rehabilitation. The penalty forswears altogether the rehabilitative ideal. By denying the defendant the right to reenter the community, the State makes an irrevocable judgment about that person’s value and place in society. This judgment is not appropriate in light of a juvenile nonhomicide offender’s capacity for change and limited moral culpability. A State’s rejection of rehabilitation, moreover, goes beyond a mere expressive judgment. As one amicus notes, defendants serving life without parole sentences are often denied access to vocational training and other rehabilitative services that are available to other inmates. See Brief for Sentencing Project 11-13. For juvenile offenders, who are most in need of and receptive to rehabilitation, see Brief for J. Lawrence Aber et al. as Amici Curiae 28-31 (hereinafter Aber Brief), the absence of rehabilitative opportunities or treatment makes the disproportionality of the sentence all the more evident.

In sum, penological theory is not adequate to justify life without parole for juvenile nonhomicide offenders. This determination; the limited culpability of juvenile nonhomicide offenders; and the severity of life without parole sentences all lead to the conclusion that the sentencing practice under consideration is cruel and unusual. This Court now holds that for a juvenile offender who did not commit homicide the Eighth Amendment forbids the sentence of life without parole. This clear line is necessary to prevent the possibility that life without parole sentences will be imposed on juvenile nonhomicide offenders who are not sufficiently culpable to merit that punishment. Because “[t]he age of 18 is the point where society draws the line for many purposes between childhood and adulthood,” those who were below that age when the offense was committed may not be sentenced to *75life without parole for a nonhomicide crime. Roper, 543 U. S., at 574.

A State is not required to guarantee eventual freedom to a juvenile offender convicted of a nonhomicide crime. What the State must do, however, is give defendants like Graham some meaningful opportunity to obtain release based on demonstrated maturity and rehabilitation. It is for the State, in the first instance, to explore the means and mechanisms for compliance. It bears emphasis, however, that while the Eighth Amendment prohibits a State from imposing a life without parole sentence on a juvenile nonhomicide offender, it does not require the State to release that offender during his natural life. Those who commit truly horrifying crimes as juveniles may turn out to be irredeemable, and thus deserving of incarceration for the duration of their lives. The Eighth Amendment does not foreclose the possibility that persons convicted of nonhomicide crimes committed before adulthood will remain behind bars for life. It does prohibit States from making the judgment at the outset that those offenders never will be fit to reenter society.

C

Categorical rules tend to be imperfect, but one is necessary here. Two alternative approaches are not adequate to address the relevant constitutional concerns. First, the State argues that the laws of Florida and other States governing criminal procedure take sufficient account of the age of a juvenile offender. Here, Florida notes that under its law prosecutors are required to charge 16- and 17-year-old offenders as adults only for certain serious felonies; that prosecutors have discretion to charge those offenders as adults for other felonies; and that prosecutors may not charge nonrecidivist 16- and 17-year-old offenders as adults for misdemeanors. Brief for Respondent 54 (citing Fla. Stat. § 985.227 (2003)). The State also stresses that “in only the narrowest of circumstances” does Florida law impose no *76age limit whatsoever for prosecuting juveniles in adult court. Brief for Respondent 54.

Florida is correct to say that state laws requiring consideration of a defendant’s age in charging decisions are salutary. An offender’s age is relevant to the Eighth Amendment, and criminal procedure laws that fail to take defendants’ youthfulness into account at all would be flawed. Florida, like other States, has made substantial efforts to enact comprehensive rules governing the treatment of youthful offenders by its criminal justice system. See generally Fla. Stat. § 958 et seq. (2007).

The provisions the State notes are, nonetheless, by themselves insufficient to address the constitutional concerns at issue. Nothing in Florida’s laws prevents its courts from sentencing a juvenile nonhomicide offender to life without parole based on a subjective judgment that the defendant’s crimes demonstrate an “irretrievably depraved character.” Roper, supra, at 570. This is inconsistent with the Eighth Amendment. Specific cases are illustrative. In Graham’s case the sentencing judge decided to impose life without parole — a sentence greater than that requested by the prosecutor — for Graham’s armed burglary conviction. The judge did so because he concluded that Graham was incorrigible: “[Y]ou decided that this is how you were going to lead your life and that there is nothing that we can do for you. . . . We can’t do anything to deter you.” App. 394.

Another example comes from Sullivan v. Florida, No. 08-7621. Sullivan was argued the same day as this case, but the Court has now dismissed the writ of certiorari in Sullivan as improvidently granted. Post, p. 181. The facts, however, demonstrate the flaws of Florida’s system. The petitioner, Joe Sullivan, was prosecuted as an adult for a sexual assault committed when he was 13 years old. Noting Sullivan’s past encounters with the law, the sentencing judge concluded that, although Sullivan had been “given opportunity after opportunity to upright himself and take advantage *77of the second and third chances he’s been given,” he had demonstrated himself to be unwilling to follow the law and needed to be kept away from society for the duration of his life. Brief for Respondent in Sullivan v. Florida, O. T. 2009, No. 08-7621, p. 6. The judge sentenced Sullivan to life without parole. As these examples make clear, existing state laws, allowing the imposition of these sentences based only on a discretionary, subjective judgment by a judge or jury that the offender is irredeemably depraved, are insufficient to prevent the possibility that the offender will receive a life without parole sentence for which he or she lacks the moral culpability.

Another possible approach would be to hold that the Eighth Amendment requires courts to take the offender’s age into consideration as part of a case-specific gross dispro-portionality inquiry, weighing it against the seriousness of the crime. This approach would allow courts to account for factual differences between cases and to impose life without parole sentences for particularly heinous crimes. Few, perhaps no, judicial responsibilities are more difficult than sentencing. The task is usually undertaken by trial judges who seek with diligence and professionalism to take account of the human existence of the offender and the just demands of a wronged society.

The case-by-case approach to sentencing must, however, be confined by some boundaries. The dilemma of juvenile sentencing demonstrates this. For even if we were to assume that some juvenile nonhomicide offenders might have “sufficient psychological maturity, and at the same time dem-onstrat[e] sufficient depravity,” Roper, 543 U. S., at 572, to merit a life without parole sentence, it does not follow that courts taking a case-by-case proportionality approach could with sufficient accuracy distinguish the few incorrigible juvenile offenders from the many that have the capacity for change. Roper rejected the argument that the Eighth Amendment required only that juries be told they must con*78sider the defendant’s age as a mitigating factor in sentencing. The Court concluded that an “unacceptable likelihood exists that the brutality or cold-blooded nature of any particular crime would overpower mitigating arguments based on youth as a matter of course, even where the juvenile offender’s objective immaturity, vulnerability, and lack of true depravity should require a sentence less severe than death.” Id., at 573. Here, as with the death penalty, “[t]he differences between juvenile and adult offenders are too marked and well understood to risk allowing a youthful person to receive” a sentence of life without parole for a nonhomicide crime “despite insufficient culpability.” Id., at 572-573.

Another problem with a case-by-case approach is that it does not take account of special difficulties encountered by counsel in juvenile representation. As some amici note, the features that distinguish juveniles from adults also put them at a significant disadvantage in criminal proceedings. Juveniles mistrust adults and have limited understandings of the criminal justice system and the roles of the institutional actors within it. They are less likely than adults to work effectively with their lawyers to aid in their defense. Brief for NAACP Legal Defense & Educational Fund et al. as Amici Curiae 7-12; Henning, Loyalty, Paternalism, and Rights: Client Counseling Theory and the Role of Child’s Counsel in Delinquency Cases, 81 Notre Dame L. Rev. 245, 272-273 (2005). Difficulty in weighing long-term consequences; a corresponding impulsiveness; and reluctance to trust defense counsel, seen as part of the adult world a rebellious youth rejects, all can lead to poor decisions by one charged with a juvenile offense. Aber Brief 35. These factors are likely to impair the quality of a juvenile defendant’s representation. Cf. Atkins, 536 U. S., at 320 (“Mentally retarded defendants may be less able to give meaningful assistance to their counsel”). A categorical rule avoids the risk that, as a result of these difficulties, a court or jury will *79erroneously conclude that a particular juvenile is sufficiently culpable to deserve life without parole for a nonhomicide.

Finally, a categorical rule gives all juvenile nonhomicide offenders a chance to demonstrate maturity and reform. The juvenile should not be deprived of the opportunity to achieve maturity of judgment and self-recognition of human worth and potential. In Roper, that deprivation resulted from an execution that brought life to its end. Here, though by a different dynamic, the same concerns apply. Life in prison without the possibility of parole gives no chance for fulfillment outside prison walls, no chance for reconciliation with society, no hope. Maturity can lead to that considered reflection which is the foundation for remorse, renewal, and rehabilitation. A young person who knows that he or she has no chance to leave prison before life’s end has little incentive to become a responsible individual. In some prisons, moreover, the system itself becomes complieit in the lack of development. As noted above, see supra, at 74, it is the policy in some prisons to withhold counseling, education, and rehabilitation programs for those who are ineligible for parole consideration. A categorical rule against life without parole for juvenile nonhomicide offenders avoids the perverse consequence in which the lack of maturity that led to an offender’s crime is reinforced by the prison term.

Terrance Graham’s sentence guarantees he will die in prison without any meaningful opportunity to obtain release, no matter what he might do to demonstrate that the bad acts he committed as a teenager are not representative of his true character, even if he spends the next half century attempting to atone for his crimes and learn from his mistakes. The State has denied him any chance to later demonstrate that he is fit to rejoin society based solely on a nonhomicide crime that he committed while he was a child in the eyes of the law. This the Eighth Amendment does not permit.

*80D

There is support for our conclusion in the fact that, in continuing to impose life without parole sentences on juveniles who did not commit homicide, the United States adheres to a sentencing practice rejected the world over. This observation does not control our decision. The judgments of other nations and the international community are not dis-positive as to the meaning of the Eighth Amendment. But “‘[t]he climate of international opinion concerning the acceptability of a particular punishment’ ” is also “ ‘not irrelevant.’” Enmund, 458 U. S., at 796, n. 22. The Court has looked beyond our Nation’s borders for support for its independent conclusion that a particular punishment is cruel and unusual. See, e. g., Roper, 543 U. S., at 575-578; Atkins, supra, at 316-318, n. 21; Thompson, 487 U. S., at 830 (plurality opinion); Enmund, supra, at 796-797, n. 22; Coker, 433 U. S., at 596, n. 10 (same); Trop, 356 U. S., at 102-103 (same).

Today we continue that longstanding practice in noting the global consensus against the sentencing practice in question. A recent study concluded that only 11 nations authorize life without parole for juvenile offenders under any circumstances; and only 2 of them, the United States and Israel, ever impose the punishment in practice. See M. Leighton & C. de la Vega, Sentencing Our Children To Die in Prison: Global Law and Practice 4 (2007). An updated version of the study concluded that Israel’s “laws allow for parole review of juvenile offenders serving life terms,” but expressed reservations about how that parole review is implemented. De la Vega & Leighton, Sentencing Our Children To Die in Prison: Global Law and Practice, 42 U. S. F. L. Rev. 983, 1002-1003 (2008). But even if Israel is counted as allowing life without parole for juvenile offenders, that nation does not appear to impose that sentence for nonhomicide crimes; all of the seven Israeli prisoners whom commentators have identified as serving life sentences for juvenile crimes were *81convicted of homicide or attempted homicide. See Amnesty International, Human Rights Watch, The Rest of Their Lives; Life Without Parole for Child Offenders in the United States 106, n. 322 (2005); Memorandum and Attachment from Ruth Levush, Law Library of Congress, to Supreme Court Library (Feb. 16, 2010) (available in Clerk of Court’s case file).

Thus, as petitioner contends and respondent does not contest, the United States is the only Nation that imposes life without parole sentences on juvenile nonhomicide offenders. We also note, as petitioner and his amici emphasize, that Article 37(a) of the United Nations Convention on the Rights of the Child, Nov. 20, 1989, 1577 U. N. T. S. 3 (entered into force Sept. 2, 1990), ratified by every nation except the United States and Somalia, prohibits the imposition of “life imprisonment without possibility of release ... for offences committed by persons below eighteen years of age.” Brief for Petitioner 66; Brief for Amnesty International et al. 15-17. As we concluded in Roper with respect to the juvenile death penalty, “the United States now stands alone in a world that has turned its face against” life without parole for juvenile nonhomicide offenders. 543 U. S., at 577.

The State’s amici stress that no international legal agreement that is binding on the United States prohibits life without parole for juvenile offenders and thus urge us to ignore the international consensus. See Brief for Solidarity Center for Law and Justice et al. 14-16; Brief for Sixteen Members of United States House of Representatives 40-43. These arguments miss the mark. The question before us is not whether international law prohibits the United States from imposing the sentence at issue in this case. The question is whether that punishment is cruel and unusual. In that inquiry, “the overwhelming weight of international opinion against” life without parole for nonhomicide offenses committed by juveniles “provide[s] respected and significant confirmation for our own conclusions.” Roper, supra, at 578.

*82The debate between petitioner’s and respondent’s amici over whether there is a binding^ cogens norm against this sentencing practice is likewise of no import. See Brief for Amnesty International 10-23; Brief for Sixteen Members of United States House of Representatives 4-40. The Court has treated the laws and practices of other nations and international agreements as relevant to the Eighth Amendment not because those norms are binding or controlling but because the judgment of the world’s nations that a particular sentencing practice is inconsistent with basic principles of decency demonstrates that the Court’s rationale has respected reasoning to support it.

The Constitution prohibits the imposition of a life without parole sentence on a juvenile offender who did not commit homicide. A State need not guarantee the offender eventual release, but if it imposes a sentence of life it must provide him or her with some realistic opportunity to obtain release before the end of that term. The judgment of the First District Court of Appeal of Florida is reversed, and the case is remanded for further proceedings not inconsistent with this opinion.

It is so ordered.

APPENDIX

I. JURISDICTIONS THAT PERMIT LIFE WITHOUT PAROLE FOR JUVENILE NONHOMICIDE OFFENDERS

Alabama Ala. Code § 12-15-20S (Supp. 2009); §§ 13A-3-3,13A-

5-9(c), 13A-6-61 (2005); §13A-7-5 (Supp. 2009) Arizona Ariz. Rev. Stat. Ann. §§ 13-501, 13-1423 (West 2010)

Arkansas Ark. Code §9-27-318(b) (2009); §5-4-501(c) (Supp. 2009)

California Cal. Penal Code Ann. § 667.7(a)(2) (West 1999); § 1170.17 (West 2004)

Del. Code Ann., Tit. 10, §1010 (Supp. 2008); id., Tit. 11, § 773(e) (2003) Delaware

*83District of Columbia D. C. Code § 16-2307 (2009 Supp. Pamphlet); §22-3020 (Supp. 2007)

Florida Fla. Stat. §§810.02, 921.002(l)(e), 985.557 (2007)

Georgia Georgia Code Ann. §15-11-30.2 (2008); §16-6-l(b) (2007)

Idaho Idaho Code §18-6503 (Lexis 2005); §§19-2513, 20-509 (Lexis Supp. 2009)

Illinois 111. Comp. Stat., ch. 705, §§405/5-805,405/5-130 (West 2008); id., ch. 720, § 5/12-13(b)(3) (West 2008); id., ch. 730, §5/3-3-3(d) (West 2008)

Indiana Ind. Code §§31-30-3-6(1), 35-50-2-8.5(a) (West 2004)

Iowa Iowa Code §§232.45(6), 709.2, 902.1 (2009)

Louisiana La. Child. Code Ann., Arts. 305, 857(A), (B) (West Supp. 2010); La Rev. Stat. Ann. § 14:44 (West 2007)

Maryland Md. Cts. & Jud. Proc. Code Ann. §§ 3-8A-03(d)(l), 3-8A-06(a)(2) (Lexis 2006); Md. Crim. Law Code Ann. §§ 3-303(d)(2), (3) (Lexis Supp. 2009)

Michigan Mich. Comp. Laws Ann. §712A.4 (West 2002); § 750.520b(2)(c) (West Supp. 2009); §769.1 (West 2000)

Minnesota Minn. Stat. §§260B.125(1), 609.3455(2) (2008)

Mississippi Miss. Code Ann. §43-21-157 (2009); §§97-3-53, 99-19-81 (2007); §99-19-83 (2006)

Missouri Mo. Rev. Stat. §§211.071, 558.018 (2000)

Nebraska Neb. Rev. Stat. §§28-105,28-416(8)(a), 29-2204(1), (3), 43-247, 43-276 (2008)

Nevada Nev. Rev. Stat. §§62B.330, 200.366 (2009)

New Hampshire N. H. Rev. Stat. Ann. §§169-B:24, 628:1 (2007); §§ 632-A:2, 651:6 (Supp. 2009)

New York N. Y. Penal Law Ann. §§30.00, 60.06 (West 2009); §490.55 (West 2008)

North Carolina N. C. Gen. Stat. Ann. §§7B-2200, 15A-1340.16B(a) (Lexis 2009)

North Dakota N. D. Cent. Code Ann. §12.1-04-01 (Lexis 1997); § 12.1-20-03 (Lexis Supp. 2009); § 12.1-32-01 (Lexis 1997)

Ohio Ohio Rev. Code Ann. §2152.10 (Lexis 2007); §2907.02 (Lexis 2006); § 2971.03(A)(2) (2010 Lexis Supp. Pamphlet)

Oklahoma Okla Stat., Tit. 10A, §§2-5-204, 2-5-205, 2-5-206 (2009 West Supp.); id., Tit. 21, § 1115 (2007 West Supp.)

Oregon Ore. Rev. Stat. §§137.707, 137.719(1) (2009)

Pennsylvania 42 Pa. Cons. Stat. § 6355(a) (2000); 18 id., § 3121(e)(2) (2008) ; 61 id., § 6137(a) (2009)

*84Rhode Island R. I. Gen. Laws §§ 14-1-7,14-1-7.1, 11-47-3.2 (Lexis 2002)

South Carolina S. C. Code Ann. §63-19-1210 (2008 Supp. Pamphlet); § 16-11-311(B) (Westlaw 2009)

South Dakota S. D. Codified Laws §26-11-3.1 (Supp. 2009); §26-11-4 (2004); §§22-3-1, 22-6-1(2), (3) (2006); §24-15-4 (2004); §§22-19-1,22-22-1 (2006)

Tennessee Tenn. Code Ann. §§37-1-134, 40-35-120(g) (Westlaw 2010)

Utah Utah Code Ann. §§78A-6-602, 78A-6-703, 76-5-302 (Lexis 2008)

Virginia Va. Code Ann. §§16.1-269.1, 18.2-61, 53.1-151(B1) (2009)

Washington Wash. Rev. Code § 13.40.110 (2009 Supp.); §§9A.04.050, 9.94A.030(34), 9.94A.570 (2008)

West Virginia W. Va. Code Ann. §49-5-10 (Lexis 2009); §61-2-14a(a) (Lexis 2005)

Wisconsin Wis. Stat. §§938.18,938.183 (2007-2008); §939.62(2m)(c) (Westlaw 2005)

Wyoming Wyo. Stat. Ann. §§6-2-306(d), (e), 14-6-203 (2009)

Federal 18 U. S. C. §2241 (2006 ed. and Supp. II); §5032 (2006 ed.)

II. JURISDICTIONS THAT PERMIT LIFE WITHOUT PAROLE FOR JUVENILE OFFENDERS CONVICTED OF HOMICIDE CRIMES ONLY

Connecticut Conn. Gen. Stat. §53a-35a (2009)

Hawaii Haw. Rev. Stat. §571-22(d) (2006); §706-656(1) (2008 Supp. Pamphlet)

Maine Me. Rev. Stat. Ann., Tit. 15, §3101(4) (Supp. 2009); id., Tit. 17-a, §1251 (2006)

Massachusetts Mass Gen. Laws eh. 119, §74, id., ch. 265, §2 (West 2008)

New Jersey N. J. Stat. Ann. §2A:4A-26 (West Supp. 2009); §2C:11-3(b)(2) (West Supp. 2009)

New Mexico N. M. Stat. Ann. §31-18-14 (Supp. 2009); §31-18-15.2(A) (Westlaw 2010)

Vt. Stat. Ann., Tit. 33, §5204 (2009 Cum. Supp.); id., Tit. 13, §2303 (2009) Vermont

*85III. JURISDICTIONS THAT FORBID LIFE WITHOUT PAROLE FOR JUVENILE OFFENDERS

Alaska Alaska Stat. § Í2.55.015(g) (2008)

Colorado Colo. Rev. Stat. Ann. § 18-1.3-401(4)(b) (2009)

Kansas Kan. Stat. Ann. §21-4622 (West 2007)

Kentucky Ky. Rev. Stat. Ann. § 640.040 (West 2008); Shepherd v. Commonwealth, 251 S. W. 3d 309, 320-321 (Ky. 2008)

Montana Mont. Code Ann. §46-18-222(1) (2009)

Texas Tex. Penal Code Ann. § 12.31 (West Supp. 2009)

Justice Stevens,

with whom Justice Ginsburg and Justice Sotomayor join, concurring.

In his dissenting opinion, Justice Thomas argues that today’s holding is not entirely consistent with the controlling opinions in Lockyer v. Andrade, 538 U. S. 63 (2003), Ewing v. California, 538 U. S. 11 (2003), Harmelin v. Michigan, 501 U. S. 957 (1991), and Rummel v. Estelle, 445 U. S. 263 (1980). Post, at 102-105. Given that “evolving standards of decency” have played a central role in our Eighth Amendment jurisprudence for at least a century, see Weems v. United States, 217 U. S. 349, 373-378 (1910), this argument suggests the dissenting opinions in those cases more accurately describe the law today than does Justice Thomas’ rigid interpretation of the Amendment. Society changes. Knowledge accumulates. We learn, sometimes, from our mistakes. Punishments that did not seem cruel and unusual at one time may, in the light of reason and experience, be found cruel and unusual at a later time; unless we are to abandon the moral commitment embodied in the Eighth Amendment, proportionality review must never become effectively obsolete, post, at 103-104, and n. 2.

While Justice Thomas would apparently not rule out a death sentence for a $50 theft by a 7-year-old, see post, at 100, 106, n. 3, the Court wisely rejects his static approach to the law. Standards of decency have evolved since 1980. They will never stop doing so.

*86Chief Justice Roberts,

concurring in the judgment.

I agree with the Court that Terrance Graham’s sentence of life without parole violates the Eighth Amendment's prohibition on “cruel and unusual punishments.” Unlike the majority, however, I see no need to invent a new constitutional rule of dubious provenance in reaching that conclusion. Instead, my analysis is based on an application of this Court’s precedents, in particular (1) our cases requiring “narrow proportionality” review of noncapital sentences and (2) our conclusion in Roper v. Simmons, 543 U. S. 551 (2005), that juvenile offenders are generally less culpable than adults who commit the same crimes.

These cases expressly allow courts addressing allegations that a noncapital sentence violates the Eighth Amendment to consider the particular defendant and particular crime at issue. The standards for relief under these precedents are rigorous, and should be. But here Graham’s juvenile status — together with the nature of his criminal conduct and the extraordinarily severe punishment imposed — lead me to conclude that his sentence of life without parole is unconstitutional.

I

Our Court has struggled with whether and how to apply the Cruel and Unusual Punishments Clause to sentences for noncapital crimes. Some of my colleagues have raised serious and thoughtful questions about whether, as an original matter, the Constitution was understood to require any degree of proportionality between noncapital offenses and their corresponding punishments. See, e. g., Harmelin v. Michigan, 501 U. S. 957, 962-994 (1991) (principal opinion of Scalia, J.); post, at 99-100, and n. 1 (Thomas, J., dissenting). Neither party here asks us to reexamine our precedents requiring such proportionality, however, and so I approach this case by trying to apply our past decisions to the facts at hand.

*87A

Graham’s case arises at the intersection of two lines of Eighth Amendment precedent. The first consists of decisions holding that the Cruel and Unusual Punishments Clause embraces a “narrow proportionality principle” that we apply, on a case-by-case basis, when asked to review non-capital sentences. Lockyer v. Andrade, 538 U. S. 63, 72 (2003) (internal quotation marks omitted); Solem v. Helm, 463 U. S. 277, 290 (1983); Ewing v. California, 538 U. S. 11, 20 (2003) (plurality opinion); Harmelin, supra, at 996-997 (Kennedy, J., concurring in part and concurring in judgment). This “narrow proportionality principle” does not grant judges blanket authority to second-guess decisions made by legislatures or sentencing courts. On the contrary, a reviewing court will only “rarely” need “to engage in extended analysis to determine that a sentence is not constitutionally disproportionate,” Solem, supra, at 290, n. 16 (emphasis added), and “successful challenges” to noncapital sentences will be all the more “exceedingly rare,” Rummel v. Estelle, 445 U. S. 263, 272 (1980).

We have “not established a clear or consistent path for courts to follow” in applying the highly deferential “narrow proportionality” analysis. Lockyer, supra, at 72. We have, however, emphasized the primacy of the legislature in setting sentences, the variety of legitimate penological schemes, the state-by-state diversity protected by our federal system, and the requirement that review be guided by objective, rather than subjective, factors. Ewing, supra, at 23 (plurality opinion); Harmelin, supra, at 998-1001 (opinion of Kennedy, J.). Most importantly, however, we have explained that the Eighth Amendment “ ‘does not require strict proportionality between crime and sentence’ ”; rather, “ ‘it forbids only extreme sentences that are “grossly disproportionate” to the crime.’ ” Ewing, supra, at 23 (plurality opinion) (quoting Harmelin, supra, at 1001 (opinion of Kennedy, J.)).

*88Our cases indicate that courts conducting “narrow proportionality” review should begin with a threshold inquiry that compares “the gravity of the offense and the harshness of the penalty.” Solem, 463 U. S., at 290-291. This analysis can consider a particular offender’s mental state and motive in committing the crime, the actual harm caused to his victim or to society by his conduct, and any prior criminal history. Id., at 292-294, 296-297, and n. 22 (considering motive, past criminal conduct, alcoholism, and propensity for violence of the particular defendant); see also Ewing, swpra, at 28-30 (plurality opinion) (examining defendant’s criminal history); Harmelin, 501 U. S., at 1001-1004 (opinion of Kennedy, J.) (noting specific details of the particular crime of conviction).

Only in “the rare case in which a threshold comparison of the crime committed and the sentence imposed leads to an inference of gross disproportionality,” id., at 1005, should courts proceed to an “intrajurisdictional” comparison of the sentence at issue with those imposed on other criminals in the same jurisdiction, and an “interjurisdictional” comparison with sentences imposed for the same crime in other jurisdictions, Solem, supra, at 291-292. If these subsequent comparisons confirm the inference of gross disproportion-ality, courts should invalidate the sentence as a violation of the Eighth Amendment.

B

The second line of precedent relevant to assessing Graham’s sentence consists of our cases acknowledging that juvenile offenders are generally — though not necessarily in every case — less morally culpable than adults who commit the same crimes. This insight animated our decision in Thompson v. Oklahoma, 487 U. S. 815 (1988), in which we invalidated a capital sentence imposed on a juvenile who had committed his crime under the age of 16. More recently, in Roper, 543 U. S. 551, we extended the prohibition on executions to those who committed their crimes before the age of 18.

*89Both Thompson and Roper arose in the unique context of the death penalty, a punishment that our Court has recognized “must be limited to those offenders who commit ‘a narrow category of the most serious crimes’ and whose extreme culpability makes them ‘the most deserving of execution.’” 543 U. S., at 568 (quoting Atkins v. Virginia, 536 U. S. 304, 319 (2002)). Roper’s prohibition on the juvenile death penalty followed from our conclusion that “[t]hree general differences between juveniles under 18 and adults demonstrate that juvenile offenders cannot with reliability be classified among the worst offenders.” 543 U. S., at 569. These differences are a lack of maturity and an underdeveloped sense of responsibility, a heightened susceptibility to negative influences and outside pressures, and the fact that the character of a juvenile is “more transitory” and “less fixed” than that of an adult. Id., at 569-570. Together, these factors establish the “diminished culpability of juveniles,” id., at 571, and “render suspect any conclusion” that juveniles are among “the worst offenders” for whom the death penalty is reserved, id., at 570.

Today, the Court views Roper as providing the basis for a new categorical rule that juveniles may never receive a sentence of life without parole for nonhomicide crimes. I disagree. In Roper, the Court tailored its analysis of juvenile characteristics to the specific question whether juvenile offenders could constitutionally be subject to capital punishment. Our answer that they could not be sentenced to death was based on the explicit conclusion that they “cannot with reliability be classified among the worst offenders.” Id., at 569 (emphasis added).

This conclusion does not establish that juveniles can never be eligible for life without parole. A life sentence is of course far less severe than a death sentence, and we have never required that it be imposed only on the very worst offenders, as we have with capital punishment. Treating juvenile life sentences as analogous to capital punishment is at *90odds with our longstanding view that “the death penalty is different from other punishments in kind rather than degree.” Solem, supra, at 294. It is also at odds with Roper itself, which drew the line at capital punishment by blessing juvenile sentences that are “less severe than death” despite involving “forfeiture of some of the most basic liberties.” 543 U. S., at 573-574. Indeed, Roper explicitly relied on the possible imposition of life without parole on some juvenile offenders. Id., at 572.

But the fact that Roper does not support a categorical rule barring life sentences for all juveniles does not mean that a criminal defendant’s age is irrelevant to those sentences. On the contrary, our cases establish that the “narrow proportionality” review applicable to noncapital cases itself takes the personal “culpability of the offender” into account in examining whether a given punishment is proportionate to the crime. Solem, supra, at 292. There is no reason why an offender’s juvenile status should be excluded from the analysis. Indeed, given Roper’s conclusion that juveniles are typically less blameworthy than adults, 543 U. S., at 571, an offender’s juvenile status can play a central role in the inquiry.

Justice Thomas disagrees with even our limited reliance on Roper on the ground that the present case does not involve capital punishment. Post, at 121 (dissenting opinion). That distinction is important — indeed, it underlies our rejection of the categorical rule declared by the Court. But Roper’s conclusion that juveniles are typically less culpable than adults has pertinence beyond capital cases, and rightly informs the ease-specific inquiry I believe to be appropriate here.

In short, our existing precedent already provides a sufficient framework for assessing the concerns outlined by the majority. Not every juvenile receiving a life sentence will prevail under this approach. Not every juvenile should. But all will receive the protection that the Eighth Amendment requires.

*91II

Applying the “narrow proportionality” framework to the particular facts of this ease, I conclude that Graham’s sentence of life without parole violates the Eighth Amendment.*

A

I begin with the threshold inquiry comparing the gravity of Graham’s conduct to the harshness of his penalty. There is no question that the crime for which Graham received his life sentence — armed burglary of a nondomicile with an assault or battery — is “a serious crime deserving serious punishment.” Enmund v. Florida, 458 U. S. 782, 797 (1982). So too is the home invasion robbery that was the basis of Graham’s probation violation. But these crimes are certainly less serious than other crimes, such as murder or rape.

As for Graham’s degree of personal culpability, he committed the relevant offenses when he was a juvenile — a stage at which, Roper emphasized, one’s “culpability or blameworthiness is diminished, to a substantial degree, by reason of youth and immaturity.” 543 U. S., at 571. Graham’s age places him in a significantly different category from the defendants in Rummél, Harmelin, and Ewing, all of whom committed their crimes as adults. Graham’s youth made *92him relatively more likely to engage in reckless and dangerous criminal activity than an adult; it also likely enhanced his susceptibility to peer pressure. See, e. g., Roper, supra, at 569; Johnson v. Texas, 509 U. S. 350, 367 (1993); Eddings v. Oklahoma, 455 U. S. 104, 115-117 (1982). There is no reason to believe that Graham should be denied the general presumption of diminished culpability that Roper indicates should apply to juvenile offenders. If anything, Graham’s in-court statements — including his request for a second chance so that he could “do whatever it takes to get to the NFL” — underscore his immaturity. App. 380.

The fact that Graham committed the crimes that he did proves that he was dangerous and deserved to be punished. But it does not establish that he was particularly dangerous — at least relative to the murderers and rapists for whom the sentence of life without parole is typically reserved. On the contrary, his lack of prior criminal convictions, his youth and immaturity, and the difficult circumstances of his upbringing noted by the majority, ante, at 53, all suggest that he was markedly less culpable than a typical adult who commits the same offenses.

Despite these considerations, the trial court sentenced Graham to life in prison without the possibility of parole. This is the second-harshest sentence available under our precedents for any crime, and the most severe sanction available for a nonhomicide offense. See Kennedy v. Louisiana, 554 U. S. 407 (2008). Indeed, as the majority notes, Graham’s sentence far exceeded the punishment proposed by the Florida Department of Corrections (which suggested a sentence of four years, Brief for Petitioner 20), and the state prosecutors (who asked that he be sentenced to. 30 years in prison for the armed burglary, App. 388). No one in Graham’s case other than the sentencing judge appears to have believed that Graham deserved to go to prison for life.

Based on the foregoing circumstances, I conclude that there is a strong inference that Graham’s sentence of life *93imprisonment without parole was grossly disproportionate in violation of the Eighth Amendment. I therefore proceed to the next steps of the proportionality analysis.

B

Both intrajurisdictional and interjurisdietional comparisons of Graham’s sentence confirm the threshold inference of disproportionality.

Graham’s sentence was far more severe than that imposed for similar violations of Florida law, even without taking juvenile status into account. For example, individuals who commit burglary or robbery offenses in Florida receive average sentences of less than 5 years and less than 10 years, respectively. Florida Dept, of Corrections, Annual Report FY 2007-2008: The Guidebook to Corrections in Florida 35. Unsurprisingly, Florida’s juvenile criminals receive similarly low sentences — typically less than five years for burglary and less than seven years for robbery. Id., at 36. Graham's life without parole sentence was far more severe than the average sentence imposed on those convicted of murder or manslaughter, who typically receive under 25 years in prison. Id., at 35. As the Court explained in Solem, 463 U. S., at 291, “[i]f more serious crimes are subject to the same penalty, or to less serious penalties, that is some indication that the punishment at issue may be excessive.”

Finally, the inference that Graham’s sentence is disproportionate is further validated by comparison to the sentences imposed in other domestic jurisdictions. As the majority opinion explains, Florida is an outlier in its willingness to impose sentences of life without parole on juveniles convicted of nonhomieide crimes. See ante, at 62-64.

Ill

So much for Graham. But what about Milagro Cunningham, a 17-year-old who beat and raped an 8-year-old girl before leaving her to die under 197 pounds of rock in a recy*94cling bin in a remote landfill? See Musgrave, Cruel or Necessary? Life Terms for Youths Spur National Debate, Palm Beach Post, Oct. 15, 2009, p. 1A. Or Nathan Walker and Jakaris Taylor, the Florida juveniles who together with their friends gang-raped a woman and forced her to perform oral sex on her 12-year-old son? See 3 Sentenced to Life for Gang Rape of Mother, Associated Press, Oct. 14,2009. The fact that Graham cannot be sentenced to life without parole for his conduct says nothing whatever about these offenders, or others like them who commit nonhomicide crimes far more reprehensible than the conduct at issue here. The Court uses Graham’s case as a vehicle to proclaim a new constitutional rule — applicable well beyond the particular facts of Graham’s case — that a sentence of life without parole imposed on any juvenile for any nonhomicide offense is unconstitutional. This categorical conclusion is as unnecessary as it is unwise.

A holding this broad is unnecessary because the particular conduct and circumstances at issue in the case before us are not serious enough to justify Graham’s sentence. In reaching this conclusion, there is no need for the Court to decide whether that same sentence would be constitutional if imposed for other more heinous nonhomicide crimes.

A more restrained approach is especially appropriate in light of the Court’s apparent recognition that it is perfectly legitimate for a juvenile to receive a sentence of life without parole for committing murder. This means that there is nothing inherently unconstitutional about imposing sentences of life without parole on juvenile offenders; rather, the constitutionality of such sentences depends on the particular crimes for which they are imposed. But if the constitutionality of the sentence turns on the particular crime being punished, then the Court should limit its holding to the particular offenses that Graham committed here, and should decline to consider other hypothetical crimes not presented by this case.

*95In any event, the Court’s categorical conclusion is also unwise. Most importantly, it ignores the fact that some nonhomicide crimes — like , the ones committed by Milagro Cunningham, Nathan Walker, and Jakaris Taylor — are especially heinous or grotesque, and thus may be deserving of more severe punishment.

Those under 18 years old may as a general matter have “diminished” culpability relative to adults who commit the same crimes, Roper, 543 U. S., at 571, but that does not mean that their culpability is always insufficient to justify a life sentence. See generally Thompson, 487 U. S., at 853 (O’Connor, J., concurring in judgment). It does not take a moral sense that is fully developed in every respect to know that beating and raping an 8-year-old girl and leaving her to die under 197 pounds of rocks is horribly wrong. The single fact of being 17 years old would not afford Cunningham protection against life without parole if the young girl had died — as Cunningham surely expected she would — so why should it do so when she miraculously survived his barbaric brutality?

The Court defends its categorical approach on the grounds that a “clear line is necessary to prevent the possibility that life without parole sentences will be imposed on juvenile nonhomicide offenders who are not sufficiently culpable to merit that punishment.” Ante, at 74. It argues that a case-by-case approach to proportionality review is constitutionally insufficient because courts might not be able “with sufficient accuracy [to] distinguish the few incorrigible juvenile offenders from the many that have the capacity for change.” Ante, at 77.

The Court is of course correct that judges will never have perfect foresight — or perfect wisdom — in making sentencing decisions. But this is true when they sentence adults no less than when they sentence juveniles. It is also true when they sentence juveniles who commit murder no less than when they sentence juveniles who commit other crimes.

*96Our system depends upon sentencing judges applying their reasoned judgment to each case that comes before them. As we explained in Solem, the whole enterprise of proportionality review is premised on the “justified” assumption that “courts are competent to judge the gravity of an offense, at least on a relative scale.” 463 U. S., at 292. Indeed, “courts traditionally have made these judgments” by applying “generally accepted criteria” to analyze “the harm caused or threatened to the victim or society, and the culpability of the offender.” Id., at 292, 294.

* * *

Terrance Graham committed serious offenses, for which he deserves serious punishment. But he was only 16 years old, and under our Court’s precedents, his youth is one factor, among others, that should be considered in deciding whether his punishment was unconstitutionally excessive. In my view, Graham’s age — together with the nature of his criminal activity and the unusual severity of his sentence — tips the constitutional balance. I thus concur in the Court’s judgment that Graham’s sentence of life without parole violated the Eighth Amendment.

I would not, however, reach the same conclusion in every case involving a juvenile offender. Some crimes are so heinous, and some juvenile offenders so highly culpable, that a sentence of life without parole may be entirely justified under the Constitution. As we have said, “successful challenges” to noncapital sentences under the Eighth Amendment have been — and, in my view, should continue to be — “exceedingly rare.” Rummel, 445 U. S., at 272. But Graham’s sentence presents the exceptional case that our precedents have recognized will come along. We should grant Graham the relief to which he is entitled under the Eighth Amendment. The Court errs, however, in using this case as a vehicle for unsettling our established jurisprudence and fashioning a categorical rule applicable to far different cases.

*97Justice Thomas,

with whom Justice Scalia joins, and with whom Justice Alito joins as to Parts I and III, dissenting.

The Court holds today that it is “grossly disproportionate” and hence unconstitutional for any judge or jury to impose a sentence of life without parole on an offender less than 18 years old, unless he has committed a homicide. Although the text of the Constitution is silent regarding the permissibility of this sentencing practice, and although it would not have offended the standards that prevailed at the founding, the Court insists that the standards of American society have evolved such that the Constitution now requires its prohibition.

The news of this evolution will, I think, come as a surprise to the American people. Congress, the District of Columbia, and 37 States allow judges and juries to consider this sentencing practice in juvenile nonhomieide cases, and those judges and juries have decided to use it in the very worst cases they have encountered.

The Court does not conclude that life without parole itself is a cruel and unusual punishment. It instead rejects the judgments of those legislatures, judges, and juries regarding what the Court describes as the “moral” question whether this sentence can ever be “proportiona[te]” when applied to the category of offenders at issue here. Ante, at 58, 59 (internal quotation marks omitted); ante, at 85 (Stevens, J., concurring).

I am unwilling to assume that we, as Members of this Court, are any more capable of making such moral judgments than our fellow citizens. Nothing in our training as judges qualifies us for that task, and nothing in Article III gives us that authority.

I respectfully dissent.

I

The Court recounts the facts of Terrance Jamar Graham’s case in detail, so only a summary is necessary here. At age *9816 years and 6 months, Graham and two masked accomplices committed a burglary at a small Florida restaurant, during which one of Graham’s accomplices twice struck the restaurant manager on the head with a steel pipe when he refused to turn over money to the intruders. Graham was arrested and charged as an adult. He later pleaded guilty to two offenses, including armed burglary with assault or battery, an offense punishable by life imprisonment under Florida law. Fla. Stat. §§810.02(2)(a), (b) (2007). The trial court withheld adjudication on both counts, however, and sentenced Graham to probation, the first 12 months of which he spent in a county detention facility.

Graham reoffended just six months after his release. At a probation revocation hearing, a judge found by a preponderance of the evidence that, at age 17 years and 11 months, Graham invaded a home with two accomplices and held the homeowner at gunpoint for approximately 30 minutes while his accomplices ransacked the residence. As a result, the judge concluded that Graham had violated his probation and, after additional hearings, adjudicated Graham guilty on both counts arising from the restaurant robbery. The judge imposed the maximum sentence allowed by Florida law on the armed burglary count, life imprisonment without the possibility of parole.

Graham argues, and the Court holds, that this sentence violates the Eighth Amendment’s Cruel and Unusual Punishments Clause because a life-without-parole sentence is always “grossly disproportionate” when imposed on a person under 18 who commits any crime short of a homicide. Brief for Petitioner 24; ante, at 72.

II

A

The Eighth Amendment, which applies to the States through the Fourteenth, provides that “[ejxcessive bail shall *99not be required, nor excessive fines imposed, nor cruel and unusual punishments inflicted.” It is by now well established that the Cruel and Unusual Punishments Clause was originally understood as prohibiting torturous “ ‘methods of punishment/ ” Harmelin v. Michigan, 501 U. S. 957, 979 (1991) (opinion of Scalia, J.) (quoting Granucci, “Nor Cruel and Unusual Punishments Inflicted:” The Original Meaning, 57 Cal. L. Rev. 839, 842 (1969)) — specifically methods akin to those that had been considered cruel and unusual at the time the Bill of Rights was adopted, Baze v. Rees, 553 U. S. 35, 99 (2008) (Thomas, J., concurring in judgment). With one arguable exception, see Weems v. United States, 217 U. S. 349 (1910); Harmelin, supra, at 990-994 (opinion of Scalia, J.) (discussing the scope and relevance of Weems’ holding), this Court applied the Clause with that understanding for nearly 170 years after the Eighth Amendment’s ratification.

More recently, however, the Court has held that the Clause authorizes it to proscribe not only methods of punishment that qualify as “cruel and unusual,” but also any punishment that the Court deems “grossly disproportionate” to the crime committed. Ante, at 58 (internal quotation marks omitted). This latter interpretation is entirely the Court’s creation. As has been described elsewhere at length, there is virtually no indication that the Cruel and Unusual Punishments Clause originally was understood to require proportionality in sentencing. See Harmelin, 501 U. S., at 975-985 (opinion of Scalia, J.). Here, it suffices to recall just two points. First, the Clause does not expressly refer to proportionality or invoke any synonym for that term, even though the Framers were familiar with the concept, as evidenced by several founding-era state constitutions that required (albeit without defining) proportional punishments. See id., at 977-978. In addition, the penal statute adopted by the First Congress demonstrates that proportionality in sentencing was not con*100sidered a constitutional command.1 See id., at 980-981 (noting that the statute prescribed capital punishment for offenses ranging from “ ‘run[ning] away with... goods or merchandise to the value of fifty dollars/” to “murder on the high seas” (quoting 1 Stat. 114)); see also Preyer, Penal Measures in the American Colonies: An Overview, 26 Am. J. Legal Hist. 326, 348-349, 353 (1982) (explaining that crimes in the late 18th-century Colonies generally were punished either by fines, whipping, or public “shaming,” or by death, as intermediate sentencing options such as incarceration were not common).

The Court has nonetheless invoked proportionality to declare that capital punishment — though not unconstitutional per se — is categorically too harsh a penalty to apply to certain types of crimes and certain classes of offenders. See Coker v. Georgia, 433 U. S. 584 (1977) (plurality opinion) (rape of an adult woman); Kennedy v. Louisiana, 554 U. S. 407 (2008) (rape of a child); Enmund v. Florida, 458 U. S. 782 (1982) (felony murder in which the defendant participated in the felony but did not kill or intend to kill); Thompson v. Oklahoma, 487 U. S. 815 (1988) (plurality opinion) (juveniles *101under 16); Roper v. Simmons, 543 U. S. 551 (2005) (juveniles under 18); Atkins v. Virginia, 536 U. S. 304 (2002) (mentally retarded offenders). In adopting these categorical proportionality rules, the Court intrudes upon areas that the Constitution reserves to other (state and federal) organs of government. The Eighth Amendment prohibits the government from inflicting a cruel and unusual method of punishment upon a defendant. Other constitutional provisions ensure the defendant’s right to fair process before any punishment is imposed. But, as members of today’s majority note, “[sjociety changes,” ante, at 85 (Stevens, J., concurring), and the Eighth Amendment leaves the unavoidably moral question of who “deserves” a particular nonprohibited method of punishment to the judgment of the legislatures that authorize the penalty, the prosecutors who seek it, and the judges and juries that impose it under circumstances they deem appropriate.

The Court has nonetheless adopted categorical rules that shield entire classes of offenses and offenders from the death penalty on the theory that “evolving standards of decency” require this result. Ante, at 58 (internal quotation marks omitted). The Court has offered assurances that these standards can be reliably measured by “ ‘objective indicia’ ” of “national consensus,” such as state and federal legislation, jury behavior, and (surprisingly, given that we are talking about “national” consensus) international opinion. Ante, at 61 (quoting Roper, supra, at 563); see also ante, at 62-67, 80-82. Yet even assuming that is true, the Framers did not provide for the constitutionality of a particular type of punishment to turn on a “snapshot of American public opinion” taken at the moment a case is decided. Roper, supra, at 629 (Scalia, J., dissenting). By holding otherwise, the Court pretermits in all but one direction the evolution of the standards it describes, thus “calling a constitutional halt to what may well be a pendulum swing in social attitudes,” Thompson, supra, at 869 (Scalia, J., dissenting), and “stunt[ing] *102legislative consideration” of new questions of penal policy as they emerge, Kennedy, supra, at 448 (Auto, J., dissenting).

But the Court is not content to rely on snapshots of community consensus in any event. Ante, at 67 (“Community consensus, while ‘entitled to great weight/ is not itself determinative” (quoting Kennedy, supra, at 434)). Instead, it reserves the right to reject the evidence of consensus it finds whenever its own “independent judgment” points in a different direction. Ante, at 67. The Court thus openly claims the power not only to approve or disapprove of democratic choices in penal policy based on evidence of how society’s standards have evolved, but also on the basis of the Court’s “independent” perception of how those standards should evolve, which depends on what the Court concedes is “ ‘ “necessarily ... a moral judgment” ’ ” regarding the propriety of a given punishment in today’s society. Ante, at 58 (quoting Kennedy, supra, at 419).

The categorical proportionality review the Court employs in capital cases thus lacks a principled foundation. The Court’s decision today is significant because it does not merely apply this standard — it remarkably expands its reach. For the first time in its history, the Court declares an entire class of offenders immune from a noncapital sentence using the categorical approach it previously reserved for death penalty cases alone.

B

Until today, the Court has based its categorical proportionality rulings on the notion that the Constitution gives special protection to capital defendants because the death penalty is a uniquely severe punishment that must be reserved for only those who are “most deserving of execution.” Atkins, supra, at 319; see Roper, supra, at 568; Eddings v. Oklahoma, 455 U. S. 104 (1982); Lockett v. Ohio, 438 U. S. 586 (1978). Of course, the Eighth Amendment itself makes no *103distinction between capital and noncapital sentencing, but the “ ‘bright line’ ” the Court drew between the two penalties has for many years served as the principal justification for the Court’s willingness to reject democratic choices regarding the death penalty. See Rummel v. Estelle, 445 U. S. 263, 275 (1980).

Today’s decision eviscerates that distinction. “Death is different” no longer. The Court now claims not only the power categorically to reserve the “most severe punishment” for those the Court thinks are “ ‘the most deserving of execution/ ” Roper, supra, at 568 (quoting Atkins, supra, at 319), but also to declare that “less culpable” persons are categorically exempt from the “second most severe penalty.” Ante, at 72 (emphasis added). No reliable limiting principle remains to prevent the Court from immunizing any class of offenders from the law’s third, fourth, fifth, or fiftieth most severe penalties as well.

The Court’s departure from the “death is different” distinction is especially mystifying when one considers how long it has resisted crossing that divide. Indeed, for a time the Court declined to apply proportionality principles to noncapital sentences at all, emphasizing that “a sentence of death differs in kind from any sentence of imprisonment, no matter how long.” Rummel, 445 U. S., at 272 (emphasis added). Based on that rationale, the Court found that the exeessiveness of one prison term as compared to another was “properly within the province of legislatures, not courts,” id., at 275-276, precisely because it involved an “invariably . . . subjective determination, there being no clear way to make ‘any constitutional distinction between one term of years and a shorter or longer term of years/ ” Hutto v. Davis, 454 U. S. 370, 373 (1982) (per curiam) (quoting Rummel, supra, at 275; emphasis added).

Even when the Court broke from that understanding in its 5-to-4 decision in Solem v. Helm, 463 U. S. 277 (1983) (strik*104ing down as “grossly disproportionate” a life-without-parole sentence imposed on a defendant for passing a worthless check), the Court did so only as applied to the facts of that case; it announced no categorical rule. Id., at 288, 303. Moreover, the Court soon cabined Solem’s rationale. The controlling opinion in the Court’s very next noncapital proportionality case emphasized that principles of federalism require substantial deference to legislative choices regarding the proper length of prison sentences. Harmelin, 501 U. S., at 999 (Kennedy, J., concurring in part and concurring in judgment) (“[Mjarked divergences both in underlying theories of sentencing and in the length of prescribed prison terms are the inevitable, often beneficial, result of the federal structure”); id., at 1000 (“[DJiffering attitudes and perceptions of local conditions may yield different, yet rational, conclusions regarding the appropriate length of prison terms for particular crimes”). That opinion thus concluded that “successful challenges to the proportionality of [prison] sentences [would be] exceedingly rare.” Id., at 1001 (internal quotation marks omitted).

They have been rare indeed. In the 28 years since Solem, the Court has considered just three such challenges and has rejected them all, see Ewing v. California, 538 U. S. 11 (2003); Lockyer v. Andrade, 538 U. S. 63 (2003); Harmelin, supra, largely on the theory that criticisms of the “wisdom, cost-efficiency, and effectiveness” of term-of-years prison sentences are “appropriately directed at the legislature^],” not the courts, Ewing, supra, at 27, 28 (plurality opinion). The Court correctly notes that those decisions were “closely divided,” ante, at 59, but so was Solem itself, and it is now fair to describe Solem as an outlier.2

*105Remarkably, the Court today does more than return to So lew’s case-by-ease proportionality standard for noncapital sentences; it hurtles past it to impose a categorical proportionality rule banning life-without-parole sentences not just in this case, but in every case involving a juvenile nonhomi-cide offender, no matter what the circumstances. Neither the Eighth Amendment nor the Court’s precedents justify this decision.

Ill

The Court asserts that categorical proportionality review is necessary here merely because Graham asks for a categorical rule, see ante, at 61, and because the Court thinks clear lines are a good idea, see ante, at 75. I find those factors wholly insufficient to justify the Court’s break from past practice. First, the Court fails to acknowledge that a petitioner seeking to exempt an entire category of offenders from a sentencing practice carries a much heavier burden than one seeking case-specific relief under Solem. Unlike the petitioner in Solem, Graham must establish not only that his own life-without-parole sentence is “grossly disproportionate,” but also that such a sentence is always grossly disproportionate whenever it is applied to a juvenile nonhomi-cide offender, no matter how heinous his crime. Cf. United States v. Salerno, 481 U. S. 739 (1987). Second, even applying the Court’s categorical “evolving standards” test, neither objective evidence of national consensus nor the notions of culpability on which the Court’s “independent judgment” relies can justify the categorical rule it declares here.

*106A

According to the Court, proper Eighth Amendment analysis “begins with objective indicia of national consensus,”* *3 and “[t]he clearest and most reliable objective evidence of contemporary values is the legislation enacted by the country’s legislatures,” ante, at 62 (internal quotation marks omitted). As such, the analysis should end quickly, because a national “consensus” in favor of the Court’s result simply does not exist. The laws of all 50 States, the Federal Government, and the District of Columbia provide that juveniles over a certain age may be tried in adult court if charged with certain crimes.4 See ante, at 82-85 (appendix to opinion of the Court). Forty-five States, the Federal Government, and the District of Columbia expose juvenile offenders charged *107in adult court to the very same range of punishments faced by adults charged with the same crimes. See ante, at 82-84 (Part I). Eight of those States do not make life-without-parole sentences available for any nonhomicide offender, regardless of age.* ***5 All remaining jurisdictions — the Federal Government, the other 37 States, and the District — authorize life-without-parole sentences for certain nonhomicide offenses, and authorize the imposition of such sentences on persons under 18. See ibid. Only five States prohibit juvenile offenders from receiving a life-without-parole sentence that could be imposed on an adult convicted of the same crime.6 ,

No plausible claim of a consensus against this sentencing practice can be made in light of this overwhelming legislative evidence. The sole fact that federal law authorizes this practice singlehandedly refutes the claim that our Nation finds it morally repugnant. The additional reality that 37 out of 50 States (a supermajority of 74%) permit the practice makes the claim utterly implausible. Not only is there no consensus against this penalty, there is a clear legislative consensus in favor of its availability.

Undaunted, however, the Court brushes this evidence aside as “incomplete and unavailing,” declaring that “ ‘[t]here *108are measures of consensus other than legislation.’” Ante, at 62 (quoting Kennedy, 554 U. S., at 433). This is nothing short of stunning. Most importantly, federal civilian law approves this sentencing practice.7 And although the Court has never decided how many state laws are necessary to show consensus, the Court has never banished into constitutional exile a sentencing practice that the laws of a majority, let alone a supermajority, of States expressly permit.8

Moreover, the consistency and direction of recent legislation — a factor the Court previously has relied upon when crafting categorical proportionality rules, see Atkins, 536 U. S., at 315-316; Roper, 543 U. S., at 565-566 — underscores *109the consensus against the rule the Court announces here. In my view, the Court cannot point to a national consensus in favor of its rule without assuming a consensus in favor of the two penological points it later discusses: (1) Juveniles are always less culpable than similarly-situated adults, and (2) juveniles who commit nonhomicide crimes should always receive an opportunity to demonstrate rehabilitation through parole. Ante, at 68-69,74-75. But legislative trends make that assumption untenable.

First, States over the past 20 years have consistently increased the severity of punishments for juvenile offenders. See 1999 DOJ National Report 89 (referring to the 1990's as “a time of unprecedented change as State legislatures crackled] down on juvenile crime”); ibid, (noting that, during that period, “legislatures in 47 States and the District of Columbia enacted laws that made their juvenile justice systems more punitive,” principally by “ma[king] it easier to transfer juvenile offenders from the juvenile justice system to the [adult] criminal justice system”); id., at 104. This, in my view, reveals the States’ widespread agreement that juveniles can sometimes act with the same culpability as adults and that the law should permit judges and juries to consider adult sentences — including life without parole — in those rare and unfortunate cases. See Feld, Unmitigated Punishment: Adolescent Criminal Responsibility and LWOP Sentences, 10 J. Law & Family Studies 11, 69-70 (2007) (noting that life-without-parole sentences for juveniles have increased since the 1980’s); Amnesty International & Human Rights Watch, The Rest of Their Lives: Life Without Parole for Child Offenders in the United States 2, 31 (2005) (same).

Second, legislatures have moved away from parole over the same period. Congress abolished parole for federal offenders in 1984 amid criticism that it was subject to “gamesmanship and cynicism,” Breyer, Federal Sentencing Guidelines Revisited, 11 Fed. Sentencing Rep. 180 (1999) (discussing the Sentencing Reform Act of 1984, 98 Stat. *1101987), and several States have followed suit, see T. Hughes, D. Wilson, & A. Beck, Dept, of Justice, Bureau of Justice Statistics, Trends in State Parole, 1990-2000, p. 1 (2001) (noting that, by the end of 2000, 16 States had abolished parole for all offenses, while another 4 States had abolished it for certain ones). In light of these developments, the argument that there is nationwide consensus that parole must be available to offenders less than 18 years old in every nonhomicide case simply fails.

B

The Court nonetheless dismisses existing legislation, pointing out that life-without-parole sentences are rarely imposed on juvenile nonhomicide offenders — 123 times in recent memory9 by the Court’s calculation, spread out across 11 States.10 Ante, at 62-64. Based on this rarity of use, *111the Court proclaims a consensus against the practice, implying that laws allowing it either reflect the consensus of a prior, less civilized time or are the work of legislatures tone-deaf to the moral values of their constituents that this Court claims to have easily discerned from afar. See ante, at 62.

This logic strains credulity. It has been rejected before. Gregg v. Georgia, 428 U. S. 153, 182 (1976) (joint opinion of Stewart, Powell, and Stevens, JJ.) (“[T]he relative infrequency of jury verdicts imposing the death sentence does not indicate rejection of capital punishment per se. Rather, [it] ... may well reflect the humane feeling that this most irrevocable of sanctions should be reserved for a small number of extreme cases”)- It should also be rejected here. That a punishment is rarely imposed demonstrates nothing more than a general consensus that it should be just that — rarely imposed. It is not proof that the punishment is one the Nation abhors.

The Court nonetheless insists that the 26 States that authorize this penalty, but are not presently incarcerating a juvenile nonhomicide offender on a life-without-parole sentence, cannot be counted as approving its use. The mere fact that the laws of a jurisdiction permit this penalty, the Court explains, “does not indicate that the penalty has been endorsed through deliberate, express, and full legislative consideration.” Ante, at 67.

But this misapplies the Court’s own evolving standards test. Under that test, “[i]t is not the burden of [a State] to establish a national consensus approving what their citizens have voted to do; rather, it is the ‘heavy burden’ of petitioners to establish a national consensus against it.” Stanford v. Kentucky, 492 U. S. 361, 373 (1989) (quoting Gregg, supra, at 175 (joint opinion of Stewart, Powell, and Stevens, JJ.); some emphasis added; citation omitted). In light of this fact, the Court is wrong to equate a jurisdiction’s disuse of a *112legislatively authorized penalty with its moral opposition to it. The fact that the laws of a jurisdiction permit this sentencing practice demonstrates, at a minimum, that the citizens of that jurisdiction find tolerable the possibility that a jury of their peers could impose a life-without-parole sentence on a juvenile whose nonhomicide crime is sufficiently depraved.

The recent case of 16-year-old Keighton Budder illustrates this point. Just weeks before the release of this opinion, an Oklahoma jury sentenced Budder to life without parole after hearing evidence that he viciously attacked a 17-year-old girl who gave him a ride home from a party. See Stogsdill, Teen Gets Life Terms in Stabbing, Rape Case, Tulsa World, Apr. 2, 2010, p. A10; Stogsdill, Delaware County Teen Sentenced in Rape, Assault Case, Tulsa World, May 4, 2010, p. A12. Budder allegedly put the girl’s head “ ‘into a headlock and sliced her throat,’ ” raped her, stabbed her about 20 times, beat her, and pounded her face into the rocks alongside a dirt road. Teen Gets Life Terms in Stabbing, Rape Case, at A10. Miraculously,- the victim survived. Ibid.

Budder’s crime was rare in its brutality. The sentence the jury imposed was also rare. According to the study relied upon by this Court, Oklahoma had no such offender in its prison system before Budder’s offense. P. Annino, D. Rasmussen, & C. Rice, Juvenile Life Without Parole for Non-Homicide Offenses: Florida Compared to Nation 2, 14 (Sept. 14, 2009) (Table A). Without his conviction, therefore, the Court would have counted Oklahoma’s citizens as morally opposed to life-without-parole. sentences for juvenile nonhomicide offenders.

Yet Oklahoma’s experience proves the inescapable flaw in that reasoning: Oklahoma citizens have enacted laws that allow Oklahoma juries to consider life-without-parole sentences in juvenile nonhomicide cases. Oklahoma juries invoke those laws rarely — in the unusual cases that they find exceptionally depraved. I cannot agree with the Court that *113Oklahoma citizens should be constitutionally disabled from using this sentencing practice merely because they have not done so more frequently. If anything, the rarity of this penalty’s use underscores just how judicious sentencing judges and juries across the country have been in invoking it.

This fact is entirely consistent with the Court’s intuition that juveniles generally are less culpable and more capable of growth than adults. See infra, at 116-118. Graham’s own case provides another example. Graham was statutorily eligible for a life-without-parole sentence after his first crime. But the record indicates that the trial court did not give such a sentence serious consideration at Graham’s initial plea hearing. It was only after Graham subsequently violated his parole by invading a home at gunpoint that the maximum sentence was imposed.

In sum, the Court’s calculation that 123 juvenile nonhomi-cide life-without-parole sentences have been imposed nationwide in recent memory, even if accepted, hardly amounts to strong evidence that the sentencing practice offends our common sense of decency.11

*114Finally, I cannot help but note that the statistics the Court finds inadequate to justify the penalty in this case are stronger than those supporting at least one other penalty this Court has upheld. Not long ago, this Court, joined by the author of today’s opinion, upheld the application of the death penalty against a 16-year-old, despite the fact that no such punishment had been carried out on a person of that age in this country in nearly 30 years. See Stanford, 492 U. S., at 374. Whatever the statistical frequency with which life-without-parole sentences have been imposed on juvenile nonhomicide offenders in the last 30 years, it is surely greater than zero.

In the end, however, objective factors such as legislation and the frequency of a penalty’s use are merely ornaments in the Court’s analysis, window dressing that accompanies its judicial fiat.12 By the Court’s own decree, “[cjommunity *115consensus ... is not itself determinative.” Ante, at 67. Only the independent moral judgment of this Court is sufficient to. decide the question. See ibid.

C

Lacking any plausible claim to consensus, the Court shifts to the heart of its argument: its “independent judgment” that this sentencing practice does not “serv[e] legitimate penological goals.” Ibid. The Court begins that analysis with the obligatory preamble that “‘[t]he Eighth Amendment does not mandate adoption of any one penological theory,’ ” ante, at 71 (quoting Harmelin, 501 U. S., at 999 (Kennedy, J., concurring in part and concurring in judgment)), then promptly mandates the adoption of the theories the Court deems best.

First, the Court acknowledges that, at a minimum, the imposition of life-without-parole sentences on juvenile nonhom-icide offenders serves two “legitimate” penological, goals: incapacitation and deterrence. Ante, at 72-74. By definition, such sentences serve the goal of incapacitation by ensuring that juvenile offenders who commit armed burglaries, or those who commit the types of grievous sex crimes described by The Chief Justice, no longer threaten their communities. See ante, at 93-94 (opinion concurring in judgment). That should settle the matter, since the Court acknowledges *116that incapacitation is an “important” penological goal. Ante, at 72. Yet, the Court finds this goal “inadequate” to justify the life-without-parole sentences here. Ibid, (emphasis added). A similar fate befalls deterrence. The Court acknowledges that such sentences will deter future juvenile offenders, at least to some degree, but rejects that peno-logical goal, not as illegitimate, but as insufficient. Ibid. (“[A]ny limited deterrent effect provided by life without parole is not enough to justify the sentence” (emphasis added)).

The Court looks more favorably on rehabilitation, but laments that life-without-parole sentences do little to promote this goal because they result in the offender’s permanent incarceration. Ante, at 74. Of course, the Court recognizes that rehabilitation's “utility and proper implementation” are subject to debate. Ante, at 73. But that does not stop it from declaring that a legislature may not “forswea[r] . . . the rehabilitative ideal.” Ante, at 74. In other words, the Eighth Amendment does not mandate “any one penological theory,” ante, at 71 (internal quotation marks omitted), just one the Court approves.

Ultimately, however, the Court’s “independent judgment” and the proportionality rule itself center on retribution — the notion that a criminal sentence should be proportioned to “‘the personal culpability of the criminal offender.’” Ante, at 67, 71 (quoting Tison v. Arizona, 481 U. S. 137, 149 (1987)). The Court finds that retributive purposes are not served here for two reasons.

1

First, quoting Roper, 543 U. S., at 569-570, the Court concludes that juveniles are less culpable than adults because, as compared to adults, they “have a ‘ “lack of maturity and an underdeveloped sense of responsibility,”’” and “their characters are ‘not as well formed.’” Ante, at 68. As a general matter, this statement is entirely consistent with the *117evidence recounted above that judges and juries impose the sentence at issue quite infrequently, despite legislative authorization to do so in many more cases. See Part III-B, supra. Our society tends to treat the average juvenile as less culpable than the average adult. But the question here does not involve the average juvenile. The question, instead, is whether the Constitution prohibits judges and juries from ever concluding that an offender under the age of 18 has demonstrated sufficient depravity and incorrigibility to warrant his permanent incarceration.

In holding that the Constitution imposes such a ban, the Court cites “developments in psychology and brain science” indicating that juvenile minds “continue to mature through late adolescence,” ante, at 68 (citing Brief for American Medical Association et al. as Amici Curiae 16-24; Brief for American Psychological Association et al. as Amici Curiae 22-27 (hereinafter APA Brief)), and that juveniles are “more likely [than adults] to engage in risky behaviors,” id., at 7. But even if such generalizations from social science were relevant to constitutional rulemaking, the Court misstates the data on which it relies.

The Court equates the propensity of a fairly substantial number of youths to engage in “risky” or antisocial behaviors with the propensity of a much smaller group to commit violent crimes. Ante, at 76-79. But research relied upon by the amici cited in the Court’s opinion differentiates between adolescents for whom antisocial behavior is a fleeting symptom and those for whom it is a lifelong pattern. See Moffitt, Adolescence-Limited and Life-Course-Persistent Antisocial Behavior: A Developmental Taxonomy, 100 Psychological Rev. 674, 678 (1993) (cited in APA Brief 8, 17, 20) (distinguishing between adolescents who are “antisocial only during adolescence” and a smaller group who engage in antisocial behavior “at every life stage” despite “drifting] through successive systems aimed at curbing their deviance”). That research further suggests that the pattern of behavior in the *118latter group often sets in before 18. See Moffitt, swpra, at 684 (“The well-documented resistance of antisocial personality disorder to treatments of all kinds seems to suggest that the life-eourse-persistent style is fixed sometime before age 18”). And, notably, it suggests that violence itself is evidence that an adolescent offender’s antisocial behavior is not transient. See Moffitt, A Review of Research on the Taxonomy of Life-Course Persistent Versus Adolescence-Limited Antisocial Behavior, in Taking Stock: the Status of Criminological Theory 277, 292-293 (F. Cullen, J. Wright, & K. Blevins eds. 2006) (observing that “life-course persistent” males “tended to specialize in serious offenses (carrying a hidden weapon, assault, robbery, violating court orders), whereas adolescence-limited” ones “specialized in non-serious offenses (theft less than $5, public drunkenness, giving false information on application forms, pirating computer software, etc.)”).

In sum, even if it were relevant, none of this psychological or sociological data is sufficient to support the Court’s “ ‘moral’ ” conclusion that youth defeats culpability in every case. Ante, at 68 (quoting Roper, 543 U. S., at 570); see id., at 618 (Scalia, J., dissenting); R. Epstein, The Case Against Adolescence 171 (2007) (reporting on a study of juvenile reasoning skills and concluding that “most teens are capable of conventional, adult-like moral reasoning”).

The Court responds that a categorical rule is nonetheless necessary to prevent the “ ‘unacceptable likelihood’ ” that a judge or jury, unduly swayed by “‘the brutality or coldblooded nature’ ” of a juvenile’s nonhomicide crime, will sentence him to a life-without-parole sentence for which he possesses “ ‘insufficient culpability,’ ” ante, at 78 (quoting Roper, supra, at 572-573). I find that justification entirely insufficient. The integrity of our criminal justice system depends on the ability of citizens to stand between the defendant and an outraged public and dispassionately determine his guilt and the proper amount of punishment based on the evidence *119presented. That process necessarily admits of human error. But so does the process of judging in which we engage. As between the two, I find far. more “unacceptable” that this Court, swayed by studies reflecting the general tendencies of youth, decree that the people of this country are not fit to decide for themselves when the rare case requires different treatment.

2

That is especially so because, in the end, the Court does not even believe its pronouncements about the juvenile mind. If it did, the categorical rule it announces today would be most peculiar because it leaves intact state and federal laws that permit life-without-parole sentences for juveniles who commit homicides. See ante, at 68-69. The Court thus acknowledges that there is nothing inherent in the psyche of a person less than 18 that prevents him from acquiring the moral agency necessary to warrant a life-without-parole sentence. Instead, the Court rejects overwhelming legislative consensus only on the question of which acts are sufficient to demonstrate that moral agency.

The Court is quite willing to accept that a 17-year-old who pulls the trigger on a firearm can demonstrate sufficient depravity and irredeemability to be denied reentry into society, but insists that a 17-year-old who rapes an 8-year-old and leaves her for dead does not. See ibid.; cf. ante, at 93-94 (Roberts, C. J., concurring in judgment) (describing the crime of life-without-parole offender Milagro Cunningham). Thus, the Court’s conclusion that life-without-parole sentences are “grossly disproportionate” for juvenile nonhomi-cide offenders in fact has very little to do with its view of juveniles, and much more to do with its perception that “defendants who do not kill, intend to kill, or foresee that life will be taken are categorically less deserving of the most serious forms of punishment than are murderers.” Ante, at 69.

*120That the Court is willing to impose such an exacting constraint on democratic sentencing choices based on such an untestable philosophical conclusion is remarkable. The question of what acts are “deserving” of what punishments is bound so tightly with questions of morality and social conditions as to make it, almost by definition, a question for legislative resolution. It is true that the Court previously has relied on the notion of proportionality in holding certain classes of offenses categorically exempt from capital punishment. See supra, at 100-101. But never before today has the Court relied on its own view of just deserts to impose a categorical limit on the imposition of a lesser punishment. Its willingness to cross that well-established boundary raises the question whether any democratic choice regarding appropriate punishment is safe from the Court’s ever-expanding constitutional veto.

IV

Although The Chief Justice’s concurrence avoids the problems associated with expanding categorical proportionality review to noncapital cases, it employs noncapital proportionality analysis in a way that raises the same fundamental concern. Although I do not believe Solem merits stare decisis treatment, Graham’s claim cannot prevail even under that test (as it has been limited by the Court’s subsequent precedents). Solem instructs a court first to compare the “gravity” of an offender’s conduct to the “harshness of the penalty” to determine whether an “inference” of gross disproportionality exists. 463 U. S., at 290-291. Only in “the rare case” in which such an inference is present should the court proceed to the “objective” part of the inquiry — an intrajurisdictional and interjurisdictional comparison of the defendant's sentence with others similarly situated. Har-melin, 501 U. S., at 1000, 1005 (Kennedy, J., concurring in part).

*121Under the Court’s precedents, I fail to see how an “inference” of gross disproportionality arises here. The concurrence notes several arguably mitigating facts — Graham’s “lack of prior criminal convictions, his youth and immaturity, and the difficult circumstances of his upbringing.” Ante, at 92 (opinion of Roberts, C. J.). But the Court previously has upheld a life-without-parole sentence imposed on a first-time offender who committed a nonviolent drug crime. See Harmelin, supra, at 1002-1004. Graham’s conviction for an actual violent felony is surely more severe than that offense. As for Graham’s age, it is true that Roper held juveniles categorically ineligible for capital punishment, but as the concurrence explains, Roper was based on the “explicit conclusion that [juveniles] ‘cannot with reliability be classified among the worst offenders’ ”; it did “not establish that juveniles can never be eligible for life without parole.” Ante, at 89 (opinion of Roberts, C. J.) (quoting Roper, 543 U. S., at 569; emphasis added in opinion of Roberts, C. J.). In my view, Roper's principles are thus not generally applicable outside the capital sentencing context.

By holding otherwise, the concurrence relies on the same type of subjective judgment as the Court, only it restrains itself to a ease-by-case rather than a categorical ruling. The concurrence is quite ready to hand Graham “the general presumption of diminished culpability” for juveniles, ante, at 92, apparently because it believes that Graham’s armed burglary and home invasion crimes were “certainly less serious” than murder or rape, ante, at 91. It recoils only from the prospect that the Court would extend the same presumption to a juvenile who commits a sex crime. See ante, at 94-95. I simply cannot accept that these subjective judgments of proportionality are ones the Eighth Amendment authorizes us to make.

The “objective” elements of the Solem test provide no additional support for the concurrence’s conclusion. The concurrence compares Graham’s sentence to “similar” sentences *122in Florida and concludes that Graham’s sentence was “far more severe.” Ante, at 93 (Roberts, C. J., concurring in judgment). But strangely, the concurrence uses average sentences for burglary or robbery offenses as examples of “similar” offenses, even though it seems that a run-of-the-mill burglary or robbery is not at all similar to Graham’s criminal history, which includes a charge for armed burglary with assault, and a probation violation for invading a home at gunpoint.

And even if Graham’s sentence is higher than ones he might have received for an armed burglary with assault in other jurisdictions, see ibid., this hardly seems relevant if one takes seriously the principle that “ ‘[ajbsent a constitutionally imposed uniformity inimical to traditional notions of federalism, some State will always bear the distinction of treating particular offenders more severely than any other State,”' Harmelin, supra, at 1000 (opinion of Kennedy, J.) (quoting Rummel, 445 U. S., at 282; emphasis added). Applying Solem, the Court has upheld a 25-years-to-life sentence for theft under California’s recidivist statute, despite the fact that the State and its amici could cite only “a single instance of a similar sentence imposed outside the context of California’s three strikes law, out of a prison population [then] approaching two million individuals.” Ewing, 538 U. S., at 47 (Breyer, J., dissenting). It has also upheld a life-without-parole sentence for a first-time drug offender in Michigan charged with possessing 672 grams of cocaine despite the fact that only one other State would have authorized such a stiff penalty for a first-time drug offense, and even that State required a far greater quantity of cocaine (10 kilograms) to trigger the penalty. See Harmelin, supra, at 1026 (White, J., dissenting). Graham’s sentence is certainly less rare than the sentences upheld in these cases, so his claim fails even under Solem.

*123Both the Court and the concurrence claim their decisions to be narrow ones, but both invite a host of line-drawing problems to which courts must seek answers beyond the strictures of the Constitution. The Court holds that “[a] State is not required to guarantee eventual freedom to a juvenile offender convicted of a nonhomicide crime,” but must provide the offender with “some meaningful opportunity to obtain release based on demonstrated maturity and rehabilitation.” Ante, at 75. But what, exactly, does such a “meaningful” opportunity entail? When must it occur? And what Eighth Amendment principles will govern review by the parole boards the Court now demands that States empanel? The Court provides no answers to these questions, which will no doubt embroil the courts for years.13

Y

The ultimate question in this case is not whether a life-without-parole sentence “fits” the crime at issue here or the crimes of juvenile nonhomicide offenders more generally, but to whom the Constitution assigns that decision. The Florida Legislature has concluded that such sentences should be available for persons under 18 who commit certain crimes, and the trial judge in this case decided to impose that legislatively authorized sentence here. Because a life-without-parole prison sentence is not a “cruel and unusual” method *124of punishment under any standard, the Eighth Amendment gives this Court no authority to reject those judgments.

It would be unjustifiable for the Court to declare otherwise even if it could claim that a bare majority of state laws supported its independent moral view. The fact that the Court categorically prohibits life-without-parole sentences for juvenile nonhomicide offenders in the face of an overwhelming legislative majority in favor of leaving that sentencing option available under certain cases simply illustrates how far beyond any cognizable constitutional principle the Court has reached to ensure that its own sense of morality and retributive justice pre-empts that of the people and their representatives.

I agree with Justice Stevens that “[w]e learn, sometimes, from our mistakes.” Ante, at 85 (concurring opinion). Perhaps one day the Court will learn from this one.

I respectfully dissent.

Justice Alito,

dissenting.

I join Parts I and III of Justice Thomas' dissenting opinion. I write separately to make two points.

First, the Court holds only that “for a juvenile offender who did not commit homicide the Eighth Amendment forbids the sentence of life without parole.” Ante, at 74 (emphasis added). Nothing in the Court's opinion afreets the imposition of a sentence to a term of years without the possibility of parole. Indeed, petitioner conceded at oral argument that a sentence of as much as 40 years without the possibility of parole “probably” would be constitutional. Tr. of Oral Arg. 6-7; see also ante, at 123, n. 13 (Thomas, J., dissenting).

Second, the question whether petitioner’s sentence violates the narrow, as-applied proportionality principle that applies to noncapital sentences is not properly before us in this case. Although petitioner asserted an as-applied proportionality challenge to his sentence before the Florida courts, see 982 So. 2d 43, 51-53 (Fla. App. 2008), he did not include *125an as-applied claim in his petition for certiorari or in his merits briefs before this Court. Instead, petitioner argued for only a categorical rule banning the imposition of life without parole on any juvenile convicted of a nonhomicide offense. Because petitioner abandoned his as-applied claim, I would not reach that issue. See this Court’s Rule 14.1(a); Yee v. Escondido, 503 U. S. 519, 534-538 (1992).

Justice Kennedy

delivered the opinion of the Court.

This case arises from serious constitutional violations in California’s prison system. The violations have persisted for years. They remain uncorrected. The appeal comes to this Court from a three-judge District Court order directing California to remedy two ongoing violations of the Cruel and Unusual Punishments Clause, a guarantee binding on the States by the Due Process Clause of the Fourteenth Amend-*500men! The violations are the subject of two class actions in two Federal District Courts. The first involves the class of prisoners with serious mental disorders. That case is Coleman v. Brown. The second involves prisoners with serious medical conditions. That ease is Plata v. Brown. The order of the three-judge District Court is applicable to both cases.

After years of litigation, it became apparent that a remedy for the constitutional violations would not be effective absent a reduction in the prison system population. The authority to order release of prisoners as a remedy to cure a systemic violation of the Eighth Amendment is a power reserved to a three-judge district court, not a single-judge district court. 18 U. S. C. § 3626(a). In accordance with that rule, the Coleman and Plata District Judges independently requested that a three-judge court be convened. The Chief Judge of the Court of Appeals for the Ninth Circuit convened a three-judge court composed of the Coleman and Plata District Judges and a third, Ninth Circuit Judge. Because the two cases are interrelated, their limited consolidation for this purpose has a certain utility in avoiding conflicting decrees and aiding judicial consideration and enforcement. The State in this Court has not objected to consolidation, although the State does argue that the three-judge court was prematurely convened. The State also objects to the substance of the three-judge court order, which requires the State to reduce overcrowding in its prisons.

The appeal presents the question whether the remedial order issued by the three-judge court is consistent with requirements and procedures set forth in a congressional statute, the Prison Litigation Reform Act of 1995 (PLRA). 18 U. S. C. § 3626; see Appendix A, infra. The order leaves the choice of means to reduce overcrowding to the discretion of state officials. But absent compliance through new construction, out-of-state transfers, or other means — or modification of the order upon a further showing by the State— *501the State will be required to release some number of prisoners before their full sentences have been served. High recidivism rates must serve as a warning that mistaken or premature release of even one prisoner can cause injury and harm. The release of prisoners in large numbers — assuming the State finds no other way to comply with the order — is a matter of undoubted, grave concern.

At the time of trial, California’s correctional facilities held some 156,000 persons. This is nearly double the number that California’s prisons were designed to hold, and California has been ordered to reduce its prison population to 137.5% of design capacity. By the three-judge court’s own estimate, the required population reduction could be as high as 46,000 persons. Although the State has reduced the population by at least 9,000 persons during the pendency of this appeal, this means a further reduction of 37,000 persons could be required. As will be noted, the reduction need not be accomplished in an indiscriminate manner or in these substantial numbers if satisfactory, alternative remedies or means for compliance are devised. The State may employ measures, including good-time credits and diversion of low-risk offenders and technical parole violators to community-based programs, that will mitigate the order’s impact. The population reduction potentially required is nevertheless of unprecedented sweep and extent.

Yet so too is the continuing injury and harm resulting from these serious constitutional violations. For years the medical and mental health care provided by California’s prisons has fallen short of minimum constitutional requirements and has failed to meet prisoners’ basic health needs. Needless suffering and death have been the well-documented result. Over the whole course of years during which this litigation has been pending, no other remedies have been found to be sufficient. Efforts to remedy the violation have been frustrated by severe overcrowding in California’s prison system. Short-term gains in the provision of care have been *502eroded by the long-term effects of severe and pervasive overcrowding.

Overcrowding has overtaken the limited resources of prison staff; imposed demands well beyond the capacity of medical and mental health facilities; and created unsanitary and unsafe conditions that make progress in the provision of care difficult or impossible to achieve. The overcrowding is the “primary cause of the violation of a Federal right,” 18 U. S. C. § 3626(a)(B)(E)(i), specifically the severe and unlawful mistreatment of prisoners through grossly inadequate provision of medical and mental health care.

This Court now holds that the PLRA does authorize the relief afforded in this case and that the court-mandated population limit is necessary to remedy the violation of prisoners’ constitutional rights. The order of the three-judge court, subject to the right of the State to seek its modification in appropriate circumstances, must be affirmed.

I

A

The degree of overcrowding in California’s prisons is exceptional. California’s prisons are designed to house a population just under 80,000, but at the time of the three-judge court’s decision the population was almost double that. The State’s prisons had operated at around 200% of design capacity for at least 11 years. Prisoners are crammed into spaces neither designed nor intended to house inmates. As many as 200 prisoners may live in a gymnasium, monitored by as few as two or three correctional officers. App. 1337-1338, 1350; see Appendix B, infra. As many as 54 prisoners may share a single toilet. App. 1337.

The Corrections Independent Review Panel, a body appointed by the Governor and composed of correctional consultants and representatives from state agencies, concluded that California’s prisons are “ ‘severely overcrowded, imperiling the safety of both correctional employees and in*503mates.’ ”1 App. to Juris. Statement, O. T. 2009, No. 09-416, p. 56a (hereinafter Juris. App.). In 2006, then-Governor Schwarzenegger declared a state of emergency in the prisons, as “ ‘immediate action is necessary to prevent death and harm caused by California’s severe prison overcrowding.’” Id., at 61a. The consequences of overcrowding identified by the Governor include “ ‘increased, substantial risk for transmission of infectious illness’ ” and a suicide rate “ ‘approaching an average of one per week.’ ” Ibid.

Prisoners in California with serious mental illness do not receive minimal, adequate care. Because of a shortage of treatment beds, suieidal inmates may be held for prolonged periods in telephone-booth-sized cages without toilets. See *504Appendix C, infra. A psychiatric expert reported observing' an inmate who had been held in such a cage for nearly 24 hours, standing in a pool of his own urine, unresponsive and nearly catatonic. Prison officials explained they had “ 'no place to put him.’ ” App. 593. Other inmates awaiting care may be held for months in administrative segregation, where they endure harsh and isolated conditions and receive only limited mental health services. Wait times for mental health care range as high as 12 months. Id., at 704. In 2006, the suicide rate in California’s prisons was nearly 80% higher than the national average for prison populations; and a court-appointed Special Master found that 72.1% of suicides involved “some measure of inadequate assessment, treatment, or intervention, and were therefore most probably foreseeable and/or preventable.”2 Id., at 1781.

Prisoners suffering from physical illness also receive severely deficient care. California’s prisons were designed to meet the medical needs of a population at 100% of design capacity and so have only half the clinical space needed to treat the current population. Id., at 1024. A correctional officer testified that, in one prison, up to 50 sick inmates may be held together in a 12- by 20-foot cage for up to five hours awaiting treatment. Tr. 597-599. The number of staff is *505inadequate, and prisoners face significant delays in access to care. A prisoner with severe abdominal pain died after a 5-week delay in referral to a specialist; a prisoner with “ ‘constant and extreme’ ” chest pain died after an 8-hour delay in evaluation by a doctor; and a prisoner died of testicular cancer after a “failure of MDs to work up for cancer in a young man with 17 months of testicular pain.”3 California Prison Health Care Receivership Corp., K. Imai, Analysis of CDCR Death Reviews 2006, pp. 6-7 (Aug. 2007). Doctor Ronald Shansky, former medical director of the Illinois state prison system, surveyed death reviews for California prisoners. He concluded that extreme departures from the standard of care were “widespread,” Tr. 430, and that the proportion of “possibly preventable or preventable” deaths was “extremely high,” id., at 429.4 Many more prisoners, suffering *506from severe but not life-threatening conditions, experience prolonged illness and unnecessary pain.

B

These conditions are the subject of two federal cases. The first to commence, Coleman v. Brown, was filed in 1990. Coleman involves the class of seriously mentally ill persons in California prisons. Over 15 years ago, in 1995, after a 39-day trial, the Coleman District Court found “overwhelming evidence of the systemic failure to deliver necessary care to mentally ill inmates” in California prisons. Coleman v. Wilson, 912 F. Supp. 1282, 1316 (ED Cal.). The prisons were “seriously and chronically understaffed,” id., at 1306, and had “no effective method for ensuring . . . the competence of their staff,” id., at 1308. The prisons had failed to implement necessary suicide-prevention procedures, “due in large measure to the severe under staffing.” Id., at 1315. Mentally ill inmates “languished for months, or even years, without access to necessary care.” Id., at 1316. “They suffer from severe hallucinations, [and] they decompensate into catatonic states.” Ibid. The court appointed a Special Master to oversee development and implementation of a remedial plan of action.

*507In 2007,12 years after his appointment, the Special Master in Coleman filed a report stating that, after years of slow improvement, the state of mental health care in California’s prisons was deteriorating. App. 489. The Special Master ascribed this change to increased overcrowding. The rise in population had led to greater demand for care, and existing programming space and staffing levels were inadequate to keep pace. Prisons had retained more mental health staff, but the “growth of the resource [had] not matched the rise in demand.” Id., at 482. At the very time the need for space was rising, the need to house the expanding population had also caused a “reduction of programming space now occupied by inmate bunks.” Id., at 479. The State was “facing a four to five-year gap in the availability of sufficient beds to meet the treatment needs of many inmates/patients.” Id., at 481. “[increasing numbers of truly psychotic inmate/patients are trapped in [lower levels of treatment] that cannot meet their needs.” Ibid. The Special Master concluded that many early “achievements have succumbed to the inexorably rising tide of population, leaving behind growing frustration and despair.” Id., at 489.

C

The second action, Plata v. Brown, involves the class of state prisoners with serious medical conditions. After this action commenced in 2001, the State conceded that deficiencies in prison medical care violated prisoners’ Eighth Amendment rights. The State stipulated to a remedial injunction. The State failed to comply with that injunction, and in 2005 the court appointed a Receiver to oversee remedial efforts. The court found that “the California prison medical care system is broken beyond repair,” resulting in an “unconscionable degree of suffering and death.” App. 917. The court found: “[I]t is an uneontested fact that, on average, an inmate in one of California’s prisons needlessly dies every six to seven days due to constitutional deficiencies in the *508[California prisons’] medical delivery system.” Ibid. And the court made findings regarding specific instances of neglect, including the following:

“[A] San Quentin prisoner with hypertension, diabetes and renal failure was prescribed two different medications that actually served to exacerbate his renal failure. An optometrist noted the patient’s retinal bleeding due to very high blood pressure and referred him for immediate evaluation, but this evaluation never took place. It was not until a year later that the patient’s renal failure was recognized, at which point he was referred to a nephrologist on an urgent basis; he should have been seen by the specialist within 14 days but the consultation never happened and the patient died three months later,” Id., at 928 (citations omitted).

Prisons were unable to retain sufficient numbers of competent medical staff, id., at 937, and would “hire any doctor who had ‘a license, a pulse and a pair of shoes,' ” id., at 926. Medical facilities lacked “necessary medical equipment” and did “not meet basic sanitation standards.” Id., at 944. “Exam tables and counter tops, where prisoners with . . . communicable diseases are treated, [were] not routinely disinfected.” Ibid.

In 2008, three years after the District Court’s decision, the Receiver described continuing deficiencies in the health care provided by California prisons:

“Timely access is not assured. The number of medical personnel has been inadequate, and competence has not been assured.... Adequate housing for the disabled and aged does not exist. The medical facilities, when they exist at all, are in an abysmal state of disrepair. Basic medical equipment is often not available or used. Medications and other treatment options are too often not available when needed. . . . Indeed, it is a misnomer to call the existing chaos a ‘medical delivery system' — it is *509more an act of desperation than a system.” Record in No. 3:01-cv-01351-TEH (ND Cal.), Doc. 1136, p. 9.

A report by the Receiver detailed the impact of overcrowding on efforts to remedy the violation. The Receiver explained that “overcrowding, combined with staffing shortages, has created a culture of cynicism, fear, and despair which makes hiring and retaining competent clinicians extremely difficult.” App. 1031. “[0]vercrowding, and the resulting day to day operational chaos of the [prison system], creates regular 'crisis’ situations which . . . take time [and] energy . . . away from important remedial programs.” Id., at 1035. Overcrowding had increased the incidence of infectious disease, id., at 1037-1038, and had led to rising prison violence and greater reliance by custodial staff on lockdowns, which “inhibit the delivery of medical care and increase the staffing necessary for such care,” id., at 1037. “Every day,” the Receiver reported, “California prison wardens and health care managers make the difficult decision as to which of the class actions, Coleman ... or Plata they will fail to comply with because of staff shortages and patient loads.” Id., at 1038.

D

The Coleman and Plata plaintiffs, believing that a remedy for unconstitutional medical and mental health care could not be achieved without reducing overcrowding, moved their respective District Courts to convene a three-judge court empowered under the PLRA to order reductions in the prison population. The judges in both actions granted the request, and the cases were consolidated before a single three-judge court. The State has not challenged the validity of the consolidation in proceedings before this Court, so its propriety is not presented by this appeal.

The three-judge court heard 14 days of testimony and issued a 184-page opinion, making extensive findings of fact. The court ordered California to reduce its prison population *510to 137.5% of the prisons’ design capacity within two years. Assuming the State does not increase capacity through new construction, the order requires a population reduction of 38,000 to 46,000 persons. Because it appears all but certain that the State cannot complete sufficient construction to comply fully with the order, the prison population will have to be reduced to at least some extent. The court did not order the State to achieve this reduction in any particular manner. Instead, the court ordered the State to formulate a plan for compliance and submit its plan for approval by the court.

The State appealed to this Court pursuant to 28 U. S. C. § 1253, and the Court postponed consideration of the question of jurisdiction to the hearing on the merits. Schwarzenegger v. Plata, 560 U. S. 964 (2010).

II

As a consequence of their own actions, prisoners may be deprived of rights that are fundamental to liberty. Yet the law and the Constitution demand recognition of certain other rights. Prisoners retain the essence of human dignity inherent in all persons. Respect for that dignity animates the Eighth Amendment prohibition against cruel and unusual punishment. “‘The basic concept underlying the Eighth Amendment is nothing less than the dignity of man.’” Atkins v. Virginia, 536 U. S. 304, 311 (2002) (quoting Prop v. Dulles, 356 U. S. 86, 100 (1958) (plurality opinion)).

To incarcerate, society takes from prisoners the means to provide for their own needs. Prisoners are dependent on the State for food, clothing, and necessary medical care. A prison’s failure to provide sustenance for inmates “may actually produce physical ‘torture or a lingering death.’” Estelle v. Gamble, 429 U. S. 97, 103 (1976) (quoting In re Kemmler, 136 U. S. 436, 447 (1890)); see generally A. Eisner, Gates of Injustice: The Crisis in America’s Prisons (2004), Just as a prisoner may starve if not fed, he or she may suffer or die if *511not provided adequate medical care. A prison that deprives prisoners of basic sustenance, including adequate medical care, is incompatible with the concept of human dignity and has no place in civilized society.

If government fails to fulfill this obligation, the courts have a responsibility to remedy the resulting Eighth Amendment violation. See Hutto v. Finney, 437 U. S. 678, 687, n. 9 (1978). Courts must be sensitive to the State’s interest in punishment, deterrence, and rehabilitation, as well as the need for deference to experienced and expert prison administrators faced with the difficult and dangerous task of housing large numbers of convicted criminals. See Bell v. Wolfish, 441 U. S. 520, 547-548 (1979). Courts nevertheless must not shrink from their obligation to “enforce the constitutional rights of all ‘persons/ including prisoners.” Cruz v. Beto, 405 U. S. 319, 321 (1972) (per curiam). Courts may not allow constitutional violations to continue simply because a remedy would involve intrusion into the realm of prison administration.

Courts faced with the sensitive task of remedying unconstitutional prison conditions must consider a range of available options, including appointment of special masters or receivers and the possibility of consent decrees. When necessary to ensure compliance with a constitutional mandate, courts may enter orders placing limits on a prison’s population. By its terms, the PLEA restricts the circumstances in which a court may enter an order “that has the purpose or effect of reducing or limiting the prison population.” 18 U. S. C. § 3626(g)(4). The order in this case does not necessarily require the State to release any prisoners. The State may comply by raising the design capacity of its prisons or by transferring prisoners to county facilities or facilities in other States. Because the order limits the prison population as a percentage of design capacity, it nonetheless has the “effect of reducing or limiting the prison population.” Ibid.

*512Under the PLRA, only a three-judge court may enter an order limiting a prison population. § 3626(a)(3)(B). Before a three-judge court may he convened, a district court first must have entered an order for less intrusive relief that failed to remedy the constitutional violation and must have given the defendant a reasonable time to comply with its prior orders. § 3626(a)(3)(A). The party requesting a three-judge court must then submit “materials sufficient to demonstrate that [these requirements] have been met.” § 3626(a)(3)(C). If the district court concludes that the materials are, in fact, sufficient, a three-judge court may be convened. Ibid.; see also 28 U. S. C. § 2284(b)(1) (stating that a three-judge court may not be convened if the district court “determines that three judges are not required”); 17A C. Wright, A. Miller, E. Cooper, & V. Amar, Federal Practice and Procedure § 4235 (3d ed. 2007).

The three-judge court must then find by clear and convincing evidence that “crowding is the primary cause of the violation of a Federal right” and that “no other relief will remedy the violation of the Federal right.” 18 U. S. C. § 3626(a)(3)(E). As with any award of prospective relief under the PLRA, the relief “shall extend no further than necessary to correct the violation of the Federal right of a particular plaintiff or plaintiffs.” §3626(a)(1)(A). The three-judge court must therefore find that the relief is “narrowly drawn, extends no further than necessary . . ., and is the least intrusive means necessary to correct the violation of the Federal right.” Ibid. In making this determination, the three-judge court must give “substantial weight to any adverse impact on public safety or the operation of a criminal justice system caused by the relief.” Ibid. Applying these standards, the three-judge court found a population limit appropriate, necessary, and authorized in this case.

This Court’s review of the three-judge court’s legal determinations is de novo, but factual findings are reviewed for clear error. See Anderson v. Bessemer City, 470 U. S. 564, *513573-574 (1985). Deference to trial court factfinding reflects an understanding that “[t]he trial judge’s major role is the determination of fact, and with experience in fulfilling that role comes expertise.” Id., at 574. The three-judge court oversaw two weeks of trial and heard at considerable length from California prison officials, as well as experts in the field of correctional administration. The judges had the opportunity to ask relevant questions of those witnesses. Two of the judges had overseen the ongoing remedial efforts of the Receiver and Special Master. The three-judge court was well situated to make the difficult factual judgments necessary to fashion a remedy for this complex and intractable constitutional violation. The three-judge court’s findings of fact may be reversed only if this Court is left with a “ ‘definite and firm conviction that a mistake has been committed.’ ” Id., at 573 (quoting United States v. United States Gypsum Co., 333 U. S. 364, 395 (1948)).

A

The State contends that it was error to convene the three-judge court without affording it more time to comply with the prior orders in Coleman and Plata.

1

The parties dispute this Court’s jurisdiction to review the determinations of the Coleman and Plata District Courts that a three-judge court should be convened. Plaintiffs claim the State was required to raise this issue first in the Court of Appeals by appealing the orders of the District Courts. When exercising jurisdiction under 28 U. S. C. §1253, however, this Court ‘‘has not hesitated to exercise jurisdiction ‘to determine the authority of the court below,’” including whether the three-judge court was properly constituted. Gonzalez v. Automatic Employees Credit Union, 419 U. S. 90, 95, n. 12 (1974) (quoting Bailey v. Patterson, 369 U. S. 31, 34 (1962) (per curiam)); see also Gully v. Interstate *514Natural Gas Co., 292 U. S. 16, 18 (1934) (per curiam) (“The case is analogous to those in which this Court, finding that the court below has acted without jurisdiction, exercises its appellate jurisdiction to correct the improper action”). The merits of the decision to convene the three-judge court, therefore, are properly before this Court.

2

Before a three-judge court may be convened to consider whether to enter a population limit, the PLRA requires that the court have “previously entered an order for less intrusive relief that has failed to remedy the deprivation of the Federal right sought to be remedied.” 18 U. S. C. § 3626(a)(3)(A)(i). This provision refers to “an order.” It is satisfied if the court has entered one order, and this single order has “failed to remedy” the constitutional violation. The defendant must also have had “a reasonable amount of time to comply with the previous court orders.” § 3626(a)(3)(A)(ii). This provision refers to the court’s “orders.” It requires that the defendant have been given a reasonable time to comply with all of the court’s orders. Together, these requirements ensure that the “ ‘last resort remedy’” of a population limit is not imposed “‘as a first step.’” Inmates of Occoquan v. Barry, 844 F. 2d 828, 843 (CADC 1988).

The first of these conditions, the previous order requirement of § 3626(a)(3)(A)(i), was satisfied in Coleman by appointment of a Special Master in 1995, and it was satisfied in Plata by approval of a consent decree and stipulated injunction in 2002. Both orders were intended to remedy the constitutional violations. Both were given ample time to succeed. When the three-judge court was convened, 12 years had passed since the appointment of the Coleman Special Master, and 5 years had passed since the approval of the Plata consent decree. The State does not claim that either *515order achieved a remedy. Although the PLRA entitles a State to terminate remedial orders such as these after two years unless the district court finds that the relief “remains necessary to correct a current and ongoing violation of the Federal right,” § 3626(b)(3), California has not attempted to obtain relief on this basis.

The State claims instead that the second condition, the reasonable time requirement of § 3626(a)(3)(A)(ii), was not met because other, later remedial efforts should have been given more time to succeed. In 2006, the Coleman District Judge approved a revised plan of action calling for construction of new facilities, hiring of new staff, and implementation of new procedures. That same year, the Plata District Judge selected and appointed a Receiver to oversee the State’s ongoing remedial efforts. When the three-judge court was convened, the Receiver had filed a preliminary plan of action calling for new construction, hiring of additional staff, and other procedural reforms.

Although both the revised plan of action in Coleman and the appointment of the Receiver in Plata were new developments in the courts’ remedial efforts, the basic plan to solve the crisis through construction, hiring, and procedural reforms remained unchanged. These efforts had been ongoing for years; the failed consent decree in Plata had called for implementation of new procedures and hiring of additional staff; and the Coleman Special Master had issued over 70 orders directed at achieving a remedy through construction, hiring, and procedural reforms. The Coleman Special Master and Plata Receiver were unable to provide assurance that further, substantially similar efforts would yield success absent a population reduction. Instead, the Coleman Special Master explained that “many of the clinical advances ... painfully accomplished over the past decade are slip-sliding away” as a result of overcrowding. App. 481-482. And the Plata Receiver indicated that, absent a reduction in over*516crowding, a successful remedial effort could “all but bankrupt” the State of California. App. 1053.

Having engaged in remedial efforts for 5 years in Plata and 12 in Coleman, the District Courts were not required to wait to see whether their more recent efforts would yield equal disappointment. When a court attempts to remedy an entrenched constitutional violation through reform of a complex institution, such as this statewide prison system, it may be necessary in the ordinary course to issue multiple orders directing and adjusting ongoing remedial efforts. Each new order must be given a reasonable time to succeed, but reasonableness must be assessed in light of the entire history of the court’s remedial efforts. A contrary reading of the reasonable time requirement would in effect require district courts to impose a moratorium on new remedial orders before issuing a population limit. This unnecessary period of inaction would delay an eventual remedy and would prolong the courts’ involvement, serving neither the State nor the prisoners. Congress did not require this unreasonable result when it used the term “reasonable."

The Coleman and Plata courts had a solid basis to doubt that additional efforts to build new facilities and hire new staff would achieve a remedy. Indeed, although five years have now passed since the appointment of the Plata Receiver and approval of the revised plan of action in Coleman, there is no indication that the constitutional violations have been cured. A report filed by the Coleman Special Master in July 2009 describes ongoing violations, including an “absence of timely access to appropriate levels of care at every point in the system.” App. 807. A report filed by the Plata Receiver in October 2010 likewise describes ongoing deficiencies in the provision of medical care and concludes that there are simply “too many prisoners for the healthcare infrastructure.” App. 1655. The Coleman and Plata courts acted reasonably when they convened a three-judge court without further delay.

*517B

Once a three-judge court has been convened, the court must find additional requirements satisfied before it may impose a population limit. The first of these requirements is that “crowding is the primary cause of the violation of a Federal right.” 18 U. S. C. §3626(a)(3)(E)(i).

1

The three-judge court found the primary cause requirement satisfied by the evidence at trial. The court found that overcrowding strains inadequate medical and mental health facilities; overburdens limited clinical and custodial staff; and creates violent, unsanitary, and chaotic conditions that contribute to the constitutional violations and frustrate efforts to fashion a remedy. The three-judge court also found that “until the problem of overcrowding is overcome it will be impossible to provide constitutionally compliant care to California’s prison population.” Juris. App. 141a.

The parties dispute the standard of review applicable to this determination. With respect to the three-judge court’s factual findings, this Court’s review is necessarily deferential. It is not this Court’s place to “duplicate the role” of the trial court. Anderson, 470 U. S., at 573. The ultimate issue of primary cause presents a mixed question of law and fact; but there, too, “the mix weighs heavily on the 'fact’ side.” Lilly v. Virginia, 527 U. S. 116, 148 (1999) (Rehnquist, C. J., concurring in judgment). Because the “district court is 'better positioned’... to decide the issue,” our review of the three-judge court’s primary cause determination is deferential. Salve Regina College v. Russell, 499 U. S. 225, 233 (1991).

The record documents the severe impact of burgeoning demand on the provision of care. At the time of trial, vacancy rates for medical and mental health staff ranged as high as 20% for surgeons, 25% for physicians, 39% for nurse prac*518titioners, and 54.1% for psychiatrists. Juris. App. 105a, 108a. These percentages are based on the number of positions budgeted by the State. Dr. Ronald Shansky, former medical director of the Illinois prison system, concluded that these numbers understate the severity of the crisis because the State has not budgeted sufficient staff to meet demand.5 According to Dr. Shansky, “even if the prisons were able to fill all of their vacant health care positions, which they have not been able to do to date, . . . the prisons would still be unable to handle the level of need given the current overcrowding.” Record in No. 2:90-cv-00520-LKK-JFM (ED Cal.), Doc. 3231-13, p. 19 (hereinafter Doc. 3231-13). Dr. Craig Haney, a professor of psychology, reported that mental health staff are “managing far larger caseloads than is appropriate or effective.” App. 596. A prison psychiatrist told Dr. Haney that “ ‘we are doing about 50% of what we should be doing.’ ” Ibid. In the context of physical care Dr. Shansky agreed that “demand for care, particularly for the high priority cases, continues to overwhelm the resources available.” Id., at 1408.

Even on the assumption that vacant positions could be filled, the evidence suggested there would be insufficient space for the necessary additional staff to perform their jobs. The Plata Receiver, in his report on overcrowding, concluded that even the “newest and most modern prisons” had been “designed with clinic space which is only one-half that necessary for the real-life capacity of the prisons.” App. 1023 (emphasis deleted). Dr. Haney reported that “[ejach one of the facilities I toured was short of significant amounts of space needed to perform otherwise critical tasks and re*519sponsibilities.” Id., at 597-598. In one facility, staff cared for 7,525 prisoners in space designed for one-third as many. Juris. App. 93a. Staff operate out of converted storage rooms, closets, bathrooms, shower rooms, and visiting centers. These makeshift facilities impede the effective delivery of care and place the safety of medical professionals in jeopardy, compounding the difficulty of hiring additional staff.

This shortfall of resources relative to demand contributes to significant delays in treatment. Mentally ill prisoners are housed in administrative segregation while awaiting transfer to scarce mental health treatment beds for appropriate care. One correctional officer indicated that he had kept mentally ill prisoners in segregation for ‘“6 months or more.’” App. 594. Other prisoners awaiting care are held in tiny, phone-booth-sized cages. The record documents instances of prisoners committing suicide while awaiting treatment.6

Delays are no less severe in the context of physical care. Prisons have backlogs of up to 700 prisoners waiting to see a doctor. Doc. 3231-13, at 21. A review of referrals for urgent specialty care at one prison revealed that only 105 of 316 pending referrals had a scheduled appointment, and only 2 had an appointment scheduled to occur within 14 days. Id., at 25-26. Urgent specialty referrals at one prison had been pending for six months to a year. Id., at 30.

Crowding also creates unsafe and unsanitary living conditions that hamper effective delivery of medical and mental health care. A medical expert described living quarters in converted gymnasiums or dayrooms, where large numbers of prisoners may share just a few toilets and showers, as *520“‘breeding grounds for disease.’”7 Juris. App. 102a. Cramped conditions promote unrest and violence, making it difficult for prison officials to monitor and control the prison population. On any given day, prisoners in the general prison population may become ill, thus entering the plaintiff class; and overcrowding may prevent immediate medical attention necessary to avoid suffering, death, or spread of disease. After one prisoner was assaulted in a crowded gymnasium, prison staff did not even learn of the injury until the prisoner had been dead for several hours. Tr. 382. Living in crowded, unsafe, and unsanitary conditions can cause prisoners with latent mental illnesses to worsen and develop overt symptoms. Crowding may also impede efforts to improve delivery of care. Two prisoners committed suicide by hanging after being placed in cells that had been identified as requiring a simple fix to remove attachment points that could support a noose. The repair was not made because doing so would involve removing prisoners from the cells, and there was no place to put them. Id., at 769-777. More generally, Jeanne Woodford, the former acting secretary of California’s prisons, testified that there “‘are simply too many issues that arise from such a large number of prisoners,’” and that, as a result, “‘management spends virtually all of its time fighting fires instead of engaging in thoughtful decision-making and planning’” of the sort needed to fashion an effective remedy for these constitutional violations. Juris. App. 82a.

*521Increased violence also requires increased reliance on lock-downs to keep order, and lockdowns further impede the effective delivery of care. In 2006, prison officials instituted 449 lockdowns. Id., at 116a. The average lockdown lasted 12 days, and 20 lockdowns lasted 60 days or longer. Ibid. During lockdowns, staff must either escort prisoners to medical facilities or bring medical staff to the prisoners. Either procedure puts additional strain on already overburdened medical and custodial staff. Some programming for the mentally ill even may be canceled altogether during lock-downs, and staff may be unable to supervise the delivery of psychotropic medications.

The effects of overcrowding are particularly acute in the prisons’ reception centers, intake areas that process 140,000 new or returning prisoners every year. Id., at 85a. Crowding in these areas runs as high as 300% of design capacity. Id., at 86a. Living conditions are “‘toxic,’” and a lack of treatment space impedes efforts to identify inmate medical or mental health needs and provide even rudimentary care. Id., at 92a. The former warden of San Quentin reported that doctors in that prison’s reception center “ ‘were unable to keep up with physicals or provid[e] any kind of chronic care follow-up.’ ” Id., at 90a. Inmates spend long periods of time in these areas awaiting transfer to the general population. Some prisoners are held in the reception centers for their entire period of incarceration.

Numerous experts testified that crowding is the primary cause of the constitutional violations. The former warden of San Quentin and former acting secretary of the California prisons concluded that crowding “makes it ‘virtually impossible for the organization to develop, much less implement, a plan to provide prisoners with adequate care.’ ” Id., at 83a. The former executive director of the Texas Department of Criminal Justice testified that “‘[e] very thing revolves around overcrowding’ ” and that “ ‘overcrowding is the primary cause of the medical and mental health care viola*522tions.’” Id., at 127a. The former head of corrections in Pennsylvania, Washington, and Maine testified that overcrowding is “ 'overwhelming the system both in terms of sheer numbers, in terms of the space available, in terms of providing healthcare.'” Ibid. And the current secretary of the Pennsylvania Department of Corrections testified that “ “the biggest inhibiting factor right now in California being able to deliver appropriate mental health and medical care is the severe overcrowding.'” Id., at 82a.

2

The State attempts to undermine the substantial evidence presented at trial, and the three-judge court's findings of fact, by complaining that the three-judge court did not allow it to present evidence of current prison conditions. This suggestion lacks a factual basis.

The three-judge court properly admitted evidence of current conditions as relevant to the issues before it. The three-judge court allowed discovery until a few months before trial; expert witnesses based their conclusions on recent observations of prison conditions; the court admitted recent reports on prison conditions by the Plata Receiver and Coleman Special Master; and both parties presented testimony related to current conditions, including understaffing, inadequate facilities, and unsanitary and unsafe living conditions. See supra, at 504-507, 517-521 and this page. Dr. Craig Haney, for example, based his expert report on tours of eight California prisons. App. 539. These tours occurred as late as August 2008, two weeks before Dr. Haney submitted his report and less than four months before the first day of trial. Id., at 585; see also id., at 563, 565, 580 (July tours). Other experts submitted reports based on similar observations. See, e. g., Doc. 3231-13, at 9 (Dr. Shansky); App. 646 (Dr. Stewart); id., at 1245 (Austin); id., at 1312 (Lehman).

The three-judge court's opinion cited and relied on this evidence of current conditions. The court relied extensively *523on the expert witness reports. See generally Juris. App. 85a-143a. The court cited the most current data available on suicides and preventable deaths in the California prisons. Id., at 123a, 125a. The court relied on statistics on staff vacancies that dated to three months before trial, id., at 105a, 108a, and statistics on shortages of treatment beds for the same period, id., at 97a. These are just examples of the extensive evidence of current conditions that informed every aspect of the judgment of the three-judge court. The three-judge court did not abuse its discretion when it also cited findings made in earlier decisions of the Plata and Coleman District Courts. Those findings remained relevant to establish the nature of these longstanding, continuing constitutional violations.

It is true that the three-judge court established a cutoff date for discovery a few months before trial. The order stated that site inspections of prisons would be allowed until that date, and that evidence of “changed prison conditions” after that date would not be admitted. App. 1190. The court also excluded evidence not pertinent to the issue whether a population limit is appropriate under the PLRA, including evidence relevant solely to the existence of an ongoing constitutional violation. The court reasoned that its decision was limited to the issue of remedy and that the merits of the constitutional violation had already been determined. The three-judge court made clear that all such evidence would be considered “[t]o the extent that it illuminates questions that are properly before the court.” Id., at 2339.

Both rulings were within the sound discretion of the three-judge court. Orderly trial management may require discovery deadlines and a clean distinction between litigation of the merits and the remedy. The State in fact represented to the three-judge court that it would be “appropriate” to cut off discovery before trial because “like plaintiffs, we, too, are really gearing up and going into a pretrial mode.” Id., at 1683. And if the State truly believed there was no longer *524a violation, it could have argued to the Coleman and Plata District Courts that a three-judge court should not be convened because the District Courts’ prior orders had not "failed to remedy the deprivation” of prisoners’ constitutional rights. 18 U. S. C. § 3626(a)(3)(A)(i); see also supra, at 514-515. Once the three-judge court was convened, that court was not required to reconsider the merits. Its role was solely to consider the propriety and necessity of a population limit.

The State does not point to any significant evidence that it was unable to present and that would have changed the outcome of the proceedings. To the contrary, the record and opinion make clear that the decision of the three-judge court was based on current evidence pertaining to ongoing constitutional violations.

3

The three-judge court acknowledged that the violations were caused by factors in addition to overcrowding and that reducing crowding in the prisons would not entirely cure the violations. This is consistent with the reports of the Coleman Special Master and Plata Receiver, both of whom concluded that even a significant reduction in the prison population would not remedy the violations absent continued efforts to train staff, improve facilities, and reform procedures. App. 487, 1054.8 The three-judge court nevertheless found *525that overcrowding was the primary cause in the sense of being the foremost cause of the violation.

This understanding of the primary cause requirement is consistent with the text of the PLRA. The State in fact concedes that it proposed this very definition of primary cause to the three-judge court. “Primary” is defined as “[f]irst or highest in rank, quality, or importance; principal.” American Heritage Dictionary 1393 (4th ed. 2000); see also Webster’s Third New International Dictionary 1800 (2002) (defining “primary” as “first in rank or importance”); 12 Oxford English Dictionary 472 (2d ed. 1989) (defining “primary” as “[o]f the first or highest rank or importance; that claims the first consideration; principal, chief”). Overcrowding need only be the foremost, chief, or principal cause of the violation. If Congress had intended to require that crowding be the only cause, it would have said so, assuming in its judgment that definition would be consistent with constitutional limitations.

As this case illustrates, constitutional violations in conditions of confinement are rarely susceptible of simple or straightforward solutions. In addition to overcrowding the failure of California’s prisons to provide adequate medical and mental health care may be ascribed to chronic and worsening budget shortfalls, a lack of political will in favor of reform, inadequate facilities, and systemic administrative failures. The Plata District Judge, in his order appointing the Receiver, compared the problem to “‘a spider web, in which the tension of the various strands is determined by the relationship among all the parts of the web, so that if one pulls on a single strand, the tension of the entire web is redistributed in a new and complex pattern.’” App. 966-967 (quoting Fletcher, The Discretionary Constitution: Institutional Remedies and Judicial Legitimacy, 91 Yale L. J. 635, 645 (1982)); see also Hutto, 437 U. S., at 688 (noting “the interdependence of the conditions producing the violation,” in-*526eluding overcrowding). Only a multifaceted approach aimed at many causes, including overcrowding, will yield a solution.

The PLRA should not be interpreted to place undue restrictions on the authority of federal courts to fashion practical remedies when confronted with complex and intractable constitutional violations. Congress limited the availability of limits on prison populations, but it did not forbid these measures altogether. See 18 U. S. C. § 3626. The House Report accompanying the PLRA explained:

“While prison caps must be the remedy of last resort, a court still retains the power to order this remedy despite its intrusive nature and harmful consequences to the public if, but only if, it is truly necessary to prevent an actual violation of a prisoner’s federal rights.” H. R. Rep. No. 104-21, p. 25 (1995).

Courts should presume that Congress was sensitive to the real-world problems faced by those who would remedy constitutional violations in the prisons and that Congress did not leave prisoners without a remedy for violations of their constitutional rights. A reading of the PLRA that would render population limits unavailable in practice would raise serious constitutional concerns. See, e. g., Bowen v. Michigan Academy of Family Physicians, 476 U. S. 667, 681, n. 12 (1986). A finding that overcrowding is the “primary cause” of a violation is therefore permissible, despite the fact that additional steps will be required to remedy the violation.

C

The three-judge court was also required to find by clear and convincing evidence that “no other relief will remedy the violation of the Federal right.” § 3626(a)(3)(E)(ii).

The State argues that the violation could have been remedied through a combination of new construction, transfers of prisoners out of State, hiring of medical personnel, and continued efforts by the Plata Receiver and Coleman Spe*527cial Master. The order in fact permits the State to comply with the population limit by transferring prisoners to county facilities or facilities in other States, or by constructing new facilities to raise the prisons’ design capacity. And the three-judge court’s order does not bar the State from undertaking any other remedial efforts. If the State does find an adequate remedy other than a population limit, it may seek modification or termination of the three-judge court’s order on that basis. The evidence at trial, however, supports the three-judge court’s conclusion that an order limited to other remedies would not provide effective relief.

The State’s argument that out-of-state transfers provide a less restrictive alternative to a population limit must fail because requiring out-of-state transfers itself qualifies as a population limit under the PLRA.9 Such an order “has the purpose or effect of reducing or limiting the prison population, or . . . directs the release from or nonadmission of prisoners to a prison.” § 3626(g)(4). The same is true of transfers to county facilities. Transfers provide a means to reduce the prison population in compliance with the three-judge court’s order. They are not a less restrictive alternative to that order.

Even if out-of-state transfers could be regarded as a less restrictive alternative, the three-judge court found no evidence of plans for transfers in numbers sufficient to relieve overcrowding. The State complains that the Coleman District Court slowed the rate of transfer by requiring inspections to ensure that the receiving institutions were in compliance with the Eighth Amendment, but the State has made no effort to show that it has the resources and the capacity *528to transfer significantly larger numbers of prisoners absent that condition.

Construction of new facilities, in theory, could alleviate overcrowding, but the three-judge court found no realistic possibility that California would be able to build itself out of this crisis. At the time of the court’s decision the State had plans to build new medical and housing facilities, but funding for some plans had not been secured and funding for other plans had been delayed by the legislature for years. Particularly in light of California’s ongoing fiscal crisis, the three-judge court deemed “chimerical” any “remedy that requires significant additional spending by the state.” Juris. App. 151a. Events subsequent to the three-judge court’s decision have confirmed this conclusion. In October 2010, the State notified the Coleman District Court that a substantial component of its construction plans had been delayed indefinitely by the legislature. And even if planned construction were to be completed, the Plata Receiver found that many so-called “expansion” plans called for cramming more prisoners into existing prisons without expanding administrative and support facilities. Juris. App. 151a-152a. The former acting secretary of the California prisons explained that these plans would “‘compound the burdens imposed on prison administrators and line staff'” by adding to the already overwhelming prison population, creating new barriers to achievement of a remedy. Id., at 152a.

The three-judge court also rejected additional hiring as a realistic means to achieve a remedy. The State for years had been unable to fill positions necessary for the adequate provision of medical and mental health care, and the three-judge court found no reason to expect a change. Although the State points to limited gains in staffing between 2007 and 2008, the record shows that the prison system remained chronically understaffed through trial in 2008. See supra, at 517-518. The three-judge court found that violence and other negative conditions caused by crowding made it diffi*529cult to hire and retain needed staff The court also concluded that there would be insufficient space for additional staff to work even if adequate personnel could somehow be retained. Additional staff cannot help to remedy the violation if they have no space in which to see and treat patients.

The three-judge court also did not err, much less commit clear error, when it concluded that, absent a population reduction, continued efforts by the Receiver and Special Master would not achieve a remedy. Both the Receiver and the Special Master filed reports stating that overcrowding posed a significant barrier to their efforts. The Plata Receiver stated that he was determined to achieve a remedy even without a population reduction, but he warned that such an effort would “all but bankrupt” the State. App. 1053. The Coleman Special Master noted even more serious concerns, stating that previous remedial efforts had “succumbed to the inexorably rising tide of population.” App. 489. Both reports are persuasive evidence that, absent a reduction in overcrowding, any remedy might prove unattainable and would at the very least require vast expenditures of resources by the State. Nothing in the long history of the Coleman and Plata actions demonstrates any real possibility that the necessary resources would be made available.

The State claims that, even if each of these measures were unlikely to remedy the violation, they would succeed in doing so if combined together. Aside from asserting this proposition, the State offers no reason to believe it is so. Attempts to remedy the violations in Plata have been ongoing for nine years. In Coleman, remedial efforts have been ongoing for 16. At one time, it may have been possible to hope that these violations would be cured without a reduction in overcrowding. A long history of failed remedial orders, together with substantial evidence of overcrowding’s deleterious effects on the provision of care, compels a different conclusion today.

*530The common thread connecting the State’s proposed remedial efforts is that they would require the State to expend large amounts of money absent a reduction in overcrowding. The Court cannot ignore the political and fiscal reality behind this case. California’s Legislature has not been willing or able to allocate the resources necessary to meet this crisis absent a reduction in overcrowding. There is no reason to believe it will begin to do so now, when the State of California is facing an unprecedented budgetary shortfall. As noted above, the legislature recently failed to allocate funds for planned new construction. Supra, at 528. Without a reduction in overcrowding, there will be no efficacious remedy for the unconstitutional care of the sick and mentally ill in California’s prisons.

D

The PLRA states that no prospective relief shall issue with respect to prison conditions unless it is narrowly drawn, extends no further than necessary to correct the violation of a federal right, and is the least intrusive means necessary to correct the violation. 18 U. S. C. § 3626(a). When determining whether these requirements are met, courts must “give substantial weight to any adverse impact on public safety or the operation of a criminal justice system.” Ibid.

1

The three-judge court acknowledged that its order “is likely to affect inmates without medical conditions or serious mental illness." Juris. App. 172a. This is because reducing California’s prison population will require reducing the number of prisoners outside the class through steps such as parole reform, sentencing reform, use of good-time credits, or other means to be determined by the State. Reducing overcrowding will also have positive effects beyond facilitating timely and adequate access to medical care, including reducing the incidence of prison violence and ameliorating unsafe living conditions. According to the State, these collateral *531consequences are evidence that the order sweeps more broadly than necessary.

The population limit imposed by the three-judge court does not fail narrow tailoring simply because it will have positive effects beyond the plaintiff class. Narrow tailoring requires a “ ‘ “fit” between the [remedy’s] ends and the means chosen to accomplish those ends.’” Board of Trustees of State Univ. of N. Y. v. Fox, 492 U. S. 469, 480 (1989). The scope of the remedy must be proportional to the scope of the violation, and the order must extend no further than necessary to remedy the violation. This Court has rejected remedial orders that unnecessarily reach out to improve prison conditions other than those that violate the Constitution. Lewis v. Casey, 518 U. S. 343, 357 (1996). But the precedents do not suggest that a narrow and otherwise proper remedy for a constitutional violation is invalid simply because it will have collateral effects.

Nor does anything in the text of the PLRA require that result. The PLRA states that a remedy shall extend no further than necessary to remedy the violation of the rights of a “particular plaintiff or plaintiffs.” 18 U. S. C. § 3626(a)(1)(A). This means only that the scope of the order must be determined with reference to the constitutional violations established by the specific plaintiffs before the court.

This case is unlike cases where courts have impermissibly reached out to control the treatment of persons or institutions beyond the scope of the violation. See Dayton Bd. of Ed. v. Brinkman, 433 U. S. 406, 420 (1977). Even prisoners with no present physical or mental illness may become afflicted, and all prisoners in California are at risk so long as the State continues to provide inadequate care. Prisoners in the general population will become sick, and will become members of the plaintiff classes, with routine frequency; and overcrowding may prevent the timely diagnosis and care necessary to provide effective treatment and to prevent fur*532ther spread of disease. Relief targeted only at present members of the plaintiff classes may therefore fail to adequately protect future class members who will develop serious physical or mental illness. Prisoners who are not sick or mentally ill do not yet have a claim that they have been subjected to care that violates the Eighth Amendment, but in no sense are they remote bystanders in California’s medical care system. They are that system’s next potential victims.

A release order limited to prisoners within the plaintiff classes would, if anything, unduly limit the ability of state officials to determine which prisoners should be released. As the State acknowledges in its brief, “release of seriously mentally ill inmates [would be] likely to create special dangers because of their recidivism rates." Consolidated Reply Brief for Appellants 34. The order of the three-judge court gives the State substantial flexibility to determine who should be released. If the State truly believes that a release order limited to sick and mentally ill inmates would be preferable to the order entered by the three-judge court, the State can move the three-judge court for modification of the order on that basis. The State has not requested this relief from this Court.

The order also is not overbroad because it encompasses the entire prison system, rather than separately assessing the need for a population limit at every institution. The Coleman court found a systemwide violation when it first afforded relief, and in Plata the State stipulated to system-wide relief when it conceded the existence of a violation. Both the Coleman Special Master and the Plata Receiver have filed numerous reports detailing systemwide deficiencies in medical and mental health care. California’s medical care program is run at a systemwide level, and resources are shared among the correctional facilities.

Although the three-judge court’s order addresses the entire California prison system, it affords the State flexibility *533to accommodate differences between institutions. There is no requirement that every facility comply with the 137.5% limit. Assuming no constitutional violation results, some facilities may retain populations in excess of the limit provided other facilities fall sufficiently below it so the system as a whole remains in compliance with the order. This will allow prison officials to shift prisoners to facilities that are better able to accommodate overcrowding, or out of facilities where retaining sufficient medical staff has been difficult. The alternative — a series of institution-specific population limits — would require federal judges to make these choices. Leaving this discretion to state officials does not make the order overbroad.

Nor is the order overbroad because it limits the State’s authority to run its prisons, as the State urges in its brief. While the order does in some respects shape or control the State’s authority in the realm of prison administration, it does so in a manner that leaves much to the State’s discretion. The State may choose how to allocate prisoners between institutions; it may choose whether to increase the prisons’ capacity through construction or reduce the population; and, if it does reduce the population, it may decide what steps to take to achieve the necessary reduction. The order’s limited scope is necessary to remedy a constitutional violation.

As the State implements the order of the three-judge court, time and experience may reveal targeted and effective remedies that will end the constitutional violations even without a significant decrease in the general prison population. The State will be free to move the three-judge court for modification of its order on that basis, and these motions would be entitled to serious consideration. See infra, at 543-545. At this time, the State has not proposed any realistic alternative to the order. The State’s desire to avoid a population limit, justified as according respect to state authority, creates a certain and unacceptable risk of continuing *534violations of the rights of sick and mentally ill prisoners, with the result that many more will die or needlessly suffer. The Constitution does not permit this wrong.

2

In reaching its decision, the three-judge court gave “substantial weight” to any potential adverse impact on public safety from its order. The court devoted nearly 10 days of trial to the issue of public safety, and it gave the question extensive attention in its opinion. Ultimately, the court concluded that it would be possible to reduce the prison population “in a manner that preserves public safety and the operation of the criminal justice system.” Juris. App. 247a-248a.

The PLRA’s requirement that a court give “substantial weight” to public safety does not require the court to certify that its order has no possible adverse impact on the public. A contrary reading would depart from the statute’s text by replacing the word “substantial” with “conclusive.” Whenever a eourt issues an order requiring the State to adjust its incarceration and criminal justice policy, there is a risk that the order will have some adverse impact on public safety in some sectors. This is particularly true when the order requires release of prisoners before their sentence has been served. Persons incarcerated for even one offense may have committed many other crimes prior to arrest and conviction, and some number can be expected to commit further crimes upon release. Yet the PLRA contemplates that courts will retain authority to issue orders necessary to remedy constitutional violations, including authority to issue population limits when necessary. See supra, at 527. A court is required to consider the public safety consequences of its order and to structure, and monitor, its ruling in a way that mitigates those consequences while still achieving an effective remedy of the constitutional violation.

*535This inquiry necessarily involves difficult predictive judgments regarding the likely effects of court orders. Although these judgments are normally made by state officials, they necessarily must be made by courts when those courts fashion injunctive relief to remedy serious constitutional violations in the prisons. These questions are difficult and sensitive, but they are factual questions and should be treated as such. Courts can, and should, rely on relevant and informed expert testimony when making factual findings. It was proper for the three-judge court to rely on the testimony of prison officials from California and other States. Those experts testified on the basis of empirical evidence and extensive experience in the field of prison administration.

The three-judge court credited substantial evidence that prison populations can be reduced in a manner that does not increase crime to a significant degree. Some evidence indicated that reducing overcrowding in California’s prisons could even improve public safety. Then-Governor Schwarzenegger, in his emergency proclamation on overcrowding, acknowledged that “‘overcrowding causes harm to people and property, leads to inmate unrest and misconduct, . . . and increases recidivism as shown within this state and in others.’” Juris. App. 191a-192a. The former warden of San Quentin and acting secretary of the California prison system testified that she “‘absolutely believe[s] that we make people worse, and that we are not meeting public safety by the way we treat people.’ ”10 Id., at 129a. And the head of *536Pennsylvania’s correctional system testified that measures to reduce prison population may “actually improve on public safety because they address the problems that brought people to jail.” Tr. 1552-1553.

Expert witnesses produced statistical evidence that prison populations had been lowered without adversely affecting public safety in a number of jurisdictions, including certain counties in California, as well as Wisconsin, Illinois, Texas, Colorado, Montana, Michigan, Florida, and Canada. Juris. App. 245a.11 Washington’s former secretary of corrections testified that his State had implemented population-reduction methods, including parole reform and expansion of good-time credits, without any “deleterious effect on crime.” Tr. 2008-2009. In light of this evidence, the three-judge court concluded that any negative impact on public safety *537would be “substantially offset, and perhaps entirely eliminated, by the public safety benefits” of a reduction in overcrowding. Juris. App. 248a.

The court found that various available methods of reducing overcrowding would have little or no impact on public safety. Expansion of good-time credits would allow the State to give early release to only those prisoners who pose the least risk of reoffending. Diverting low-risk offenders to community programs such as drug treatment, day reporting centers, and electronic monitoring would likewise lower the prison population without releasing violent convicts.12 The State now sends large numbers of persons to prison for violating a technical term or condition of their parole, and it could reduce the prison population by punishing technical parole violations through community-based programs. This last measure would be particularly beneficial as it would reduce crowding in the reception centers, which are especially hard hit by overcrowding. See supra, at 521. The court’s order took account of public safety concerns by giving the State substantial flexibility to select among these and other means of reducing overcrowding.

The State submitted a plan to reduce its prison population in accordance with the three-judge court’s order, and it complains that the three-judge court approved that plan without considering whether the specific measures contained within it would substantially threaten public safety. The three-judge court, however, left the choice of how best to comply with its population limit to state prison officials. The court *538was not required to second-guess the exercise of that discretion. Courts should presume that state officials are in a better position to gauge how best to preserve public safety and balance competing correctional and law enforcement concerns. The decision to leave details of implementation to the State’s discretion protected public safety by leaving sensitive policy decisions to responsible and competent state officials.

During the pendency of this appeal, the State in fact began to implement measures to reduce the prison population. See Supp. Brief for Appellants 1. These measures will shift “thousands” of prisoners from the state prisons to the county jails by “mak[ing] certain felonies punishable by imprisonment in county jail” and “requiring] that individuals returned to custody for violating their conditions of parole ‘serve any custody term in county jail.’” Ibid. These developments support the three-judge court's conclusion that the prison population can be reduced in a manner calculated to avoid an undue negative effect on public safety.

Ill

Establishing the population at which the State could begin to provide constitutionally adequate medical and mental health care, and the appropriate timeframe within which to achieve the necessary reduction, requires a degree of judgment. The inquiry involves uncertain predictions regarding the effects of population reductions, as well as difficult determinations regarding the capacity of prison officials to provide adequate care at various population levels. Courts have substantial flexibility when making these judgments. “ ‘Once invoked, “the scope of a district court’s equitable powers ... is broad, for breadth and flexibility are inherent in equitable remedies.” ’ ” Hutto, 437 U. S., at 687, n. 9 (quoting Milliken v. Bradley, 433 U. S. 267, 281 (1977), in turn quoting Swann v. Charlotte-Mecklenburg Bd. of Ed., 402 U. S. 1, 15 (1971)).

*539Nevertheless, the PLEA requires a court to adopt a remedy that is “narrowly tailored” to the constitutional violation and that gives “substantial weight” to public safety. 18 U. S. C. § 3626(a). When a court is imposing a population limit, this means the court must set the limit at the highest population consistent with an efficacious remedy. The court must also order the population reduction achieved in the shortest period of time reasonably consistent with public safety.

A

The three-judge court concluded that the population of California’s prisons should be capped at 137.5% of design capacity. This conclusion is supported by the record. Indeed, some evidence supported a limit as low as 100% of design capacity. The chief deputy secretary of Correctional Healthcare Services for the California prisons testified that California’s prisons “ ‘were not designed and made no provision for any expansion of medical care space beyond the initial 100% of capacity.’” Juris. App. 176a. Other evidence supported a limit as low as 130%. The head of the State’s Facilities Strike Team recommended reducing the population to 130% of design capacity as a long-term goal. Id., at 179a-180a. A former head of correctional systems in Washington State, Maine, and Pennsylvania testified that a 130% limit would “ ‘give prison officials and staff the ability to provide the necessary programs and services for California’s prisoners.’” Id., at 180a. A former executive director of the Texas prisons testified that a limit of 130% was “ ‘realistic and appropriate’” and would “‘ensure that [California’s] prisons are safe and provide legally required services.’ ” Ibid. And a former acting secretary of the California prisons agreed with a 130% limit with the caveat that a 130% limit might prove inadequate in some older facilities. Ibid.

According to the State, this testimony expressed the witnesses’ policy preferences, rather than their views as to what would cure the constitutional violation. Of course, courts *540must not confuse professional standards with constitutional requirements. Rhodes v. Chapman, 452 U. S. 337, 348, n. 13 (1981). But expert opinion may be relevant when determining what is obtainable and what is acceptable in corrections philosophy. See supra, at 536-537. Nothing in the record indicates that the experts in this case imposed their own policy views or lost sight of the underlying violations. To the contrary, the witnesses testified that a 130% population limit would allow the State to remedy the constitutionally inadequate provision of medical and mental health care. When expert opinion is addressed to the question of how to remedy the relevant constitutional violations, as it was here, federal judges can give it considerable weight.

The Federal Bureau of Prisons (BOP) has set 130% as a long-term goal for population levels in the federal prison system. Brief for Appellants 43-44. The State suggests the expert witnesses impermissibly adopted this professional standard in their testimony. But courts are not required to disregard expert opinion solely because it adopts or accords with professional standards. Professional standards may be “helpful and relevant with respect to some questions.” Chapman, supra, at 348, n. 13. The witnesses testified that a limit of 130% was necessary to remedy the constitutional violations, not that it should be adopted because it is a BOP standard. If anything, the fact that the BOP views 130% as a manageable population density bolsters the three-judge court’s conclusion that a population limit of 130% would alleviate the pressures associated with overcrowding and allow the State to begin to provide constitutionally adequate care.

Although the three-judge court concluded that the “evidence in support of a 130% limit is strong,” it found that some upward adjustment was warranted in light of “the caution and restraint required by the PLRA.” Juris. App. 183a, 184a. The three-judge court noted evidence supporting a higher limit. In particular, the State’s Corrections Independent Review Panel had found that 145% was the maxi*541mum “operable capacity” of California’s prisons, id., at 181a-182a, although the relevance of that determination was undermined by the fact that the panel had not considered the need to provide constitutionally adequate medical and mental health care, as the State itself concedes. Brief for Coleman Appellees 45. After considering, but discounting, this evidence, the three-judge court concluded that the evidence supported a limit lower than 145%, but higher than 130%. It therefore imposed a limit of 137.5%.

This weighing of the evidence was not clearly erroneous. The adversary system afforded the court an opportunity to weigh and evaluate evidence presented by the parties. The plaintiffs’ evidentiary showing was intended to justify a limit of 130%, and the State made no attempt to show that any other number would allow for a remedy. There are also no scientific tools available to determine the precise population reduction necessary to remedy a constitutional violation of this sort. The three-judge court made the most precise determination it could in light of the record before it. The PLRA’s narrow tailoring requirement is satisfied so long as these equitable, remedial judgments are made with the objective of releasing the fewest possible prisoners consistent with an efficacious remedy. In light of substantial evidence supporting an even more drastic remedy, the three-judge court complied with the requirement of the PLRA in this case.

B

The three-judge court ordered the State to achieve this reduction within two years. At trial and closing argument before the three-judge court, the State did not argue that reductions should occur over a longer period of time. The State later submitted a plan for court approval that would achieve the required reduction within five years, and that would reduce the prison population to 151% of design capacity in two years. The State represented that this plan would “safely reach a population level of 137.5% over time.” *542Juris. App. 317a. The three-judge court rejected this plan because it did not comply with the deadline set by its order.

The State first had notice that it would be required to reduce its prison population in February 2009, when the three-judge court gave notice of its tentative ruling after trial. The 2-year deadline, however, will not begin to run until this Court issues its judgment. When that happens, the State will have already had over two years to begin complying with the order of the three-judge court. The State has used the time productively. At oral argument, the State indicated it had reduced its prison population by approximately 9,000 persons since the decision of the three-judge court. After oral argument, the State filed a supplemental brief indicating that it had begun to implement measures to shift “thousands” of additional prisoners to county facilities. Supp. Brief for Appellants 1.

Particularly in light of the State’s failure to contest the issue at trial, the three-judge court did not err when it established a 2-year deadline for relief. Plaintiffs proposed a 2-year deadline, and the evidence at trial was intended to demonstrate the feasibility of a 2-year deadline. See Tr. 2979. Notably, the State has not asked this Court to extend the 2-year deadline at this time.

The three-judge court, however, retains the authority, and the responsibility, to make further amendments to the existing order or any modified decree it may enter as warranted by the exercise of its sound discretion. “The power of a court of equity to modify a decree of injunctive relief is long-established, broad, and flexible.” New York State Assn. for Retarded Children, Inc. v. Carey, 706 F. 2d 956, 967 (CA2 1983) (Friendly, J.). A court that invokes equity’s power to remedy a constitutional violation by an injunction mandating systemic changes to an institution has the continuing duty and responsibility to assess the efficacy and consequences of its order. Id., at 969-971. Experience may teach the necessity for modification or amendment of an earlier decree. *543To that end, the three-judge court must remain open to a showing or demonstration by either party that the injunction should be altered to ensure that the rights and interests of the parties are given all due and necessary protection.

Proper respect for the State and for its governmental processes requires that the three-judge court exercise its jurisdiction to accord the State considerable latitude to find mechanisms and make plans to correct the violations in a prompt and effective way consistent with public safety. In order to “give substantial weight to any adverse impact on public safety,” 18 U. S. C. § 3626(a)(1)(A), the three-judge court must give due deference to informed opinions as to what public safety requires, including the considered determinations of state officials regarding the time in which a reduction in the prison population can be achieved consistent with public safety. An extension of time may allow the State to consider changing political, economic, and other circumstances and to take advantage of opportunities for more effective remedies that arise as the Special Master, the Receiver, the prison system, and the three-judge court itself evaluate the progress being made to correct unconstitutional conditions. At the same time, both the three-judge court and state officials must bear in mind the need for a timely and efficacious remedy for the ongoing violation of prisoners’ constitutional rights.

The State may wish to move for modification of the three-judge court’s order to extend the deadline for the required reduction to five years from the entry of the judgment of this Court, the deadline proposed in the State’s first population-reduction plan. The three-judge court may grant such a request provided that the State satisfies necessary and appropriate preconditions designed to ensure that measures are taken to implement the plan without undue delay. Appropriate preconditions may include a requirement that the State demonstrate that it has the authority and the resources necessary to achieve the required reduction within *544a 5-year period and to meet reasonable interim directives for population reduction. The three-judge court may also condition an extension of time on the State’s ability to meet interim benchmarks for improvement in provision of medical and mental health care.

The three-judge court, in its discretion, may also consider whether it is appropriate to order the State to begin without delay to develop a system to identify prisoners who are unlikely to reoffend or who might otherwise be candidates for early release. Even with an extension of time to construct new facilities and implement other reforms, it may become necessary to release prisoners to comply with the court’s order. To do so safely, the State should devise systems to select those prisoners least likely to jeopardize public safety. An extension of time may provide the State a greater opportunity to refine and elaborate those systems.

The State has already made significant progress toward reducing its prison population, including reforms that will result in shifting “thousands” of prisoners to county jails. See Supp. Brief for Appellants 1. As the State makes further progress, the three-judge court should evaluate whether its order remains appropriate. If significant progress is made toward remedying the underlying constitutional violations, that progress may demonstrate that further population reductions are not necessary or are less urgent than previously believed. Were the State to make this showing, the three-judge court in the exercise of its discretion could consider whether it is appropriate to extend or modify this timeline.

Experience with the three-judge court’s order may also lead the State to suggest other modifications. The three-judge court should give any such requests serious consideration. The three-judge court should also formulate its orders to allow the State and its officials the authority necessary to address contingencies that may arise during the remedial process.

*545These observations reflect the fact that the three-judge court’s order, like all continuing equitable decrees, must remain open to appropriate modification. They are not intended to cast doubt on the validity of the basic premise of the existing order. The medical and mental health care provided by California’s prisons falls below the standard of decency that inheres in the Eighth Amendment. This extensive and ongoing constitutional violation requires a remedy, and a remedy will not be achieved without a reduction in overcrowding. The relief ordered by the three-judge court is required by the Constitution and was authorized by Congress in the PLRA. The State shall implement the order without further delay.

The judgment of the three-judge court is affirmed.

It is so ordered.

APPENDIXES

A

18 U. S. C. §8626:

“(a) Requirements for Relief.—

“(1) Prospective relief. — (A) Prospective relief in any civil action with respect to prison conditions shall extend no further than necessary to correct the violation of the Federal right of a particular plaintiff or plaintiffs. The court shall not grant or approve any prospective relief unless the court finds that such relief is narrowly drawn, extends no further than necessary to correct the violation of the Federal right, and is the least intrusive means necessary to correct the violation of the Federal right. The court shall give substantial weight to any adverse impact on public safety or the operation of a criminal justice system caused by the relief.

“(3) Prisoner release order. — (A) In any civil action with respect to prison conditions, no court shall enter a prisoner release order unless—

*546“(i) a court has previously entered an order for less intrusive relief that has failed to remedy the deprivation of the Federal right sought to be remedied through the prisoner release order; and

“(ii) the defendant has had a reasonable amount of time to comply with the previous court orders.

“(B) In any civil action in Federal court with respect to prison conditions, a prisoner release order shall be entered only by a three-judge court in accordance with section 2284 of title 28, if the requirements of subparagraph (E) have been met.

“(C) A party seeking a prisoner release order in Federal court shall file with any request for such relief, a request for a three-judge court and materials sufficient to demonstrate that the requirements of subparagraph (A) have been met.

“(D) If the requirements under subparagraph (A) have been met, a Federal judge before whom a civil action with respect to prison conditions is pending who believes that a prison release order should be considered may sua sponte request the convening of a three-judge court to determine whether a prisoner release order should be entered.

“(E) The three-judge court shall enter a prisoner release order only if the court finds by clear and convincing evidence that—

“(i) crowding is the primary cause of the violation of a Federal right; and

“(ii) no other relief will remedy the violation of the Federal right.

“(F) Any State or local official including a legislator or unit of government whose jurisdiction or function includes the appropriation of funds for the construction, operation, or maintenance of prison facilities, or the prosecution or custody of persons who may be released from, or not admitted to, a prison as a result of a prisoner release order shall have standing to oppose the imposition or continuation in effect of such relief and to seek termination of such relief, and shall *547have the right to intervene in any proceeding relating to such relief.

“(g) Definitions. — As used in this section—

“(4) the term ‘prisoner release order’ includes any order, including a temporary restraining order or preliminary injunctive relief, that has the purpose or effect of reducing or limiting the prison population, or that directs the release from or nonadmission of prisoners to a prison ....”

[Appendix B is on p. 548.]

*548

*549

*550

*551

*550Justice Scalia,

with whom Justice Thomas joins, dissenting.

Today the Court affirms what is perhaps the most radical injunction issued by a court in our Nation’s history: an order requiring California to release the staggering number of 46,000 convicted criminals.

There comes before us, now and then, a case whose proper outcome is so clearly indicated by tradition and common sense, that its decision ought to shape the law, rather than viee versa. One would think that, before allowing the decree of a federal district court to release 46,000 convicted felons, this Court would bend every effort to read the law in such a way as to avoid that outrageous result. Today, quite to the contrary, the Court disregards stringently drawn provisions of the governing statute, and traditional constitutional limitations upon the power of a federal judge, in order to uphold the absurd.

The proceedings that led to this result were a judicial travesty. I dissent because the institutional reform the District Court has undertaken violates the terms of the governing statute, ignores bedrock limitations on the power of Article III judges, and takes federal courts wildly beyond their institutional capacity.

I

A

The Prison Litigation Reform Act of 1995 (PLRA) states that “[prospective relief in any civil action with respect to prison conditions shall extend no further than necessary to correct the violation of the Federal right of a particular plaintiff or plaintiffs”; that such relief must be “narrowly drawn, [and] exten[d] no further than necessary to correct the violation of the Federal right”; and that it must be "the least intrusive means necessary to correct the violation of the Federal right.” 18 U. S. C. § 3626(a)(1)(A). In deciding whether these multiple limitations have been complied with, *551it is necessary to identify with precision what is the “violation of the Federal right of a particular plaintiff or plaintiffs” that has been alleged. What has been alleged here, and what the injunction issued by the court is tailored (narrowly or not) to remedy, is the running of a prison system with inadequate medical facilities. That may result in the denial of needed medical treatment to “a particular [prisoner] or [prisoners],” thereby violating (according to our cases) his or their Eighth Amendment rights. But the mere existence of the inadequate system does not subject to cruel and unusual punishment the entire prison population in need of medical care, including those who receive it.

The Court acknowledges that the plaintiffs “do not base their case on deficiencies in care provided on any one occasion”; rather, “[p]laintiffs rely on systemwide deficiencies in the provision of medical and mental health care that, taken as a whole, subject sick and mentally ill prisoners in California to ‘substantial risk of serious harm’ and cause the delivery of care in the prisons to fall below the evolving standards of decency that mark the progress of a maturing society.” Ante, at 505, n. 3. But our judge-empowering “evolving standards of decency” jurisprudence (with which, by the way, I heartily disagree, see, e. g., Roper v. Simmons, 543 U. S. 551, 615-616 (2005) (Scalia, J., dissenting)) does not prescribe (or at least has not until today prescribed) rules for the “decent” running of schools, prisons, and other government institutions. It forbids “indecent” treatment of individuals — in the context of this case, the denial of medical care to those who need it. And the persons who have a constitutional claim for denial of medical care are those who are denied medical care — not all who face a “substantial risk” (whatever that is) of being denied medical care.

The Coleman litigation involves “the class of seriously mentally ill persons in California prisons,” ante, at 506, and the Plata litigation involves “the class of state prisoners with serious medical conditions,” ante, at 507. The plaintiffs *552do not appear to claim — and it would be absurd to suggest— that every single one of those prisoners has personally experienced “torture or a lingering death,” ante, at 510 (internal quotation marks omitted), as a consequence of that bad medical system. Indeed, it is inconceivable that anything more than a small proportion of prisoners in the plaintiff classes have personally received sufficiently atrocious treatment that their Eighth Amendment right was violated — which, as the Court recognizes, is why the plaintiffs do not premise their claim on “deficiencies in care provided on any one occasion.” Ante, at 505, n. 3. Rather, the plaintiffs’ claim is that they are all part of a medical system so defective that some number of prisoners will inevitably be injured by incompetent medical care, and that this number is sufficiently high so as to render the system, as a whole, unconstitutional.

But what procedural principle justifies certifying a class of plaintiffs so they may assert a claim of systemic unconstitutionality? I can think of two possibilities, both of which are untenable. The first is that although some or most plaintiffs in the class do not individually have viable Eighth Amendment claims, the class as a whole has collectively suffered an Eighth Amendment violation. That theory is contrary to the bedrock rule that the sole purpose of classwide adjudication is to aggregate claims that are individually viabla “A class action, no less than traditional joinder (of which it is a species), merely enables a federal court to adjudicate claims of multiple parties at once, instead of in separate suits. And like traditional joinder, it leaves the parties’ legal rights and duties intact and the rules of decision unchanged.” Shady Grove Orthopedic Associates, P. A. v. Allstate Ins. Co., 559 U. S. 393, 408 (2010) (plurality opinion).

The second possibility is that every member of the plaintiff class has suffered an Eighth Amendment violation merely by virtue of being a patient in a poorly-run prison system, and the purpose of the class is merely to aggregate all those individually viable claims. This theory has the virtue of *553being consistent with procedural principles, but at the cost of a gross substantive departure from our case law. Under this theory, each and every prisoner who happens to be a patient in a system that has systemic weaknesses — such as “hir[ing] any doctor who had a license, a pulse and a pair of shoes,” ante, at 508 (internal quotation marks omitted) — has suffered cruel or unusual punishment, even if that person cannot make an individualized showing of mistreatment. Such a theory of the Eighth Amendment is preposterous. And we have said as much in the past: “If ... a healthy inmate who had suffered no deprivation of needed medical treatment were able to claim violation of his constitutional right to medical care ... simply on the ground that the prison medical facilities were inadequate, the essential distinction between judge and executive would have disappeared: it would have become the function of the courts to assure adequate medical care in prisons.” Lewis v. Casey, 518 U. S. 343, 350 (1996).

Whether procedurally wrong or substantively wrong, the notion that the plaintiff class can allege an Eighth Amendment violation based on “systemwide deficiencies” is assuredly wrong. It follows that the remedy decreed here is also contrary to law, since the theory of systemic unconstitutionality is central to the plaintiffs’ case. The PLRA requires plaintiffs to establish that the systemwide injunction entered by the District Court was “narrowly drawn” and “extends no further than necessary” to correct “the violation of the Federal right of a particular plaintiff or plaintiffs.” If (as is the case) the only viable constitutional claims consist of individual instances of mistreatment, then a remedy reforming the system as a whole goes far beyond what the statute allows.

It is also worth noting the peculiarity that the vast majority of inmates most generously rewarded by the release order — the 46,000 whose incarceration will be ended — do not form part of any aggrieved class even under the Court’s ex*554pansive notion of constitutional violation. Most of them will not be prisoners with medical conditions or severe mental illness; and many will undoubtedly be fine physical specimens who have developed intimidating muscles pumping iron in the prison gym.

B

Even if I accepted the implausible premise that the plaintiffs have established a systemwide violation of the Eighth Amendment, I would dissent from the Court’s endorsement of a decrowding order. That order is an example of what has become known as a “structural injunction.” As'I have previously explained, structural injunctions are radically different from the injunctions traditionally issued by courts of equity, and presumably part of the “judicial Power” conferred on federal courts by Article III:

“The mandatory injunctions issued upon termination of litigation usually required ‘a single simple act.’ H. Mc-Clintock, Principles of Equity §15, pp. 82-33 (2d ed. 1948). Indeed, there was a ‘historical prejudice of the court of chancery against rendering decrees which called for more than a single affirmative act.’ Id., §61, at 160. And where specific performance of contracts was sought, it was the categorical rule that no decree would issue that required ongoing supervision.... Compliance with these ‘single act’ mandates could, in addition to being simple, be quick; and once it was achieved the contemnor’s relationship with the court came to an end, at least insofar as the subject of the order was concerned. Once the document was turned over or the land conveyed, the litigant’s obligation to the court, and the court’s coercive power over the litigant, ceased. . . . The court did not engage in any ongoing supervision of the litigant’s conduct, nor did its order continue to regulate his behavior.” Mine Workers v. Bagwell, 512 U. S. 821, 841-842 (1994) (Scalia, J., concurring).

*555Structural injunctions depart from that historical practice, turning judges into long-term administrators of complex social institutions such as schools, prisons, and police departments. Indeed, they require judges to play a role essentially indistinguishable from the role ordinarily played by executive officials. Today’s decision not only affirms the structural injunction but vastly expands its use, by holding that an entire system is unconstitutional because it may produce constitutional violations.

The drawbacks of structural injunctions have been described at great length elsewhere. See, e. g., Lewis, supra, at 385-393 (Thomas, J., concurring); Missouri v. Jenkins, 515 U. S. 70, 124-133 (1995) (Thomas, J., concurring); Horowitz, Decreeing Organizational Change: Judicial Supervision of Public Institutions, 1983 Duke L. J. 1265. This case illustrates one of their most pernicious aspects: that they force judges to engage in a form of factfinding-as-policymaking that is outside the traditional judicial role. The factfinding judges traditionally engage in involves the determination of past or present facts based (except for a limited set of materials of which courts may take “judicial notice”) exclusively upon a closed trial record. That is one reason why a district judge’s factual findings are entitled to clear-error review: because having viewed the trial first hand he is in a better position to evaluate the evidence than a judge reviewing a cold record. In a very limited category of cases, judges have also traditionally been called upon to make some predictive judgments: which custody will best serve the interests of the child, for example, or whether a particular one-shot injunction will remedy the plaintiff’s grievance. When a judge manages a structural injunction, however, he will inevitably be required to make very broad empirical predictions necessarily based in large part upon policy views — the sort of predictions regularly made by legislators and executive officials, but inappropriate for the Third Branch.

*556This feature of structural injunctions is superbly illustrated by the District Court’s proceeding concerning the deerowding order’s effect on public safety. The PLRA requires that, before granting “[prospective relief in [a] civil action with respect to prison conditions,” a court must “give substantial weight to any adverse impact on public safety or the operation of a criminal justice system caused by the relief.” 18 U. S. C. § 3626(a)(1)(A). Here, the District Court discharged that requirement by making the “factual finding” that “the state has available methods by which it could readily reduce the prison population to 137.5% design capacity or less without an adverse impact on public safety or the operation of the criminal justice system.” App. to Juris. Statement, O. T. 2009, No. 09-416, p. 253a. It found the evidence “clear” that prison overcrowding would “perpetuate a criminogenic prison system that itself threatens public safety,” id., at 186a, and volunteered its opinion that “[t]he population could be reduced even further with the reform of California’s antiquated sentencing policies and other related changes to the laws.” Id., at 253a. It “rejected] the testimony that inmates released early from prison would commit additional new crimes,” id., at 200a, finding that “shortening the length of stay through earned credits would give inmates incentives to participate in programming designed to lower recidivism,” id., at 204a, and that “slowing the flow of technical parole violators to prison, thereby substantially reducing the churning of parolees, would by itself improve both the prison and parole systems, and public safety,” id., at 209a. It found that “the diversion of offenders to community correctional programs has significant beneficial effects on public safety,” id., at 214a, and that “additional rehabilitative programming would result in a significant population reduction while improving public safety,” id., at 216a.

The District Court cast these predictions (and the Court today accepts them) as “factual findings,” made in reliance on the procession of expert witnesses who testified at trial. *557Because these “findings” have support in the record, it is difficult to reverse them under a plain-error standard of review. Ante, at 535. And given that the District Court devoted nearly 10 days of trial and 70 pages of its opinion to this issue, it is difficult to dispute that the District Court has discharged its statutory obligation to give “substantial weight to any adverse impact on public safety.”

But the idea that the three District Judges in this case relied solely on the credibility of the testifying expert witnesses is fanciful. Of course they were relying largely on their own beliefs about penology and recidivism. And of course different district judges, of different policy views, would have “found” that rehabilitation would not work and that releasing prisoners would increase the crime rate. I am not saying that the District Judges rendered their factual findings in bad faith. I am saying that it is impossible for judges to make “factual findings” without inserting their own policy judgments, when the factual findings are policy judgments. What occurred here is no more judicial fact-finding in the ordinary sense than would be the factual findings that deficit spending will not lower the unemployment rate, or that the continued occupation of Iraq will decrease the risk of terrorism. Yet, because they have been branded “factual findings” entitled to deferential review, the policy preferences of three District Judges now govern the operation of California's penal system.

It is important to recognize that the dressing-up of policy judgments as factual findings is not an error peculiar to this case. It is an unavoidable concomitant of institutional-reform litigation. When a district court issues an injunction, it must make a factual assessment of the anticipated consequences of the injunction. And when the injunction undertakes to restructure a social institution, assessing the factual consequences of the injunction is necessarily the sort of predictive judgment that our system of government allocates to other government officials.

*558But structural injunctions do not simply invite judges to indulge policy preferences. They invite judges to indulge incompetent policy preferences. Three years of law school and familiarity with pertinent Supreme Court precedents give no insight whatsoever into the management of social institutions. Thus, in the proceeding below the District Court determined that constitutionally adequate medical services could be provided if the prison population was 137.5% of design capacity. This was an empirical finding it was utterly unqualified to make. Admittedly, the court did not generate that number entirely on its own; it heard the numbers 130% and 145% bandied about by various witnesses and decided to split the difference. But the ability of judges to spit back or even average out numbers spoon fed to them by expert witnesses does not render them competent decisionmakers in areas in which they are otherwise unqualified.

The District Court also relied heavily on the views of the Receiver and Special Master, and those reports play a starring role in the Court's opinion today. The Court notes that “the Receiver and the Special Master filed reports stating that overcrowding posed a significant barrier to their efforts” and deems those reports “persuasive evidence that, absent a reduction in overcrowding, any remedy might prove unattainable and would at the very least require vast expenditures of resources by the State.” Ante, at 529. The use of these reports is even less consonant with the traditional judicial role than the District Court's reliance on the expert testimony at trial. The latter, even when, as here, it is largely the expression of policy judgments, is at least subject to cross-examination. Relying on the un-crossexamined findings of an investigator, sent into the field to prepare a factual report and give suggestions on how to improve the prison system, bears no resemblance to ordinary judicial decisionmaking. It is true that the PLRA contemplates the appointment of special masters (although not receivers), but special masters are authorized only to “conduct *559hearings and prepare proposed findings of fact” and “assist in the development of remedial plans,” 18 U. S. C. § 3626(f)(6). This does not authorize them to make factual findings (unconnected to hearings) that are given seemingly wholesale deference. Neither the Receiver nor the Special Master was selected by California to run its prisons, and the fact that they may be experts in the field of prison reform does not justify the judicial imposition of their perspectives on the State.

C

My general concerns associated with judges’ running social institutions are magnified when they run prison systems, and doubly magnified when they force prison officials to release convicted criminals. As we have previously recognized:

“[Cjourts are ill equipped to deal with the increasingly urgent problems of prison administration and reform. . . . [Tjhe problems of prisons in America are complex and intractable, and, more to the point, they are not readily susceptible of resolution by decree. . .. Running a prison is an inordinately difficult undertaking that requires expertise, planning, and the commitment of resources, all of which are peculiarly within the province of the legislative and executive branches of government. Prison administration is, moreover, a task that has been committed to the responsibility of those branches, and separation of powers concerns counsel a policy of judicial restraint. Where a state penal system is involved, federal courts have .. . additional reason to accord deference to the appropriate prison authorities.” Turner v. Safley, 482 U. S. 78, 84-85 (1987) (internal quotation marks omitted).

These principles apply doubly to a prisoner-release order. As the author of today’s opinion explained earlier this Term, granting a writ of habeas corpus “ ‘disturbs the State’s sig*560nificant interest in repose for concluded litigation, denies society the right to punish some admitted offenders, and intrudes on state sovereignty to a degree matched by few exercises of federal judicial authority.’” Harrington v. Richter, 562 U. S. 86, 103 (2011) (quoting Harris v. Reed, 489 U. S. 255, 282 (1989) (Kennedy, J., dissenting)). Recognizing that habeas relief must be granted sparingly, we have reversed the Ninth Circuit's erroneous grant of habeas relief to individual California prisoners four times this Term alone. Cullen v. Pinholster, ante, p. 170; Felkner v. Jackson, 562 U. S. 594 (2011) (per curiam); Swarthout v. Cooke, 562 U. S. 216 (2011) (per curiam); Harrington, supra. And yet here, the Court affirms an order granting the fimctional equivalent of 46,000 writs of habeas corpus, based on its paean to courts’ “substantial flexibility when making these judgments.” Ante, at 538. It seems that the Court’s respect for state sovereignty has vanished in the case where it most matters.

II

The Court’s opinion includes a bizarre coda noting that “[t]he State may wish to move for modification of the three-judge court’s order to extend the deadline for the required reduction to five years.” Ante, at 543. The District Court, it says, “may grant such a request provided that the State satisfies necessary and appropriate preconditions designed to ensure that measures are taken to implement the plan without undue delay”; and it gives vague suggestions of what these preconditions “may include,” such as “interim benchmarks.” Ante, at 543-544. It also invites the District Court to “consider whether it is appropriate to order the State to begin without delay to develop a system to identify prisoners who are unlikely to reoffend,” and informs the State that it “should devise systems to select those prisoners least likely to jeopardize public safety.” Ante, at 544. (What a good idea!)

*561The legal effect of this passage is unclear — I suspect intentionally so. If it is nothing but a polite reminder to the State and to the District Court that the injunction is subject to modification, then it is entirely unnecessary. As both the State and the District Court are undoubtedly aware, a party is always entitled to move to modify an equitable decree, and the PLRA contains an express provision authorizing District Courts to modify or terminate prison injunctions. See 18 U.S. C. § 3626(b).

I suspect, however, that this passage is a warning shot across the bow, telling the District Court that it had better modify the injunction if the State requests what we invite it to request. Such a warning, if successful, would achieve the benefit of a marginal reduction in the inevitable murders, robberies, and rapes to be committed by the released inmates. But it would achieve that at the expense of intellectual bankruptcy, as the Court’s “warning” is entirely alien to ordinary principles of appellate review of injunctions. When a party moves for modification of an injunction, the district court is entitled to rule on that motion first, subject to review for abuse of discretion if it declines to modify the order. Horne v. Flores, 557 U. S. 433, 447, 456 (2009). Moreover, when a district court enters a new decree with new benchmarks, the selection of those benchmarks is also reviewed under a deferential, abuse-of-diseretion standard of review — a point the Court appears to recognize. Ante, at 542. Appellate courts are not supposed to “affirm” injunctions while preemptively noting that the State “may” request, and the District Court “may” grant, a request to extend the State’s deadline to release prisoners by three years based on some suggestions on what appropriate preconditions for such a modification “may” include.

Of course what is really happening here is that the Court, overcome by common sense, disapproves of the results reached by the District Court, but cannot remedy them (it *562thinks) by applying ordinary standards of appellate review. It has therefore selected a solution unknown in our legal system: A deliberately ambiguous set of suggestions on how to modify the injunction, just deferential enough so that it can say with a straight face that it is “affirming,” just stern enough to put the District Court on notice that it will likely get reversed if it does not follow them. In doing this, the Court has aggrandized itself, grasping authority that appellate courts are not supposed to have, and using it to enact a compromise solution with no legal basis other than the Court’s say-so. That we are driven to engage in these extralegal activities should be a sign that the entire project of permitting district courts to run prison systems is misbegotten.

But perhaps I am being too unkind. The Court, or at least a majority of the Court’s majority, must be aware that the judges of the District Court are likely to call its bluff, since they know full well it cannot possibly be an abuse of discretion to refuse to accept the State’s proposed modifications in an injunction that has just been approved (affirmed) in its present form. An injunction, after all, does not have to be perfect; only good enough for government work, which the Court today says this is. So perhaps the coda is nothing more than a ceremonial washing of the hands — making it clear for all to see, that if the terrible things sure to happen as a consequence of this outrageous order do happen, they will be none of this Court’s responsibility. After all, did we not want, and indeed even suggest, something better?

Ill

In view of the incoherence of the Eighth Amendment claim at the core of this case, the nonjudicial features of institutional-reform litigation that this case exemplifies, and the unique concerns associated with mass prisoner releases, I do not believe this Court can affirm this injunction. I will *563state my approach briefly: In my view, a court may not order a prisoner's release unless it determines that the prisoner is suffering from a violation of his constitutional rights, and that his release, and no other relief, will remedy that violation. Thus, if the court determines that a particular prisoner is being denied constitutionally required medical treatment, and the release of that prisoner (and no other remedy) would enable him to obtain medical treatment, then the court can order his release; but a court may not order the release of prisoners who have suffered no violations of their constitutional rights, merely to make it less likely that that will happen to them in the future.

This view follows from the PLRA’s text that I discussed at the outset, 18 U. S. C. § 3626(a)(1)(A). “[N]arrowly drawn” means that the relief applies only to the “particular [prisoner] or [prisoners]” whose constitutional rights are violated; “extends no further than necessary” means that prisoners whose rights are not violated will not obtain relief; and “least intrusive means necessary to correct the violation of the Federal right” means that no other relief is available.*

I acknowledge that this reading of the PLRA would severely limit the circumstances under which a court could issue structural injunctions to remedy allegedly unconstitutional prison conditions, although it would not eliminate them entirely. If, for instance, a class representing all prisoners in a particular institution alleged that the temperature in their cells was so cold as to violate the Eighth Amendment, or that they were deprived of all exercise time, a court could enter a prisonwide injunction ordering that the tem*564perature be raised or exercise time be provided. Still, my approach may invite the objection that the PLRA appears to contemplate structural injunctions in general and mass prisoner-release orders in particular. The statute requires courts to “give substantial weight to any adverse impact on public safety or the operation of a criminal justice system caused by the relief” and authorizes them to appoint special masters, § 3626 (a)(1)(A), (f), provisions that seem to presuppose the possibility of a structural remedy. It also sets forth criteria under which courts may issue orders that have “the purpose or effect of reducing or limiting the prison population,” § 3626(g)(4).

I do not believe that objection carries the day. In addition to imposing numerous limitations on the ability of district courts to order injunctive relief with respect to prison conditions, the PLRA states that “[n]othing in this section shall be construed to . . . repeal or detract from otherwise applicable limitations on the remedial powers of the courts.” § 3626(a)(1)(C). The PLRA is therefore best understood as an attempt to constrain the discretion of courts issuing structural injunctions — not as a mandate for their use. For the reasons I have outlined, structural injunctions, especially prisoner-release orders, raise grave separation-of-powers concerns and veer significantly from the historical role and institutional capability of courts. It is appropriate to construe the PLRA so as to constrain courts from entering injunctive relief that would exceed that role and capability.

* * *

The District Court’s order that California release 46,000 prisoners extends “further than necessary to correct the violation of the Federal right of a particular plaintiff or plaintiffs” who have been denied needed medical care. 18 U. S. C. § 3626(a)(1)(A). It is accordingly forbidden by the PLRA— besides defying all sound conception of the proper role of judges.

*565Justice Alito,

with whom The Chief Justice joins, dissenting.

The decree in this case is a perfect example of what the Prison Litigation Reform Act of 1995 (PLRA), 110 Stat. 1321-66, was enacted to prevent.

The Constitution does not give federal judges the authority to run state penal systems. Decisions regarding state prisons have profound public safety and financial implications, and the States are generally free to make these decisions as they choose. See Turner v. Safiey, 482 U. S. 78, 85 (1987).

The Eighth Amendment imposes an important — but limited — restraint on state authority in this field. The Eighth Amendment prohibits prison officials from depriving inmates of “the minimal civilized measure of life’s necessities.” Rhodes v. Chapman, 452 U. S. 337, 347 (1981). Federal courts have the responsibility to ensure that this constitutional standard is met, but undesirable prison conditions that do not violate the Constitution are beyond the federal courts’ reach.

In this case, a three-judge court exceeded its authority under the Constitution and the PLRA. The court ordered a radical reduction in the California prison population without finding that the current population level violates the Constitution.

Two cases were before the three-judge court, and neither targeted the general problem of overcrowding. Indeed, the plaintiffs in one of those cases readily acknowledge that the current population level is not itself unconstitutional. Brief for Coleman Appellees 56. Both of the cases were brought not on behalf of all inmates subjected to overcrowding, but rather in the interests of much more limited classes of prisoners, namely, those needing mental health treatment and those with other serious medical needs. But these cases were used as a springboard to implement a criminal justice program far different from that chosen by the state legisla*566ture. Instead of crafting a remedy to attack the specific constitutional violations that were found — which related solely to prisoners in the two plaintiff classes — the lower court issued a decree that will at best provide only modest help to those prisoners but that is very likely to have a major and deleterious effect on public safety.

The three-judge court ordered the premature release of approximately J/,6,000 criminals — the equivalent of three Army divisions.

The approach taken by the three-judge court flies in the face of the PLRA. Contrary to the PLRA, the court’s remedy is not narrowly tailored to address proven and ongoing constitutional violations. And the three-judge court violated the PLRA’s critical command that any court contemplating a prisoner release order must give “substantial weight to any adverse impact on public safety.” 18 U. S. C. § 3626(a)(1)(A). The three-judge court would have us believe that the early release of 46,000 inmates will not imperil — and will actually improve — public safety. App. to Juris. Statement, O. T. 2009, No. 09-416, pp. 248a-249a (hereinafter Juris. App.). Common sense and experience counsel greater caution.

I would reverse the decision below for three interrelated reasons. First, the three-judge court improperly refused to consider evidence concerning present conditions in the California prison system. Second, the court erred in holding that no remedy short of a massive prisoner release can bring the California system into compliance with the Eighth Amendment. Third, the court gave inadequate weight to the impact of its decree on public safety.

I

Both the PLRA and general principles concerning injunctive relief dictate that a prisoner release order cannot properly be issued unless the relief is necessary to remedy an ongoing violation. Under the PLRA, a prisoner release may *567be decreed only if crowding “is the primary cause” of an Eighth Amendment violation and only if no other relief “will remedy” the violation. § 3626(a)(3)(E) (emphasis added). This language makes it clear that proof of past violations alone is insufficient to justify a court-ordered prisoner release.

Similarly, in cases not governed by the PLRA, we have held that an inmate seeking an injunction to prevent a violation of the Eighth Amendment must show that prison officials are “knowingly and unreasonably disregarding an objectively intolerable risk of harm, and that they will continue to do so ... into the future.” Farmer v. Brennan, 511 U. S. 825, 846 (1994). The “deliberate indifference” needed to establish an Eighth Amendment violation must be examined “in light of the prison authorities’ current attitudes and conduct,” Helling v. McKinney, 509 U. S. 25, 36 (1993), which means “their attitudes and conduct at the time suit is brought and persisting thereafter,” Farmer, supra, at 845.

For these reasons, the propriety of the relief ordered here cannot be assessed without ascertaining the nature and scope of any ongoing constitutional violations. Proof of past violations will not do; nor is it sufficient simply to establish that some violations continue. The scope of permissible relief depends on the scope of any continuing violations, and therefore it was essential for the three-judge court to make a reliable determination of the extent of any violations as of the time its release order was issued. Particularly in light of the radical nature of its chosen remedy, nothing less than an up-to-date assessment was tolerable.

The three-judge court, however, relied heavily on outdated information and findings and refused to permit California to introduce new evidence. Despite evidence of improvement,1 *568the three-judge court relied on old findings made by the single-judge courts, see Juris. App. 76a-77a, including a finding made 14 years earlier, see id., at 170a (citing Coleman v. Wilson, 912 F. Supp. 1282, 1316, 1319 (ED Cal. 1995)). The three-judge court highlighted death statistics from 2005, see Juris. App. 9a, while ignoring the “significant and continuous decline since 2006,” California Prison Health Care Receivership Corp., K. Imai, Analysis of Year 2008 Death Reviews 31 (Dec. 2009) (hereinafter 2008 Death Reviews). And the court dwelled on conditions at a facility that has since been replaced. See Juris. App. 19a-20a, 24a, 89a-90a, 94a, 107a, 111a.

Prohibiting the State from introducing evidence about conditions as of the date when the prisoner release order was under consideration, id., at 76a-78a, and n. 42, the three-judge court explicitly stated that it would not “evaluate the state’s continuing constitutional violations,” id., at 77a. Instead, it based its remedy on constitutional deficiencies that, in its own words, were found “years ago.” Ibid.2

The three-judge court justified its refusal to receive up-to-date evidence on the ground that the State had not filed a motion to terminate prospective relief under a provision of the PLRA, § 3626(b). See Juris. App. 77a. Today’s opinion for this Court endorses that reasoning, ante, at 523-524. But the State’s opportunity to file such a motion did not eliminate the three-judge court’s obligation to ensure that its relief was necessary to remedy ongoing violations.3 More*569over, the lower court’s reasoning did not properly take into account the potential significance of the evidence that the State sought to introduce. Even if that evidence did not show that all violations had ceased — the showing needed to obtain the termination of relief under § 3626(b) — that evidence was highly relevant with respect to the nature and scope of permissible relief.4

The majority approves the three-judge court’s refusal to receive fresh evidence based largely on the need for “[orderly trial management.” Ibid. The majority reasons that the three-judge court had closed the book on the question of constitutional violations and had turned to the question of remedy. Ibid. As noted, however, the extent of any continuing constitutional violations was highly relevant to the question of remedy.

The majority also countenances the three-judge court’s reliance on dated findings. The majority notes that the lower court considered recent reports by the Special Master and ’Receiver, ante, at 516, but the majority provides no persuasive justification for the lower court’s refusal to receive hard, up-to-date evidence about any continuing violations. *570With the safety of the people of California in the balance, the record on this issue should not have been closed.

The majority repeats the lower court’s error of reciting statistics that are clearly out of date. The Court notes the lower court’s finding that as of 2005 “ ‘an inmate in one of California’s prisons needlessly dies every six to seven days.’ ” See ante, at 507. Yet by the date of the trial before the three-judge court, the death rate had been trending downward for 10 quarters, App. 2257, and the number of likely preventable deaths fell from 18 in 2006 to 3 in 2007, a decline of 83 percent.5 Between 2001 and 2007, the California prison system had the 13th lowest average mortality rate of all 50 state systems.6

The majority highlights past instances in which particular prisoners received shockingly deficient medical care. See ante, at 503-504, 504-505, 508 (recounting five incidents). But such anecdotal evidence cannot be given undue weight in assessing the current state of the California system. The population of the California prison system (156,000 inmates *571at the time of trial) is larger than that of many medium-sized cities,7 and an examination of the medical care provided to the residents of many such cities would likely reveal cases in which grossly deficient treatment was provided. Instances of past mistreatment in the California system are relevant, but prospective relief must be tailored to present and future, not past, conditions.

II

Under the PLRA, a court may not grant any prospective relief unless the court finds that the relief is narrowly drawn, extends no further than necessary to correct the “violation of [a] Federal right, and is the least intrusive means necessary to correct the violation of the Federal right.” § 3626(a)(1)(A). In addition, the PLRA prohibits the issuance of a prisoner release order unless the court finds “by clear and convincing evidence that . . . crowding is the primary cause of the violation of a Federal right” and that “no other relief will remedy the violation of the Federal right.” § 3626(a)(3)(E).

These statutory restrictions largely reflect general stand- ’ ards for injunctive relief aimed at remedying constitutional violations by state and local governments. “The power of the federal courts to restructure the operation of local and state governmental entities is not plenary. ... Once a constitutional violation is found, a federal court is required to tailor the scope of the remedy to fit the nature and extent of the constitutional violation.” Dayton Bd. of Ed. v. Brinkman, 433 U. S. 406, 419-420 (1977) (internal quotation marks omitted).

Here, the majority and the court below maintain that no remedy short of a massive release of prisoners from the general prison population can remedy the State’s failure to pro*572vide constitutionally adequate health care. This argument is implausible on its face and is not supported by the requisite clear and convincing evidence.

It is instructive to consider the list of deficiencies in the California prison health care system that are highlighted in today’s opinion for this Court and in the opinion of the court below. The deficiencies noted by the majority here include the following: “ ‘[e]xam tables and counter tops, where prisoners with . .. communicable diseases are treated, [are] not routinely disinfected,’ ” ante, at 508; medical facilities “ ‘are in an abysmal state of disrepair,”' ibid.; medications "‘are too often not available when needed,’ ” ibid.; “ ‘[b]asie medical equipment is often not available or used,’ ” ibid.; prisons “would ‘hire any doctor who had “a license, a pulse and a pair of shoes,” ’ ” ibid.; and medical and mental health staff positions have high vacancy rates, ante, at 517. The three-judge court pointed to similar problems. See Juris. App. 93a-121a (citing, among other things, staffing vacancies, too few beds for mentally ill prisoners, and an outmoded records management system).

Is it plausible that none of these deficiencies can be remedied without releasing 46,000 prisoners? Without taking that radical and dangerous step, exam tables and counter-tops cannot properly be disinfected? None of the system’s dilapidated facilities can be repaired? Needed medications and equipment cannot be purchased and used? Staff vacancies cannot be filled? The qualifications of prison physicians cannot be improved? A better records management system cannot be developed and implemented?

I do not dispute that general overcrowding contributes to many of the California system’s health care problems. But it by no means follows that reducing overcrowding is the only or the best or even a particularly good way to alleviate those problems. Indeed, it is apparent that the prisoner release ordered by the court below is poorly suited for this purpose. The release order is not limited to prisoners need*573ing substantial medical care but instead calls for a reduction in the system’s overall population. Under the order issued by the court below, it is not necessary for a single prisoner in the plaintiff classes to be released. Although some class members will presumably be among those who are discharged, the decrease in the number of prisoners needing mental health treatment or other forms of extensive medical care will be much smaller than the total number of prisoners released, and thus the release will produce at best only a modest improvement in the burden on the medical care system.

The record bears this out. The Special Master stated dramatically that even releasing 100,000 inmates (two-thirds of the California system’s entire inmate population!) would leave the problem of providing mental health treatment “largely unmitigated.” App. 487. Similarly, the Receiver proclaimed that “ ‘those . . . who think that population controls will solve California’s prison health care problems . . . are simply wrong.’ ” Juris. App. 282a.

The State proposed several remedies other than a massive release of prisoners, but the three-judge court, seemingly intent on attacking the broader problem of general overcrowding, rejected all of the State’s proposals. In doing so, the court made three critical errors.

First, the court did not assess those proposals and other remedies in light of conditions proved to exist at the time the release order was framed. Had more recent evidence been taken into account, a less extreme remedy might have been shown to be sufficient.

Second, the court failed to distinguish between conditions that fall below the level that may be desirable as a matter of public policy and conditions that do not meet the minimum level mandated by the Constitution. To take one example, the court criticized the California system because prison doctors must conduct intake exams in areas separated by folding screens rather than in separate rooms, creating conditions *574that “do not allow for appropriate confidentiality.” Id., at 88a. But the legitimate privacy expectations of inmates are greatly diminished, see Hudson v. Palmer, 468 U. S. 517, 525-526 (1984), and this Court has never suggested that the failure to provide private consultation rooms in prisons amounts to cruel and unusual punishment.

Third, .the court rejected alternatives that would not have provided “ ‘immediate’ ” relief. Juris. App. 148a. But nothing in the PLRA suggests that public safety may be sacrificed in order to implement an immediate remedy rather than a less dangerous one that requires a more extended but reasonable period of time.

If the three-judge court had not made these errors, it is entirely possible that an adequate but less drastic remedial plan could have been crafted. Without up-to-date information, it is not possible to specify what such a plan might provide, and in any event, that is hot a task that should be undertaken in the first instance by this Court. But possible components of such a plan are not hard to identify.

Many of the problems noted above plainly could be addressed without releasing prisoners and without incurring the costs associated with a large-scale prison construction program. Sanitary procedures could be improved; sufficient supplies of medicine and medical equipment could be purchased; an adequate system of records management could be implemented; and the number of medical and other staff positions could be increased. Similarly, it is hard to believe that staffing vacancies cannot be reduced or eliminated and that the qualifications of medical personnel cannot be improved by any means short of a massive prisoner release. Without specific findings backed by hard evidence, this Court should not accept the counterintuitive proposition that these problems cannot be ameliorated by increasing salaries, improving working conditions, and providing better training and monitoring of performance.

*575While the cost of a large-scale construction program may well exceed California’s current financial capabilities, a more targeted program, involving the repair and perhaps the expansion of current medical facilities (as opposed to general prison facilities), might be manageable. After all, any remedy in this case, including the new programs associated with the prisoner release order and other proposed relief now before the three-judge court, will necessarily involve some state expenditures.

Measures such as these might be combined with targeted reductions in critical components of the State’s prison population. A certain number of prisoners in the classes on whose behalf the two cases were brought might.be transferred to out-of-state facilities. The three-judge court rejected the State’s proposal to transfer prisoners to out-of-state facilities in part because the number of proposed transfers was too small. See id., at 160a. See also ante, at 527. But this reasoning rested on the court’s insistence on a reduction in the State’s general prison population rather than the two plaintiff classes.

When the State proposed to make a targeted transfer of prisoners in one of the plaintiff classes (i e., prisoners needing mental health treatment), one of the District Judges blocked the transfers for fear that the out-of-state facilities would not provide a sufficiently high level of care. See App. 434-440. The District Judge even refused to allow out-of-state transfers for prisoners who volunteered for relocation. See id., at 437. And the court did this even though there was not even an allegation, let alone clear evidence, that the States to which these prisoners would have been sent were violating the Eighth Amendment.

The District Judge presumed that the receiving States might fail to provide constitutionally adequate care, but “ ‘in the absence of clear evidence to the contrary, courts presume that [public officers] have properly discharged their official duties.’” United States v. Armstrong, 517 U. S. 456, 464 *576(1996) (quoting United States v. Chemical Foundation, Inc., 272 U. S. 1, 14-15 (1926)); Postal Service v. Gregory, 534 U. S. 1, 10 (2001) (“[A] presumption of regularity attaches to the actions of Government agencies”); see also McKune v. Lile, 536 U. S. 24, 51 (2002) (O’Connor, J., concurring in judgment) (“[W]e may assume that the prison is capable of controlling its inmates so that respondent’s personal safety is not jeopardized , at least in the absence of proof to the contrary”).8

Finally, as a last resort, a much smaller release of prisoners in the two plaintiff classes could be considered. Plaintiffs proposed not only a systemwide population cap, but also a lower population cap for inmates in specialized programs. Tr. 2915:12-15 (Feb. 3, 2009). The three-judge court rejected this proposal, and its response exemplified what went wrong in this case. One judge complained that this remedy would be deficient because it would protect only the members of the plaintiff classes. The judge stated:

“The only thing is we would be protecting the class members. And maybe that’s the appropriate thing to do. I mean, that’s what this ease is about, but it would be ... difficult for me to say yes, and the hell with everybody else.” Id., at 2915:23-2916:2.

Overstepping his authority, the judge was not content to provide relief for the classes of plaintiffs on whose behalf the suit before him was brought. Nor was he content to remedy the only constitutional violations that were proved — which concerned the treatment of the members of those classes. Instead, the judge saw it as his responsibility to attack the general problem of overcrowding.

*577III

Before ordering any prisoner release, the PLRA commands a court to “give substantial weight to any adverse impact on public safety or the operation of a criminal justice system caused by the relief.” §3626(a)(1)(A). This provision unmistakably reflects Congress’ view that prisoner release orders are inherently risky.

In taking this view, Congress was well aware of the impact of previous prisoner release orders. The prisoner release program carried out a few years earlier in Philadelphia is illustrative. In the early 1990’s, federal courts enforced a cap on the number of inmates in the Philadelphia prison system, and thousands of inmates were set free. Although efforts were made to release only those prisoners who were least likely to commit violent crimes, that attempt was spectacularly unsuccessful. During an 18-month period, the Philadelphia police rearrested thousands of these prisoners for committing 9,732 new crimes. Those defendants were charged with 79 murders, 90 rapes, 1,113 assaults, 959 robberies, 701 burglaries, and 2,748 thefts, not to mention thousands of drug offenses.9 Members of Congress were well aware of this experience.10

*578Despite the record of past prisoner release orders, the three-judge court in this case concluded that loosing 46,000 criminals would not produce a tally like that in Philadelphia and would actually improve public safety. Juris. App. 248a-249a. In reaching this debatable conclusion, the three-judge court relied on the testimony of selected experts, id., at 248a, and the majority now defers to what it characterizes as the lower court’s findings of fact on this controversial public policy issue, ante, at 512-513, 517, 522.

This is a fundamental and dangerous error. When a trial court selects between the competing views of experts on broad empirical questions such as the efficacy of preventing crime through the incapacitation of convicted criminals, the trial court’s choice is very different from a classic finding of fact and is not entitled to the same degree of deference on appeal.

The particular three-judge court convened in this case was “confident" that releasing 46,000 prisoners pursuant to its plan “would in fact benefit public safety.” Juris. App. 248a-249a. According to that court, “overwhelming evidence” supported this purported finding. Id., at 232a. But a more cautious court, less bent on implementing its own criminal justice agenda, would have at least acknowledged that the consequences of this massive prisoner release cannot be ascertained in advance with any degree of certainty and that it is entirely possible that this release will produce results similar to those under prior court-ordered population caps. After all, the sharp increase in the California prison population that the three-judge court lamented, see id., at 254a, has been accompanied by an equally sharp decrease in violent crime.11 These California trends mirror similar develop*579ments at the national level,12 and “[tjhere is a general consensus that the decline in crime is, at least in part, due to more and longer prison sentences.”13 If increased incarceration in California has led to decreased crime, it is entirely possible that a decrease in imprisonment will have the opposite effect.

Commenting on the testimony of an expert who stated that he could not be certain about the effect of the massive prisoner discharge on public safety, the three-judge court complained that “[sjuch equivocal testimony is not helpful.” Id., at 247a. But testimony pointing out the difficulty of assessing the consequences of this drastic remedy would have been valued by a careful court duly mindful of the overriding need to guard public safety.

The three-judge court acknowledged that it “ha[dj not evaluated the public safety impact of each individual element” of the population reduction plan it ordered the State to implement. App. to Juris. Statement 3a. The majority argues that the three-judge court nevertheless gave substantial weight to public safety because its order left “details of implementation to the State’s discretion.” Ante, at 538. Yet the State had told the three-judge court that, after studying possible population reduction measures, it concluded that “reducing the prison population to 137.5% within *580a two-year period cannot be accomplished without unacceptably compromising public safety.” Juris. App. 317a. The State found that public safety required a 5-year period in which to achieve the ordered reduction. Ibid.

Thus, the three-judge court approved a population reduction plan that neither it nor the State found could be implemented without unacceptable harm to public safety. And this Court now holds that the three-judge court discharged its obligation to “give substantial weight to any adverse impact on public safety,” § 3626(a)(1)(A), by deferring to officials who did not believe the reduction could be accomplished in a safe manner. I do not believe the PLRA’s public-safety requirement is so trivial.

The members of the three-judge court and the experts on whom they relied may disagree with key elements of the crime-reduction program that the State of California has pursued for the past few decades, including “the shift to inflexible determinate sentencing and the passage of harsh mandatory minimum and three-strikes laws.” Id., at 254a. And experts such as the Receiver are entitled to take the view that the State should “re-thin[k] the place of incarceration in its criminal justice system,” App. 489. But those controversial opinions on matters of criminal justice policy should not be permitted to override the reasonable policy view that is implicit in the PLRA — that prisoner release orders present an inherent risk to the safety of the public.

* * *

The prisoner release ordered in this case is unprecedented, improvident, and contrary to the PLRA. In largely sustaining the decision below, the majority is gambling with the safety of the people of California. Before putting public safety at risk, every reasonable precaution should be taken. The decision below should be reversed, and the case should be remanded for this to be done.

*581I fear that today’s decision, like prior prisoner release orders, will lead to a grim roster of victims. I hope that I am -wrong.

In a few years, we will see.

Mr. Justice Black

delivered the opinion of the Court.

A jury in a New York state court found appellant guilty of murder in the first degree.1 The jury recommended life imprisonment, but the trial judge imposed sentence of death.2- In giving his reasons for imposing the death sentence the judge discussed in open court the evidence upon.which the jury had convicted stating that this evidence had been considered in the light of additional information obtained through the court’s “Probation Department, and through other sources.” Consideration of *243this additional information was pursuant to § 482 of New York Criminal Code which provides:

. . Before rendering judgment or pronouncing sentence the court shall .cause the defendant’s previous criminal record to be submitted to it, including any reports that may have been made as a result of a mental, phychiatric [sic] or physical examination of such person, and may seek' any information that will aid the court in determining the proper treatment of such defendant.”

The Court of Appeals of New York affirmed the conviction and sentence over the' contention that as construed and applied the controlling penal statutes are in violation of the due process clause of the Fourteenth Amendment of the Constitution of the United States “in that the sentence of death was based upon information supplied .by witnesses with whom the accused had not been confronted and as to whom he had no opportunity for cross-examination .or rebuttal . . . .” 298 N. Y. 803, 804, 83 N. E. 2d 698, 699. Because the statutes were sustained over this constitutional challenge the case is here on appeal under 28 U. S. C. § 1257 (2). . .

The harrow contention here makes it unnecessary to set out the facts at length. The record shows a carefully conducted trial lasting more than two • weeks in which appellant was represented by three appointed lawyers who conducted his defense with fidelity and zeal. -.The evidence proved a wholly indefensible murder committed by a person engaged in a burglary. The judge instructed the jury that if it returned a verdict of guilty as charged, without recommendation for- life sentence, “The Court must impose the death penalty,” but if such recommendation was made, “the Court may impose a life sentence.” The judge went on to emphasize that “the Court is not bound to accept your recommendation.”

*244About five weeks after the verdict of guilty with recommendation of life imprisonment, and after a statutory pre-sentence investigation report to the judge, the defendant was brought to court to be sentenced. Asked what he had to say, appellant protested his innocence. After each of his three lawyers had appealed to the court to accept the jury’s recommendation of a life sentence, the judge gave reasons why he felt that the- death sén-r tence should be imposed. He narrated the shocking details of the crime as shown by the trial evidence, expressing his own complete belief in appellant’s guilt. He stated that the pre-sentence investigation revealed many material facts concerning appellant’s background which though relevant to the question of punishment, could not properly have been brought to the attention of the jury in its consideration of the question of guilt.' He referred to the experience -appellant “had had on thirty other burglaries in and about the same vicinity” where the murder had been committed. The appellant had not been convicted of these burglaries although the judge' had information that he had confessed to some and had been identified as the perpetrator of some of the others. The- judge also referred to certain activities of appellant as shown by the probation report that indicated appellant possessed “a morbid sexuality” and classified him as a “menace to society.” : The accuracy of the statements made by the judge as to appellant’s background and past practices was not challenged by appellant or his counsel, nor was the judge asked to disregard apy of them or to afford appellant a chance to refute or discredit any of them by cross-examination or otherwise.-

The case presents a serious and difficult question. The question relates to the rules of evidence- applicable to the manner in which á judge may obtain information to guide him in the imposition of sentence upon an already convicted defendant. Within limits fixed by statutes, *245New York judges are given a broad discretion to decide the type and extent of punishment for convicted defendants. Here, for example, the judge’s discretion was to sentence to life imprisonment or death. To aid a judge in exercising this discretion intelligently the New York procedural policy encourages him to consider information about the convicted person’s past life, health, habits, conduct, and mental and moral propensities. The sentencing judge may consider such informátion even though obtained outside the courtroom from persons whom a defendant has not been permitted to confront of cross-examine. It is the consideration of information obtained by a sentencing judge in this manner that is the basis for appellant’s broad constitutional challenge to the New York statutory policy.

Appellant urges that the New York statutory policy is in irreconcilable conflict with the underlying philosophy of a second procedural policy grounded in the due process of law clause of the Fourteenth Amendment. That policy as stated in In re Oliver, 333 U. S. 257, 273, is in part that no person shall be tried and convicted of an offense unless he is given reasonable notice of the charges against him and is afforded an opportunity to examine adverse witnesses.3 That the due process clause does provide these salutary and time-tested protections where the question for consideration .is the guilt of a defendant seems entirely clear from the genesis and historical evolution of the clause. See, e. g., Chambers v. Florida, 309 U. S. 227, 236-237, and authorities cited in note 10.

*246Tribunals passing on the guilt Of a defendant always have been hedged in by strict evidentiary procedural limitations.4 But both before and since the American colonies became a nation, courts in this country and in England practiced a policy under which a sentencing judge could exercise a wide discretion in the sources and types of evidence used to assist him in determining the kind and extent of punishment to be imposed within limits fixed by law.5 *Out-of-court affidavits have been used frequently, and of course in the smaller communities sentencing judges naturally have in mind their knowledge of the personalities and backgrounds of convicted offenders.6 A recent manifestation of the historical latitude allowed sentencing judges appears in Rule 32 of the Federal Rules of Criminal Procedure. That rule provides for consideration by federal judges of reports made by probation officers containing information about a convicted defendant, including such information “as may be helpful in imposing sentence or in granting probation or in the correctional treatment of the defendant . ...” 7

In addition to the historical basis for different evidentiary rules governing trial and sentencing procedures there are sound practical reasons for the distinction. In a trial before ‘verdict the issue is whether a defendant is guilty of having engaged in certain criminal conduct of which he has been specifically accused. Rules of evi*247dence have been fashioned for criminal trials which narrowly confine the trial contest to evidence that is strictly relevant to the particular offense charged. These rules rest in part on a necessity to prevent a time-consuming and confusing trial of collateral issues. They were also designed to prevent tribunals concerned solely with the issue of guilt of a particular offense from being influenced to convict for that offense by evidence that the defendant had habitually engaged in other misconduct. A sentencing judge, however, is not confined to the narrow issue of guilt. His task within fixed statutory or constitutional limits is to determine the type and extent of punishment after the issue of guilt has been determined. Highly relevant — if not essential — to his selection of an appropriate sentence is the possession of the fullest information possible concerning the defendant’s life and characteristics.8 And modern concepts individualizing punishment have made it all the more necessary that a sentencing judge not be denied an opportunity to obtain pertinent information by a requirement of rigid adherence to restrictive rules of evidence properly applicable to the trial.

Undoubtedly the New York statutes emphasize a prevalent modern philosophy of penology that the punishment should fit the offender and not merely the crime. People v. Johnson, 252 N. Y. 387, 392, 169 N. E. 619, 621. The belief no longer prevails that every offense in a like legal category calls for an identical punishment without regard to the past life and habits of a particular offender. This whole country ha? traveled far from the period in which the death sentence was an automatic and commonplace result of convictions — even for offenses today deemed *248trivial.9. --TeriayVphilastmhy of individualizing sentences makes sharp distinctions for example between~fTrst and repeated- offenders.10 Indeterminate sentences the ultimate termination of which are sometimes decided by nonjudicial'agencies have to a large extent taken the place of the old rigidly fixed punishments. 11 The practicó of probation which relies heavily on non-judicial implementation has been accepted as a wise policy.12 Execution'of the United States parole system rests on the discretion of an administrative parole board. 36 Stat. 819, 18 U. S. C. §§ 714, 716, now §§ 4202, 4203. Retribution' is no longer the dominant objective of the criminal law. Reformation and rehabilitation of offenders have become important goals of criminal jurisprudence.13

Modern changes in the treatment of offenders make it more necessary now than a century ago for observance *249of the distinctions in the evidential procedure in the trial and sentencing processes.//For indeterminate sentences and probation have resulted in an increase in the •discretionary powers exercised in fixing punishments., In general, these modern changes have not resulted in making the lot of offenders harder. On the contrary a strong motivating force for the changes has been the belief that by careful study of the lives and personalities of convicted offenders many could be less severely punished and restored sooner to complete freedom and useful citizenship. This belief to a large extent has been justified.

Under the practice of individualizing, punishments, investigational techniques have been given an important role. ^Probation workers making reports of their investigations have not been trained to — orosecuie but to aid, ^offenders?] Their reports have been given a high value by conscientious judges who want to sentence persons on the best available information rather than on guesswork and inadequate information.14 To deprive sentenc*250ing judges of this kind of information would undermine modern penological procedural policies that have been cautiously adopted throughout the nation after careful consideration and experimentation. - We must recognize that most of the information now relied upon by judges to guide them in the intelligent imposition of sentences would be unavailable if information were restricted to that given in open court by witnesses subject to cross-examination. And the modern probation report draws on information concerning every aspect of a defendant’s life.15 The type and' extent of this information make totally impractical if not impossible open court testimony with cross-exámination. Such a procedure could endlessly delay criminal administration- in a retrial of collateral issues.

The considerations we have set out admonish us against" treating the due process clause as a uniform command that courts throughout the Nation abandon their age-old *251practice of seeking information from out-of-court sources to guide their judgment toward a more enlightened and just sentence. New York criminal statutes set wide limits for maximum and minimum sentences.16 Under New York statutes a state judge cannot escape his grave responsibility of fixing sentence.' In determining whether a defendant shall receive a one-year minimum or a twenty-year maximum sentence, we do not, think the — Federal Constitution restricts the view of the sentencing judge to thelnTfMñaUon receiygcTTn open court. The due process clause should not be treated aslTdevice for freezing the evidential procedure of sentencing in the mold of trial procedure. So to treat the due process clause would hinder if not preclude all courts — state and federal — from making progressive efforts to improve the administration of criminal justice.

It is urged, however, that we should draw a constitutional distinction as to the procedure for obtaining information where the death sentence is imposed. We cannot accept the contention. Leaving a sentencing judge free to avail himself of out-of-court information in making such a fateful choice of sentences does secure to him a broad discretionary power, one susceptible of abuse. But in considering whether a rigid constitutional barrier should be created, it must be remembered that there is possibility of abuse wherever a judge must choose between life iiúprisonment and death. And it is con*252ceded that no federal constitutional objection would have been possible if the judge here had sentenced appellant to death because appellant’s trial manner impressed the judge that appellant was a bad risk for society, or. if the judge had sentenced him to death giving no reason at all. We cannot'say that the düe process clause renders a sentence void merely because a judge , gets additional out-of-court information to assist him in the exercise of this awesome power of imposing' the death' sentence. •

Appellant was found guilty after a fairly conducted trial. His sentence followed a hearing conducted by the judge. Upon the judge’s inquiry as to why sentence should not bd imposed, the defendant made statements. His counsel made extended arguments. The case went to the highest court in the state, and that court had power to reverse for abuse of discretion or legal error in the imposition of the sentence.17' That court affirmed. We hold thaf appellant was not denied due process of law.18 '

Affirmed.

Mr. Justice Rutledge dissents.

Mr. Justice Murphy,

dissenting.

A combination of factors in this case impels me to dissent.'

Petitioner was convicted of murder by a jury, and sentenced to death by the judge. The jury which heard the *253trial unanimously recommended life imprisonment as a suitable punishment for the defendant. They had observed him throughout the trial, had heard all the evidence adduced against him, and in spite of the shocking character of the crime of which they found him guilty, were unwilling to decree that his life should be taken. In our criminal courts the jury sits as the representative of the community; its voice is that of the society against which the crime was committed. A judge, even though vested with statutory authority to do so, should hesitate indeed to increase the sé verity of such a community expression.

He should be willing to increase it, moreover, only with the most scrupulous regard for the rights of the defendant. The record before us indicates that the judge exercised his discretion to deprive a man of his life, in reliance on material madp available to him in a probation report, consisting almost entirely of évidence that would have been inadmissible at the trial. Some, such ás allegations of prior crimes, was irrelevant. Much was incompetent as hearsay. All was damaging, and none was subject to scrutiny by the defendant.

Due process of law includes at least the idea that a person accused of crime shall be accorded a fair hearing through all the stages of the proceedings against him. I agree with the Court t.s to the value and humaneness of liberal use of probation reports as developed by modern penologists, but, in a capital case, against the unanimous recommendation of a jury,- where the report would concededly not have been admissible at the trial, and was not subject to examination by the defendant, I am forced to conclude that the high commands of due process were not obeyed.

*298Justice Scalia

delivered the opinion of the Court.

Petitioner Ralph Howard Blakely, Jr., pleaded guilty to the kidnaping of his estranged wife. The facts admitted in his plea, standing alone, supported a maximum sentence of 53 months. Pursuant to state law, the court imposed an “exceptional” sentence of 90 months after making a judicial determination that he had acted with “deliberate cruelty.” App. 40, 49. We consider whether this violated petitioner’s Sixth Amendment right to trial by jury.

I

Petitioner married his wife Yolanda in 1973. He was evidently a difficult man to live with, having been diagnosed at various times with psychological and personality disorders including paranoid schizophrenia. His wife ultimately filed for divorce. In 1998, he abducted her from their orchard home in Grant County, Washington, binding her with duct tape and forcing her at knifepoint into a wooden box in the bed of his pickup truck. In the process, he implored her to dismiss the divorce suit and related trust proceedings.

When the couple’s 13-year-old son Ralphy returned home from school, petitioner ordered him to follow in another car, threatening to harm Yolanda with a shotgun if he did not do so. Ralphy escaped and sought help when they stopped at a gas station, but petitioner continued on with Yolanda to a friend’s house in Montana. He was finally arrested after the friend called the police.

The State charged petitioner with first-degree kidnaping, Wash. Rev. Code Ann. § 9A.40.020(1) (200Q).1 Upon reaching a plea agreement, however, it reduced the charge to second-degree kidnaping involving domestic violence and use *299of a firearm, see §§9A.40.030(1), 10.99.020(3)(p), 9.94A.125.2 Petitioner entered a guilty plea admitting the elements of second-degree kidnaping and the domestic-violence and firearm allegations, but no other relevant facts.

The case then proceeded to sentencing. In Washington, second-degree kidnaping is a class B felony. § 9A.40.030(3). State law provides that “[n]o person convicted of a [class B] felony shall be punished by confinement... exceeding ... a term of ten years.” § 9A.20.021(1)(b). Other provisions of state law, however, further limit the range of sentences a judge may impose. Washington’s Sentencing Reform Act specifies, for petitioner’s offense of second-degree kidnaping with a firearm, a “standard range” of 49 to 53 months. See §9.94A.320 (seriousness level V for second-degree kidnaping); App. 27 (offender score 2 based on §9.94A.360); § 9.94A.310(1), box 2-V (standard range of 13-17 months); § 9.94A.310(3)(b) (36-month firearm enhancement).3 A judge may impose a sentence above the standard range if he finds “substantial and compelling reasons justifying an exceptional sentence.” § 9.94A.120(2). The Act lists aggravating factors that justify such a departure, which it recites to be illustrative rather than exhaustive. § 9.94A.390. Nevertheless, “[a] reason offered to justify an exceptional sentence can be considered only if it takes into account factors other than those which are used in computing the standard range sentence for the offense.” State v. Gore, 143 Wash. 2d 288, 315—316, 21 P. 3d 262, 277 (2001). When a judge imposes an exceptional sentence, he must set forth findings of fact and conclusions of law supporting it. § 9.94A. 120(3). A review*300ing court will reverse the sentence if it finds that “under a clearly erroneous standard there is insufficient evidence in the record to support the reasons for imposing an exceptional sentence.” Id., at 315, 21 P. 3d, at 277 (citing § 9.94A.210(4)).

Pursuant to the plea agreement, the State recommended a sentence within the standard range of 49 to 53 months. After hearing Yolanda’s description of the kidnaping, however, the judge rejected the State’s recommendation and imposed an exceptional sentence of 90 months — 37 months beyond the standard maximum. He justified the sentence on the ground that petitioner had acted with “deliberate cruelty,” a statutorily enumerated ground for departure in domestic-violence cases. § 9.94A.390(2)(h)(iii).4

Faced with an unexpected increase of more than three years in his sentence, petitioner objected. The judge accordingly conducted a 3-day bench hearing featuring testimony from petitioner, Yolanda, Ralphy, a police officer, and medical experts. After the hearing, he issued 32 findings of fact, concluding:

“The defendant’s motivation to commit kidnapping was complex, contributed to by his mental condition and personality disorders, the pressures of the divorce litigation, the impending trust litigation trial and anger over his troubled interpersonal relationships with his spouse and children. While he misguidedly intended to forcefully reunite his family, his attempt to do so was subservient to his desire to terminate lawsuits and modify title ownerships to his benefit.

*301“The defendant’s methods were more homogeneous than his motive. He used stealth and surprise, and took advantage of the victim’s isolation. He immediately employed physical violence, restrained the victim with tape, and threatened her with injury and death to herself and others. He immediately coerced the victim into providing, information by the threatening application of a knife. He violated a subsisting restraining order.” App. 48-49.

The judge adhered to his initial determination of deliberate cruelty.

Petitioner appealed, arguing that this sentencing procedure deprived him of his federal constitutional right to have a jury determine beyond a reasonable doubt all facts legally essential to his sentence. The State Court of Appeals affirmed, 111 Wash. App. 851, 870-871, 47 P. 3d 149, 159 (2002), relying on the Washington Supreme Court’s rejection of a similar challenge in Gore, supra, at 311-315, 21 P. 3d, at 275-277. The Washington Supreme Court denied discretionary review. 148 Wash. 2d 1010, 62 P. 3d 889 (2003). We granted certiorari. 540 U. S. 965 (2003).

II

This case requires us to apply the rule we expressed in Apprendi v. New Jersey, 530 U. S. 466, 490 (2000): “Other than the fact of a prior conviction, any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury, and proved beyond a reasonable doubt.” This rule reflects two longstanding tenets of common-law criminal jurisprudence: that the “truth of every accusation” against a defendant “should afterwards be confirmed by the unanimous suffrage of twelve of his equals and neighbours,” 4 W. Blackstone, Commentaries on the Laws of England 343 (1769), and that “an accusation which lacks any particular fact which the law makes essential to the punishment is ... no accusation within the requirements *302of the common law, and it is no accusation in reason,” 1 J. Bishop, Criminal Procedure § 87, p. 55 (2d ed. 1872).5 These principles have been acknowledged by courts and treatises since the earliest days of graduated sentencing; we compiled the relevant authorities in Apprendi, see 530 U. S., at 476-483, 489-490, n. 15; id., at 501-518 (Thomas, J., concurring), and need not repeat them here.6

*303Apprendi involved a New Jersey hate-crime statute that authorized a 20-year sentence, despite the usual 10-year maximum, if the judge found the crime to have been committed “‘with a purpose to intimidate . . . because of race, color, gender, handicap, religion, sexual orientation or ethnicity.’” Id., at 468-469 (quoting N. J. Stat. Ann. §2C:44-3(e) (West Supp. 1999-2000)). In Ring v. Arizona, 536 U. S. 584, 592-593, and n. 1 (2002), we applied Apprendi to an Arizona law that authorized the death penalty if the judge found 1 of 10 aggravating factors. In each case, we concluded that the defendant’s constitutional rights had been violated because the judge had imposed a sentence greater than the maximum he could have imposed under state law without the challenged factual finding. Apprendi, supra, at 491-497; Ring, supra, at 603-609.

In this case, petitioner was sentenced to more than three years above the 53-month statutory maximum of the standard range because he had acted with “deliberate cruelty.” The facts supporting that finding were neither admitted by petitioner nor found by a jury. The State nevertheless contends that there was no Apprendi violation because the relevant “statutory maximum” is not 53 months, but the 10-year maximum for class B felonies in §9A.20.021(1)(b). It observes that no exceptional sentence may exceed that limit. See § 9.94A.420. Our precedents make clear, however, that the “statutory maximum” for Apprendi purposes is the maximum sentence a judge may impose solely on the basis of the fa,cts reflected in the jury verdict or admitted by the defendant. See Ring, supra, at 602 (“ ‘the maximum he would receive if punished according to the facts reflected in the jury verdict alone’ ” (quoting Apprendi, supra, at 483)); Harris v. United States, 536 U. S. 545, 563 (2002) (plurality opinion) (same); cf. Apprendi, supra, at 488 (facts admitted by the defendant). In other words, the relevant “statutory maximum” is not the maximum sentence a judge may impose *304after finding additional facts, but the maximum he may impose without any additional findings. When a judge inflicts punishment that the jury’s verdict alone does not allow, the jury has not found all the facts “which the law makes essential to the punishment,” Bishop, supra, §87, at 55, and the judge exceeds his proper authority.

The judge in this case could not have imposed the exceptional 90-month sentence solely on the basis of the facts admitted in the guilty plea. Those facts alone were insufficient because, as the Washington Supreme Court has explained, “[a] reason offered to justify an exceptional sentence can be considered only if it takes into account factors other than those which are used in computing the standard range sentence for the offense,” Gore, 143 Wash. 2d, at 315-316, 21 P. 3d, at 277, which in this case included the elements of second-degree kidnaping and the use of a firearm, see §§ 9.94A.320, 9.94A.310(3)(b).7 Had the judge imposed the 90-month sentence solely on the basis of the plea, he would have been reversed. See § 9.94A.210(4). The “maximum sentence” is no more 10 years here than it was 20 years in Apprendi (because that is what the judge could have imposed upon finding a hate crime) or death in Ring (because that is what the judge could have imposed upon finding an aggravator).

The State defends the sentence by drawing an analogy to those we upheld in McMillan v. Pennsylvania, 477 U. S. 79 (1986), and Williams v. New York, 837 U. S. 241 (1949). Neither case is on point. McMillan involved a sentencing scheme that imposed a statutory minimum if a judge found a particular fact. 477 U. S., at 81. We specifically noted that the statute “does not authorize a sentence in excess of that otherwise allowed for [the underlying] offense.” Id., *305at 82; cf. Harris, supra, at 567. Williams involved an indeterminate-sentencing regime that allowed a judge (but did not compel him) to rely on facts outside the trial record in determining whether to sentence a defendant to death. 337 U. S., at 242-243, and n. 2. The judge could have “sentenced [the defendant] to death giving no reason at all.” Id., at 252. Thus, neither case involved a sentence greater than what state law authorized on the basis of the verdict alone.

Finally, the State tries to distinguish Apprendi and Ring by pointing out that the enumerated grounds for departure in its regime are illustrative rather than exhaustive. This distinction is immaterial. Whether the judge’s authority to impose an enhanced sentence depends on finding a specified fact (as in Apprendi), one of several specified facts (as in Ring), or any aggravating fact (as here), 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.8

Because the State’s sentencing procedure did not comply with the Sixth Amendment, petitioner’s sentence is invalid.9

Ill

Our commitment to Apprendi in this context reflects not just respect for longstanding precedent, but the need to give intelligible content to the right of jury trial. That right is *306no mere procedural formality, but a fundamental reservation of power in our constitutional structure. Just as suffrage ensures the people’s ultimate control in the legislative and executive branches, jury trial is meant to ensure their control in the judiciary. See Letter XV by the Federal Farmer (Jan. 18, 1788), reprinted in 2 The Complete Anti-Federalist 815, 320 (H. Storing ed. 1981) (describing the jury as “securing] to the people at large, their just and rightful controul in the judicial department”); John Adams, Diary Entry (Feb. 12, 1771), reprinted in 2 Works of John Adams 252, 253 (C. Adams ed. 1850) (“[T]he common people, should have as complete a control... in every judgment of a court of judicature” as in the legislature); Letter from Thomas Jefferson to the Abbé Arnoux (July 19, 1789), reprinted in 15 Papers of Thomas Jefferson 282, 283 (J. Boyd ed. 1958) (“Were I called upon to decide whether the people had best be omitted in the Legislative or Judiciary department, I would say it is better to leave them out of the Legislative”); Jones v. United States, 526 U. S. 227, 244-248 (1999). Apprendi carries out this design by ensuring that the judge’s authority to sentence derives wholly from the jury’s verdict. Without that restriction, the jury would not exercise the control that the Framers intended.

Those who would reject Apprendi we resigned to one of two alternatives. The first is that the jury need only find whatever facts the legislature chooses to label elements of the crime, and that those it labels sentencing factors — no matter how much they may increase the punishment — may be found by the judge. This would mean, for example, that a judge could sentence a man for committing murder even if the jury convicted him only of illegally possessing the firearm used to commit it — or of making an illegal lane change while fleeing the death scene. Not even Apprendi’s critics would advocate this absurd result. Cf. 530 U. S., at 552-553 (O’Connor, J., dissenting). The jury could not function as circuitbreaker in the State’s machinery of justice if it were *307relegated to making a determination that the defendant at some point did something wrong, a mere preliminary to a judicial inquisition into the facts of the crime the State actually seeks to punish.10

The second alternative is that legislatures may establish legally essential sentencing factors within limits — limits crossed when, perhaps, the ‘sentencing factor is a “tail which wags the dog of the substantive offense.” McMillan, 477 U. S., at 88. Whát this means in operation is that the law must not go too far — it must not exceed the judicial estimation of the proper role of the judge.

The subjectivity of this standard is obvious. Petitioner argued below that second-degree kidnaping with deliberate cruelty was essentially the same as first-degree kidnaping, the very charge he had avoided by pleading to a lesser offense. The court conceded this might be so but held it irrelevant. See 111 Wash. App., at 869,47 P. 3d, at 158.11 Petitioner’s 90-month sentence exceeded the 53-month standard maximum by almost 70%; the Washington Supreme Court in other cases has upheld exceptional sentences 15 times the standard maximum. See State v. Oxborrow, 106 Wash. 2d 525, 528, 533, 723 P. 2d 1123, 1125, 1128 (1986) (en banc) (15-year exceptional sentence; 1-year standard maximum sen*308tence); State v. Branch, 129 Wash. 2d 635, 650, 919 P. 2d 1228, 1235 (1996) (en banc) (4-year exceptional sentence; 3-month standard maximum sentence). Did the court go too far in any of these cases? There is no answer that legal analysis can provide. With too far as the yardstick, it is always possible to disagree with such judgments and never to refute them.

Whether the Sixth Amendment incorporates this manipulate standard rather than Apprendi’s bright-line rule depends on the plausibility of the claim that the Framers would have left definition of the scope of jury power up to judges’ intuitive sense of how far is too far. We think that claim not plausible at all, because the very reason the Framers put a jury-trial guarantee in the Constitution is that they were unwilling to trust government to mark out the role of the jury.

IV

By reversing the judgment below, we are not, as the State would have it, “finding] determinate sentencing schemes unconstitutional.” Brief for Respondent 34. This case is not about whether determinate sentencing is constitutional, only about how it can be implemented in a way that respects the Sixth Amendment. Several policies prompted Washington’s adoption of determinate sentencing, including proportionality to the gravity of the offense and parity among defendants. See Wash. Rev. Code Ann. §9.94A.010 (2000). Nothing we have said impugns those salutary objectives.

Justice O’Connor argues that, because determinate-sentencing schemes involving judicial factfinding entail less judicial, discretion than indeterminate schemes, the constitutionality of the latter implies the constitutionality of the former. Post, at 314-323. This argument is flawed on a number of levels. First, the Sixth Amendment by its terms is not a limitation on judicial power, but a reservation of jury power. It limits judicial power only to the extent that the claimed judicial power infringes on the province of the jury. *309Indeterminate sentencing does not do so. It increases judicial discretion, to be sure, but not at the expense of the jury’s traditional function of finding the facts essential to lawful imposition of the penalty. Of course indeterminate schemes involve judicial factfinding, in that a judge (like a parole board) may implicitly rule on those facts he deems important to the exercise of his sentencing discretion. But the facts do not pertain to whether the defendant has a legal right to a lesser sentence — and that makes all the difference insofar as judicial impingement upon the traditional role of the jury is concerned. In a system that says the judge may punish burglary with 10 to 40 years, every burglar knows he is risking 40 years in jail. In a system that punishes burglary with a 10-year sentence, with another 30 added for use of a gun, the burglar who enters a home unarmed is entitled to no more than a 10-year sentence — and by reason of the Sixth Amendment the facts bearing upon that entitlement must be found by a jury.

But even assuming that restraint of judicial power unrelated to the jury’s role is a Sixth Amendment objective, it is far from clear that Apprendi disserves that goal. Determinate judicial-factfinding schemes entail less judicial power than indeterminate schemes, but more judicial power than determinate jury-factfinding schemes. Whether Apprendi increases judicial power overall depends on what States with determinate judicial-factfinding schemes would do, given the choice between the two alternatives. Justice O’Connor simply assumes that the net effect will favor judges, but she has no empirical basis for that prediction. Indeed, what evidence we have points exactly the other way: When the Kansas Supreme Court found Apprendi infirmities in that State’s determinate-sentencing regime in State v. Gould, 271 Kan. 394, 404-414, 23 P. 3d 801, 809-814 (2001), the legislature responded not by reestablishing indeterminate sentencing but by applying Apprendi's requirements to its current regime. See Act of May 29, 2002, ch. 170, 2002 Kan. Sess. *310Laws pp. 1018-1023 (codified at Kan. Stat. Ann. §21-4718 (2003 Cum. Supp.)); Brief for Kansas Appellate Defender Office as Amicus Curiae 3-7. The result was less, not more, judicial power.

Justice Breyer argues that Apprendi works to the detriment of criminal defendants who plead guilty by depriving them of the opportunity to argue sentencing factors to a judge. Post, at 331. But nothing prevents a defendant from waiving his Apprendi rights. When a defendant pleads guilty, the State is free to seek judicial sentence enhancements so long as the defendant either stipulates to the relevant facts or consents to judicial factfinding. See Ap-prendi, 530 U. S., at 488; Duncan v. Louisiana, 391 U. S. 145, 158 (1968). If appropriate waivers are procured, States may continue to offer judicial factfinding as a matter of course to all defendants who plead guilty. Even a defendant who stands trial may consent to judicial factfinding as to sentence enhancements, which may well be in his interest if relevant evidence would prejudice him at trial. We do not understand how Apprendi can possibly work to the detriment of those who are free, if they think its costs outweigh its benefits, to render it inapplicable.12

*311Nor do we see any merit to Justice Breyer’s contention that Apprendi is unfair to criminal defendants because, if States respond by enacting “17-element robbery crime[s],” prosecutors will have more elements with which to bargain. Post, at 331, 335-336 (citing Bibas, Judicial Fact-Finding and Sentence Enhancements in a World of Guilty Pleas, 110 Yale L. J. 1097 (2001)). Bargaining already exists with regard to sentencing factors because defendants can either stipulate or contest the facts thát make them applicable. If there is any difference between bargaining over sentencing factors and bargaining over elements, the latter probably favors the defendant. Every new element that a prosecutor can threaten to charge is also an element that a defendant can threaten to contest at trial and make the prosecutor prove beyond a reasonable doubt. Moreover, given the sprawling scope of most criminal codes, and the power to affect sentences by making (even nonbinding) sentencing recommendations, there is already no shortage of in terrorem tools at prosecutors’ disposal. See King & Klein, Apprendi and Plea Bargaining, 54 Stan. L. Rev. 295, 296 (2001) (“Every prosecuto-rial bargaining chip mentioned by Professor Bibas existed pro-Apprendi exactly as it does post-Apprendi”).

Any evaluation of Apprendis “fairness” to criminal defendants must compare it with the regime it replaced, in which a defendant, with no warning in either his indictment or plea, would routinely see his maximum potential sentence balloon from as little as five years to as much as life imprisonment, see 21 U. S. C. §§ 841(b)(1)(A), (D),13 based not on *312facts proved to his peers beyond a reasonable doubt, but on facts extracted after trial from a report compiled by a probation officer who the judge thinks more likely got it right than got it wrong. We can conceive of no measure of fairness that would find more fault in the utterly speculative bargaining effects Justice Breyer identifies than in the regime he champions. Suffice it to say that, if such a measure exists, it is not the one the Framers left us with.

The implausibility of Justice Breyer’s contention that Apprendi is unfair to criminal defendants is exposed by the lineup of amici in this case. It is hard to believe that the National Association of Criminal Defense Lawyers was somehow duped into arguing for the wrong side. Justice Breyer’s only authority asking that defendants be protected from Apprendi is an article written not by a criminal defense lawyer but by a law professor and former prosecutor. See post, at 331 (citing Bibas, supra)', Association of American Law Schools Directory of Law Teachers 2003-2004, p. 319.

Justice Breyer also claims that Apprendi will attenuate the connection between “real criminal conduct and real punishment” by encouraging plea bargaining and by restricting alternatives to adversarial factfinding. Post, at 334, 338-339. The short answer to the former point (even assuming the questionable premise that Apprendi does encourage plea bargaining, but see supra, at 310, n. 12) is that the Sixth Amendment was not written for the benefit of those who choose to forgo its protection. It guarantees the right to jury trial. It does not guarantee that a particular number of jury trials will actually take place. That more defendants elect to waive that right (because, for example, government at the moment is not particularly oppressive) does not prove that a constitutional provision guaranteeing availability of that option is disserved.

Justice Breyer’s more general argument — that Apprendi undermines alternatives to adversarial factfinding— *313is not so much a criticism of Apprendi as an assault on jury trial generally. His esteem for “nonadversarial” truth-seeking processes, post, at 339, supports just as well an argument against either. Our Constitution and the common-law traditions it entrenches, however, do not admit the contention that facts are better discovered by judicial inquisition than by adversarial testing before a jury. See 3 Blackstone, Commentaries, at 373-374, 379-381. Justice Breyer may be convinced of the equity of the regime he favors, but his views, are not the ones we are bound to uphold.

Ultimately, our decision cannot turn on whether or to what degree trial by jury impairs the efficiency or fairness of criminal justice. One can certainly argue that both these values would be better served by leaving justice entirely in the hands of professionals; many nations of the world, particularly those following civil-law traditions, take just that course. There is not one shred of doubt, however, about the Framers’ paradigm for criminal justice: not the civil-law ideal of administrative perfection, but the common-law ideal of limited state power accomplished by strict division of authority between judge and jury. As Apprendi held, every defendant has the right to insist that the prosecutor prove to a jury all facts legally essential to the punishment. Under the dissenters’ alternative, he has no such right. That should be the end of the matter.

* * *

Petitioner was sentenced to prison for more than three years beyond what the law allowed for the crime to which he confessed, on the basis of a disputed finding that he had acted with “deliberate cruelty.” The Framers would not have thought it too much to demand that, before depriving a man of three more years of his liberty, the State should suffer the modest inconvenience of submitting its accusation to “the unanimous suffrage of twelve of his equals and neigh-*314bours,” 4 Blackstone, supra, at 343, rather than a lone employee of the State.

The judgment of the Washington Court of Appeals is reversed, and the case is remanded for further proceedings not inconsistent with this opinion.

It is so ordered.

Justice O’Connor,

with whom Justice Breyer joins, and with whom The Chief Justice and Justice Kennedy join as to all but Part IV-B, dissenting.

The legacy of today’s opinion, whether intended or not, will be the consolidation of sentencing power in the State and Federal Judiciaries. The Court says to Congress and state legislatures: If you want to constrain the sentencing discretion of judges and bring some uniformity to sentencing, it will cost you — dearly. Congress and States, faced with the burdens imposed by the extension of Apprendi v. New Jersey, 530 U. S. 466 (2000), to the present context, will either trim or eliminate altogether their sentencing guidelines schemes and, with them, 20 years of sentencing reform. It is thus of little moment that the majority does not expressly declare guidelines schemes unconstitutional, ante, at 308; for, as residents of “Apprendi-land” are fond of saying, “the relevant inquiry is one not of form, but of effect.” Ap-prendi, supra, at 494; Ring v. Arizona, 536 U. S. 584, 613 (2002) (Scalia, J., concurring). The “effect” of today’s decision will be greater judicial discretion and less uniformity in sentencing. Because I find it implausible that the Framers would have considered such a result to be required by the Due Process Clause or the Sixth Amendment, and because the practical consequences of today’s decision may be disastrous, I respectfully dissent.

I

One need look no further than the history leading up to and following the enactment of Washington’s guidelines *315scheme to appreciate the damage that today’s decision will cause. Prior to 1981, Washington, like most other States and the Federal Government, employed an indeterminate sentencing scheme. Washington’s criminal code separated all felonies into three broad categories: “class A,” carrying a sentence of 20 years to life; “class B,” carrying a sentence of 0 to 10 years; and “class C,” carrying a sentence of 0 to 5 years. Wash. Rev. Code Ann. §9A.20.020 (2000); see also Sentencing Reform Act of 1981, 1981 Wash. Laws ch. 137, p. 534. Sentencing judges, in conjunction with parole boards, had virtually unfettered discretion to sentence defendants to prison terms falling anywhere within the statutory range, including probation — ! e., no jail sentence at all.' Wash. Rev. Code Ann. §§9.95.010-9.95.011; Boerner & Lieb, Sentencing Reform in the Other Washington, 28 Crime and Justice 71, 73 (M. Tonry ed. 2001) (hereinafter Boerner & Lieb) (“Judges were authorized to choose between prison and probation with few exceptions, subject only to review for abuse of discretion”). See also D. Boerner, Sentencing in Washington § 2.4, pp. 2-27 to 2-28 (1985).

This system of unguided discretion inevitably resulted in severe disparities in sentences received and served by defendants committing the same offense and having similar criminal histories. Boerner & Lieb 126-127; cf. S. Rep. No. -98-225, p. 38 (1983) (Senate Report on precursor to federal Sentencing Reform Act of 1984) (“[Ejvery day Federal judges mete out an unjustifiably wide range of sentences to offenders with similar histories, convicted of similar crimes, committed under similar circumstances. . . . These disparities, whether they occur at the time of the initial sentencing or at the parole stage, can be traced directly to the unfettered discretion the law confers on . those judges and parole authorities responsible for imposing and implementing the sentence”). Indeed, rather than reflect legally relevant criteria, these disparities too often were correlated with constitutionally suspect variables such as race. Boerner & Lieb *316126-128. See also Breyer, The Federal Sentencing Guidelines and the Key Compromises Upon Which They Rest, 17 Hofstra L. Rev. 1, 5 (1988) (elimination of racial disparity one reason behind Congress’ creation of the Federal Sentencing Commission).

To counteract these trends, the state legislature passed the Sentencing Reform Act of 1981. The Act had the laudable purposes of “mak[ing] the criminal justice system accountable to the public,” and “[e]nsur[ing] that the punish^ ment for a criminal offense is proportionate to the seriousness of the offense . . . [and] commensurate with the punishment imposed on others committing similar offenses.” Wash. Rev. Code Ann. § 9.94A.010 (2000). The Act neither increased any of the statutory sentencing ranges for the three types of felonies (though it did eliminate the statutory mandatory minimum for class A felonies), nor reclassified any substantive offenses. 1981 Wash. Laws ch. 187, p. 534. It merely placed meaningful constraints on discretion to sentence offenders within the statutory ranges, and eliminated parole. There is thus no evidence that the legislature was attempting to manipulate the statutory elements of criminal offenses or to circumvent the procedural protections of the Bill of Rights. Rather, lawmakers were trying to bring some much-needed uniformity, transparency, and accountability to an otherwise “ ‘labyrinthine’ sentencing and corrections system that ‘lack[ed] any principle except unguided discretion.’” Boerner & Lieb 73 (quoting F. Zimring, Making the Punishment Fit the Crime: A Consumers’ Guide to Sentencing Reform, Occasional Paper No. 12, p. 6 (1977)).

II

Far from disregarding principles of due process and the jury trial right, as the majority today suggests, Washington’s reform has served them. Before passage of the Act, a defendant charged with second degree kidnaping, like petitioner, had no idea whether he would receive a 10-year sen*317tence or probation. The ultimate sentencing determination could turn as much on the idiosyncracies of a particular judge as on the specifics of the defendant’s crime or background. A defendant did not know what facts, if any, about his offense or his history would be considered relevant by the sentencing judge or by the parole board. After passage of the Act, a defendant charged with second degree kidnaping knows what his presumptive sentence will be; he has a good idea of the types of factors that a sentencing judge can and will consider when deciding whether to sentence him outside that range; he is guaranteed meaningful appellate review to protect against a.n arbitrary sentence. Boerner & Lieb 93 (“By consulting one sheet, practitioners could identify the applicable scoring rules for criminal history, the sentencing range, and the available sentencing options for each case”). Criminal defendants still face the same statutory maximum sentences, but they now at least know, much more than before, the real consequences of their actions.

Washington’s move to a system of guided discretion has served equal protection principles as well. Over the past 20 years, there has been a substantial reduction in racial disparity in sentencing across the State. Id., at 126 (Racial disparities that do exist “are accounted for by differences in legally relevant variables — the offense of conviction and prior criminal record”); id., at 127 (“[J]udicial authority to impose exceptional sentences under the court’s departure, authority shows little evidence of disparity correlated with race”). The reduction is directly traceable to the constraining effects of the guidelines — namely, their “presumptive range[s]” and limits on the imposition of “exceptional sentences” outside of those ranges. Id., at 128. For instance, sentencing judges still retain unreviewable discretion in first-time offender cases and in certain sex offender cases to impose alternative sentences that are far more lenient than those contemplated by the guidelines. To the extent that unjustifiable racial disparities have persisted in Washington, it *318has been in the imposition of such alternative sentences: “The lesson is powerful: racial disparity is correlated with unstructured and unreviewed discretion.” Ibid.; see also Washington State Minority and Justice Commission, R. Crutchfield, J. Weis, R. Engen, & R. Gainey, Raeial/Ethnie Disparities and Exceptional Sentences in Washington State, Final Report 51-53 (Sept. 1993) (“[Ejxceptional sentences are not a major source of racial disparities in sentencing”).

The majority does not, because it cannot, disagree that determinate sentencing schemes, like Washington’s, serve important constitutional values. Ante, at 308. Thus, the majority says: “This case is not about whether determinate sentencing is constitutional, only about how it can be implemented in a way that respects the Sixth Amendment.” Ibid. But extension of Apprendi to the present context will impose significant costs on a legislature’s determination that a particular fact, not historically an element, warrants a higher sentence. While not a constitutional prohibition on guidelines schemes, the majority’s decision today exacts a substantial constitutional tax.

The costs are substantial and real. Under the majority’s approach, any fact that increases the upper bound, on a judge’s sentencing discretion is an element, of the offense. Thus, facts that historically have been taken into account by sentencing judges to assess a sentence within a broad range — such as drug quantity, role in the offense, risk of bodily harm — all must now be charged in an indictment and submitted to a jury, In re Winship, 397 U. S. 358 (1970), simply because it is the legislature, rather than the judge, that constrains the extent to which such facts may be used to impose a sentence within a pre-existing statutory range.

While that, alone is enough to threaten the continued use of sentencing guidelines schemes, there are additional costs. For example, a legislature might rightly think that some factors bearing on sentencing, such as prior bad acts or criminal history, should not be considered in a jury’s determination of *319a defendant’s guilt — such “character evidence” has traditionally been off. limits during the guilt phase of criminal proceedings because of its tendency to inflame the passions of the jury. See, e. g., Fed. Rule Evid. 404; 1 E. Imwinkelried, P. Giannelli, F. Gilligan, & F. Lederer, Courtroom Criminal Evidence 285 (3d ed. 1998). If a legislature desires uniform consideration of such factors at sentencing, but does not want them to impact a jury’s initial determination of guilt, the State may have to bear the additional expense of a separate, full-blown jury trial during the penalty phase proceeding.

Some facts that bear on sentencing either will not be discovered, or are not discoverable, prior to trial. For instance, a legislature might desire that defendants who act in an obstructive manner during trial or post-trial proceedings receive a greater sentence than defendants who do not. See, e. g., United States Sentencing Commission, Guidelines Manual §3C1.1 (Nov. 2003) (hereinafter USSG) (2-point increase in offense level for obstruction of justice). In such cases, the violation arises too late for the State to provide notice to the defendant or to argue the facts to the jury. A State wanting to make such facts relevant at sentencing must now either vest sufficient discretion in the judge to account for them or bring a separate criminal prosecution for obstruction of justice or perjury. And, the latter option is available only to the extent that a defendant’s obstructive behavior is so severe as to constitute an already-existing separate offense, unless the legislature is willing to undertake the unlikely expense of criminalizing relatively minor obstructive behavior.

Likewise, not all facts that historically have been relevant to sentencing always will be known prior to trial. For instance, trial or sentencing proceedings of a drug distribution defendant might reveal that he sold primarily to children. Under the majority’s approach, a State wishing such a revelation to result in a higher sentence within a pre-existing statutory range either must vest judges with sufficient dis*320cretion to account for it (and trust that they exercise that discretion) or bring a separate criminal prosecution. Indeed, the latter choice might not be available — a separate prosecution, if it is for an aggravated offense, likely would be barred altogether by the Double Jeopardy Clause. Blockburger v. United States, 284 U. S. 299 (1932) (government cannot prosecute for separate offenses unless each offense has at least one element that the other does not).

The majority may be correct that States and the Federal Government will be willing to bear some of these costs. Ante, at 309-310. But simple economics dictate that they will not, and cannot, bear them all. To the extent that they do not, there will be an inevitable increase in judicial discretion with all of its attendant failings.1

Ill

Washington’s Sentencing Reform Act did not alter the statutory maximum sentence to which petitioner was exposed. See Wash. Rev. Code Ann. § 9A.40.030 (2003) (second *321degree kidnaping class B felony since 1975); see also State v. Pawling, 23 Wash. App. 226, 228-229, 597 P. 2d 1367, 1369 (1979) (citing second degree kidnaping provision as existed in 1977). Petitioner was informed in the charging document, his plea agreement, and during his plea hearing that he faced a potential statutory maximum of 10 years in prison. App. 63, 66, 76. As discussed above, the guidelines served due process by providing notice to petitioner of the consequences of his acts; they vindicated his jury trial right by informing him of the stakes of risking trial; they served equal protection by ensuring petitioner that invidious characteristics such as race would not impact his sentence.

Given these observations, it is difficult for me to discern what principle besides doctrinaire formalism actually motivates today’s decision. The majority chides the Apprendi dissenters for preferring a nuanced interpretation of the Due Process Clause and Sixth Amendment jury trial guarantee that would generally defer to legislative labels while acknowledging the existence of constitutional constraints— what the majority calls “the law must not go too far” approach. Ante, at 307 (emphasis deleted). If indeed the choice is between adopting a balanced case-by-case approach that takes into consideration the values underlying the Bill of Rights, as well as the history of a particular sentencing reform law, and adopting a rigid rule that destroys everything in its path, I will choose the former. See Apprendi, 530 U. S., at 552-554 (O’Connor, J., dissenting) (“Because I do not believe that the Court’s ‘increase in the maximum penalty’ rule is required by the Constitution, I would evaluate New Jersey’s sentence-enhancement statute by analyzing the factors we have examined in past cases” (citation omitted)).

But even were one to accept formalism as a principle worth vindicating for its own sake, it would not explain Ap-prendi% or today’s, result. A rule of deferring to legislative labels has no less formal pedigree. It would be more *322consistent with our decisions leading up to Apprendi, see Almendarez-Torres v. United States, 523 U. S. 224 (1998) (fact of prior conviction not an element of aggravated recidivist offense); United States v. Watts, 519 U. S. 148 (1997) (per curiam) (acquittal of offense no bar to consideration of underlying conduct for purposes of guidelines enhancement); Witte v. United States, 515 U. S. 389 (1995) (no double jeopardy bar against consideration of uncharged conduct in imposition of guidelines enhancement); Walton v. Arizona, 497 U. S. 639 (1990) (aggravating factors need not be found by a jury in capital case); Mistretta v. United States, 488 U. S. 361 (1989) (Federal Sentencing Guidelines do not violate separation of powers); McMillan v. Pennsylvania, 477 U. S. 79 (1986) (facts increasing mandatory minimum sentence are not necessarily elements); and it would vest primary authority for defining crimes in the political branches, where it belongs. Apprendi, supra, at 523-554 (O’Connor, J., dissenting). It also would be easier to administer than the majority’s rule, inasmuch as courts would not be forced to look behind statutes and regulations to determine whether a particular fact does or does not increase the penalty to which a defendant was exposed.

The majority is correct that rigid adherence to such an approach could conceivably produce absurd results, ante, at 306; but, as today’s decision demonstrates, rigid adherence to the majority’s approach does and will continue to produce results that disserve the very principles the majority purports to vindicate. The pre-Apprendi rule of deference to the legislature retains a built-in political check to prevent lawmakers from shifting the prosecution for crimes to the penalty phase proceedings of lesser included and easier-to-prove offenses — e.g., the majority’s hypothesized prosecution of murder in the guise of a traffic offense sentencing proceeding. Ante, at 306, There is no similar check, however, on application of the majority’s “any fact that increases the upper bound of judicial discretion” by courts.

*323The majority claims the mantle of history and original intent. But as I have explained elsewhere, a handful of state decisions in the mid-19th century and a criminal procedure treatise have little if any persuasive value as evidence of what the Framers of the Federal Constitution intended in the late 18th century. See Apprendi, 530 U. S., at 525-528 (O’Connor, J., dissenting). Because broad judicial sentencing discretion was foreign to the Framers, id., at 478-479 (citing J. Archbold, Pleading and Evidence in Criminal Cases 44 (15th ed. 1862)), they were never faced with the constitutional choice between submitting every fact that increases a sentence to the jury or vesting the sentencing judge with broad discretionary authority to account for differences in offenses and offenders.

IV

A

The consequences of today’s decision will be as far reaching as they are disturbing. Washington’s sentencing system is by no means unique. Numerous other States have enacted guidelines systems, as has the Federal Government. See, e. g., Alaska Stat. § 12.55.155 (2002); Ark. Code Ann. §16-90-804 (2003 Supp.); Fla. Stat. §921.0016 (2003); Kan. Stat. Ann. §21-4701 et seq. (2003); Mich. Comp. Laws Ann. §769.34 (West Supp. 2004); Minn. Stat. §244.10 (2002); N. C. Gen. Stat. §15A-1340.16 (Lexis 2003); Ore. Admin. Rule §213-008-0001 (2003); 204 Pa. Code § 303 et seq. (2004), reproduced following 42 Pa. Cons. Stat. Ann. §9721 (Purdon Supp. 2004); 18 U. S. C. §3553; 28 U. S. C. §991 et seq. Today’s decision casts constitutional doubt over them all and, in so doing, threatens an untold number of criminal judgments. Every sentence imposed under such guidelines in cases currently pending on direct appeal is in jeopardy. And, despite the fact that we hold in Schriro v. Summerlin, post, p. 348, that Ring (and a fortiori Apprendi) does not apply retroactively on habeas review, all criminal sentences imposed *324under the federal and state guidelines since Apprendi was decided in 2000 arguably remain open to collateral attack. See Teague v. Lane, 489 U. S. 288, 301 (1989) (plurality opinion) (“[Apease announces a new rule if the result was not dictated by precedent existing at the time the defendant's conviction became final”).2

The practical consequences for trial courts, starting today, will be equally unsettling: How are courts to mete out guidelines sentences? Do courts apply the guidelines as to mitigating factors, but not as to aggravating factors? Do they jettison the guidelines altogether? The Court ignores the havoc it is about to wreak on trial courts across the country.

B

It is no answer to say that today’s opinion impacts only Washington’s scheme and not others, such as, for example, the Federal Sentencing Guidelines. See ante, at 305, n. 9 (“The Federal Guidelines are not before us, and we express no opinion on them”); cf. Apprendi, supra, at 496-497 (claiming not to overrule Walton, supra, soon thereafter overruled in Ring); Apprendi, supra, at 497, n. 21 (reserving question of Federal Sentencing Guidelines). The fact that the Federal Sentencing Guidelines are promulgated by an administrative agency nominally located in the Judicial Branch is irrelevant to the majority’s reasoning. The Guidelines have the force of law, see Stinson v. United States, 508 U. S. 36 (1993); and Congress has unfettered control to reject or *325accept any particular guideline, Mistretta, 488 U. S., at 393-394.

The structure of the Federal Guidelines likewise does not, as the Government halfheartedly suggests, provide any grounds for distinction. Brief for United States as Amicus Curiae 27-29. Washington’s scheme is almost identical to the upward departure regime established by 18 U. S. C. § 3553(b) and implemented in USSG §5K2.0. If anything, the structural differences that do exist make the Federal Guidelines more vulnerable to attack. The provision struck down here provides for an increase in the upper bound of the presumptive sentencing range if the sentencing court finds, “considering the purpose of [the Act], that there are substantial and compelling reasons justifying an exceptional sentence.” Wash. Rev.Code Ann. §9.94A.120 (2000). The Act elsewhere provides a nonexhaustive list of aggravating factors that satisfy the definition. §9.94A.390. The Court flatly rejects respondent’s argument that such soft constraints, which still allow Washington judges to exercise a substantial amount of discretion, survive Apprendi. Ante, at 305. This suggests that the hard constraints found throughout chapters 2 and 3 of the Federal Sentencing Guidelines, which require an increase in the sentencing range upon specified factual findings, will meet the same fate. See, e.g., USSG §2K2.1 (increases in offense level for firearms offenses based on number of firearms involved, whether possession was in connection with another offense, whether the firearm was stolen); §2B1.1 (increase in offense level for financial crimes based on amount of money involved, number of victims, possession of weapon); §3C1.1 (general increase in offense level for obstruction of justice).

Indeed, the “extraordinary sentence” provision struck down today is as inoffensive to the holding of Apprendi as a regime of guided discretion could possibly be. The list of facts that justify an increase in the range is nonexhaustive. The State’s “real facts” doctrine precludes reliance by sen*326tencing courts upon facts that would constitute the elements of a different or aggravated offense. See Wash. Rev. Code Ann. § 9.94A.370(2) (2000) (codifying “real facts” doctrine). If the Washington scheme does not comport with the Constitution, it is hard to imagine a guidelines scheme that would.

* * *

What I have feared most has now come to pass: Over 20 years of sentencing reform are all but lost, and tens of thousands of criminal judgments are' in jeopardy. Ap-prendi, 530 U. S., at 549-559 (O’Connor, J., dissenting); Ring, 536 U. S., at 619-621 (O’Connor, J., dissenting). I respectfully dissent.

Justice Kennedy,

with whom Justice Breyer joins, dissenting.

The majority opinion does considerable damage to our laws and to the administration of the criminal justice system for all the reasons well stated in Justice O’Connor’s dissent, plus one more: The Court, in my respectful submission, disregards the fundamental principle under our constitutional system that different branches of government “converse with each other on matters of vital common interest.” Mistretta v. United States, 488 U. S. 361, 408 (1989). As the Court in Mistretta explained, the Constitution establishes a system of government that presupposes, not just “ ‘autonomy’ ” and “ ‘separateness,’ ” but also “ ‘interdependence’ ” and “ ‘reciprocity.’ ” Id., at 381 (quoting Youngstown Sheet & Tube Co. v. Sawyer, 343 U. S. 579, 635 (1952) (Jackson, J., concurring)). Constant, constructive discourse between our courts and our legislatures is an integral and admirable part of the constitutional design. Case-by-case judicial determinations often yield intelligible patterns that can be refined by legislatures and codified into statutes or rules as general standards. As these legislative enactments are followed by incremental judicial interpretation, the legis*327latures may respond again, and the cycle repeats. This recurring dialogue, an essential source for the elaboration and the evolution of the law, is basic constitutional theory in action.

Sentencing guidelines are a prime example of this collaborative process. Dissatisfied with the wide disparity in sentencing, participants in the criminal justice system, including judges, pressed for legislative reforms. In response, legislators drew from these participants’ shared experiences and enacted measures to correct the problems, which, as Justice O’Connor explains, could sometimes rise to the level of a constitutional injury. As Mistretta recognized, this interchange among different actors in the constitutional scheme is consistent with the Constitution’s structural protections.

To be sure, this case concerns the work of a state legislature, and not of Congress. If anything, however, this distinction counsels even greater judicial caution. Unlike Mistretta, the case here implicates not just the collective wisdom of legislators on the other side of the continuing dialogue over fair sentencing, but also the interest of the States to serve as laboratories for innovation and experiment. See New State Ice Co. v. Liebmann, 285 U. S. 262, 311 (1932) (Brandeis, J., dissenting). With no apparent sense of irony that the effect of today’s decision is the destruction of a sentencing scheme devised by democratically elected legislators, the majority shuts down alternative, nonjudicial, sources of ideas and experience. It does so under a faintly disguised distrust of judges and their purported usurpation of the jury’s function in criminal trials. It tells not only trial judges who have spent years studying the problem but also legislators who have devoted valuable time and resources “calling upon the accumulated wisdom and experience of the Judicial Branch ... on a matter uniquely within the ken of judges,” Mistretta, supra, at 412, that their efforts and judgments were all for naught. Numerous States that have enacted sentencing guidelines similar to the one in Washing*328ton State are now commanded to scrap everything and start over.

If the Constitution required this result, the majority’s decision, while unfortunate, would at least be understandable and defensible. As Justice O’Connor’s dissent demonstrates, however, this is simply not the case. For that reason, and because the Constitution does not prohibit the dynamic and fruitful dialogue between the judicial and legislative branches of government that has marked sentencing reform on both the state and the federal levels for more than 20 years, I dissent.

Justice Breyer,

with whom Justice O’Connor joins, dissenting.

The Court makes clear that it means what it said in Ap-prendi v. New Jersey, 530 U. S. 466 (2000). In its view, the Sixth Amendment says that “‘any fact that increases the penalty for a crime beyond the prescribed statutory maximum must be submitted to a jury.’” Ante, at 301 (quoting Apprendi, supra, at 490). “ ‘ [Prescribed statutory maximum’ ” means the penalty that the relevant statute authorizes “solely on the basis of the facts reflected in the jury verdict.” Ante, at 301, 303 (emphasis deleted). Thus, a jury must find, not only the facts that make up the crime of which the offender is charged, but also all (punishment-increasing) facts about the way in which the offender carried out that crime.

It is not difficult to understand the impulse that produced this holding. Imagine a classic example — a statute (or mandatory sentencing guideline) that provides a 10-year sentence for ordinary bank robbery, but a 15-year sentence for bank robbery committed with a gun. One might ask why it should matter for jury trial purposes whether the statute (or guideline) labels the gun’s presence (a) a sentencing fact about the way in which the offender carried out the lesser crime of ordinary bank robbery, or (b) a factual element of *329the greater crime of bank robbery with a gun? If the Sixth Amendment requires a jury finding about the gun in the latter circumstance, why should it not also require a jury to find the same fact in the former circumstance? The two sets of circumstances are functionally identical. In both instances, identical punishment follows from identical factual findings (related to, e. g., a bank, a taking, a thing-of-value, force or threat of force, and a gun). The only difference between the two circumstances concerns a legislative (or Sentencing Commission) decision about which label (“sentencing fact” or “element of a greater crime”) to affix to one of the facts, namely, the presence of the gun, that will lead to the greater sentence. Given the identity of circumstances apart from the label, the jury’s traditional factfinding role, and the law’s insistence upon treating like cases alike, why should the legislature’s labeling choice make an important Sixth Amendment difference?

The Court in Apprendi, and now here, concludes that it should not make a difference. The Sixth Amendment’s jury trial guarantee applies similarly to both. I agree with the majority’s analysis, but not with its conclusion. That is to say, I agree that, classically speaking, the difference between a traditional sentencing factor and an element of a greater offense often comes down to a legislative choice about which label to affix. But I cannot jump from there to the conclusion that the Sixth Amendment always requires identical treatment of the two scenarios. That jump is fraught with consequences that threaten the fairness of our traditional criminal justice system; it distorts historical sentencing or criminal trial practices; and it upsets settled law on which legislatures have relied in designing punishment systems.

The Justices who have dissented from Apprendi have written about many of these matters in other opinions. See 530 U. S., at 523-554 (O’Connor, J., dissenting); id., at 555-566 (Breyer, J., dissenting); Harris v. United States, 536 U. S. 545, 549-550, 556-569 (2002) (Kennedy, J.); id., at 569-572 *330(Breyer, J., concurring in part and concurring in judgment); Jones v. United States, 526 U. S. 227, 254, 264-272 (1999) (Kennedy, J., dissenting); Monge v. California, 524 U. S. 721, 728-729 (1998) (O’Connor, J.); McMillan v. Pennsylvania, 477 U. S. 79, 86-91 (1986) (Rehnquist, C. J.).' At the risk of some repetition, I shall set forth several of the most important considerations here. They lead me to conclude that I must again dissent.

I

The majority ignores the adverse consequences inherent in its conclusion. As a result of the majority’s rule, sentencing must now take one of three forms, each of which risks either impracticality, unfairness, or harm to the jury trial right the majority purports to strengthen. This circumstance shows that the majority’s Sixth Amendment interpretation cannot be right.

A

A first option for legislators is to create a simple, pure or nearly pure “charge offense” or “determinate” sentencing system. See Breyer, The Federal Sentencing Guidelines and the Key Compromises upon Which They Rest, 17 Hof-stra L. Rev. 1, 8-9 (1988). In such a system, an indictment would charge a few facts which, taken together, constitute a crime, such as robbery. Robbery would carry a single sentence, say, five years’ imprisonment. And every person convicted of robbery would receive that sentence — just as, centuries ago, everyone convicted of almost any serious crime was sentenced to death. See, e.g., Lillquist, The Puzzling Return of Jury Sentencing: Misgivings About Apprendi, 82 N. C. L. Rev. 621, 630 (2004).

Such a system assures uniformity, but at intolerable costs. First, simple determinate sentencing systems impose identical punishments on people who committed their crimes in very different ways. MTien dramatically different conduct *331ends up being punished the same way, an injustice has taken place. Simple determinate sentencing has the virtue of treating like cases alike, but it simultaneously fails to treat different cases differently. Some commentators have leveled this charge at sentencing guidelines systems themselves. See, e. g., Schulhofer, Assessing the Federal Sentencing Process: The Problem Is Uniformity, Not Disparity, 29 Am. Crim. L. Rev. 833, 847 (1992) (arguing that the “most important problem under the Guidelines system is not too much disparity, but rather excessive uniformity” and arguing for adjustments, including elimination of mandatory minimums, to make the Guidelines system more responsive to relevant differences). The charge is doubly applicable to simple “pure charge” systems that permit no departures from the prescribed sentences, even in extraordinary cases.

Second, in a world of statutorily fixed mandatory sentences for many crimes, determinate sentencing gives tremendous power to prosecutors to manipulate sentences through their choice of charges. Prosecutors can simply charge, or threaten to charge, defendants with crimes bearing higher mandatory sentences. Defendants, knowing that they will not have a chance to argue for a lower sentence in front of a judge, may plead to charges that they might otherwise contest. Considering that most criminal cases do not go to trial and resolution by plea bargaining is the norm, the rule of Apprendi, to the extent it results in a return to determinate sentencing, threatens serious unfairness. See Bibas, Judicial Fact-Finding and Sentence Enhancements in a World of Guilty Pleas, 110 Yale L. J. 1097, 1100-1101 (2001) (explaining that the rule of Apprendi hurts defendants by depriving them of sentencing hearings, “the only hearings they were likely to have”; forcing defendants to surrender sentencing issues like drug quantity when they agree to the plea; and transferring power to prosecutors).

*332B

A second option for legislators is to return to a system of indeterminate sentencing, such as California had before the recent sentencing reform movement. See Payne v. Tennessee, 501 U. S. 808, 820 (1991) (“With the increasing importance of probation, as opposed to imprisonment, as a part of the penological process, some States such as California developed the ‘indeterminate sentence/ where the time of incarceration was left almost entirely to the penological authorities rather than to the courts”); Thompson, Navigating the Hidden Obstacles to Ex-Offender Reentry, 45 Boston College L. Rev. 255, 267 (2004) (“In the late 1970s, California switched from an indeterminate criminal sentencing scheme to .determinate sentencing”). Under indeterminate systems, the length of the sentence is entirely or almost entirely within the discretion of the judge or of the parole board, which typically has broad power to decide when to release a prisoner.

When such systems were in vogue, they were criticized, and rightly so, for producing unfair disparities, including race-based disparities, in the punishment of similarly situated defendants. See, e. g., ante, at 315-316 (O’Connor, J., dissenting) (citing sources). The length of time a person spent in prison appeared to depend on “what the judge ate for breakfast” on the day of sentencing, on which judge you got, or on other factors that should not have made a difference to the length of the sentence. See Breyer, supra, at 4-5 (citing congressional and expert studies indicating that, before the United States Sentencing Commission Guidelines were promulgated, punishments for identical crimes in the Second Circuit ranged from 3 to 20 years’ imprisonment and that sentences varied depending upon region, gender of the defendant, and race of the defendant). And under such a system, the judge could vary the sentence greatly based upon his findings about how the defendant had committed the crime — findings that might not have been *333made by a “preponderance of the evidence,” much less “beyond a reasonable doubt.” See McMillan, 477 U. S., at 91 (“Sentencing courts have traditionally heard evidence and found facts without any prescribed burden of proof at all” (citing Williams v. New York, 337 U. S. 241 (1949))).

Returning to such a system would diminish the “ ‘reason’ ” the majority claims it is trying to uphold. Ante, at 302 (quoting 1 J. Bishop, Criminal Procedure §87, p. 55 (2d ed. 1872)). It also would do little to “ensur[e] [the] control” of what the majority calls “the peopl[e,]” i. e., the jury, “in the judiciary,” ante, at 306, since “the peopl[e]” would only decide the defendant’s guilt, a finding with no effect on the duration of the sentence. While “the judge’s authority to sentence” would formally derive from the jury’s verdict, the jury would exercise little or no control over the sentence itself. Ibid. It is difficult to see how such an outcome protects the structural safeguards the majority claims to be defending.

C

A third option is that which the Court seems to believe legislators will in fact take. That is the option of retaining structured schemes that attempt to punish similar conduct similarly and different conduct differently, but modifying them to conform to Apprendi's dictates. Judges would be able to depart downward from presumptive sentences upon finding that mitigating factors were present, but would not be able to depart upward unless the prosecutor charged the aggravating fact to a jury and proved it beyond a reasonable doubt. The majority argues, based on the single example of Kansas, that most legislatures will enact amendments along these lines in the face of the oncoming Apprendi train. See ante, at 309-310 (citing State v. Gould, 271 Kan. 394, 404-414, 23 P. 3d 801, 809-814 (2001); Act of May 29, 2002, ch. 170, 2002 Kan. Sess. Laws pp. 1018-1023 (codified at Kan. Stat. Ann. §21-4718 (2003 Cum. Supp.)); Brief for Kansas Appellate Defender Office as Amicus Curiae 3-7). It is therefore *334worth exploring how this option could work in practice, as well as the assumptions on which it depends.

1

This option can be implemented in one of two ways. The first way would be for legislatures to subdivide each crime into a list of complex crimes, each of which would be defined to include commonly found sentencing factors such as drug quantity, type of victim, presence of violence, degree of injury, use of gun, and so on. A legislature, for example, might enact a robbery statute, modeled on robbery sentencing guidelines, that increases punishment depending upon (1) the nature of the institution robbed, (2) the (a) presence of, (b) brandishing of, (c) other use of, a firearm, (3) making of a death threat, (4) presence of (a) ordinary, (b) serious, (c) permanent or life threatening, bodily injury, (5) abduction, (6) physical restraint, (7) taking of a firearm, (8) taking of drugs, (9) value of property loss, etc. Cf. United States Sentencing Commission, Guidelines Manual §2B3.1 (Nov. 2003) (hereinafter USSG).

This possibility is, of course, merely a highly calibrated form of the “pure charge” system discussed in Part I-A, supra. And it suffers from some of the same defects. The prosecutor, through control of the precise charge, controls the punishment, thereby marching the sentencing system directly away from, not toward, one important guideline goal: rough uniformity of punishment for those who engage in roughly the same real criminal conduct. The artificial (and consequently unfair) nature of the resulting sentence is aggravated by the fact that prosecutors must charge all relevant facts about the way. the crime was committed before a presentence investigation examines the criminal conduct, perhaps before the trial itself, i. e., before many of the facts relevant to punishment are known.

This “complex charge offense” system also prejudices defendants who seek trial, for it can put them in the untenable *335position of contesting material aggravating facts in the guilt phases of their trials. Consider a defendant who is charged, not with mere possession of cocaine, but with the specific offense of possession of more than 500 grams of cocaine. Or consider a defendant charged, not with murder, but with the new crime of murder using a machete. Or consider a defendant who the prosecution wants to claim was a “supervisor,” rather than an ordinary gang member. How can a Constitution that guarantees due process put these defendants, as a matter of course, in the position of arguing, “I did not sell drugs, and if I did, I did not sell more than 500 grams,” or “I did not kill him, and if I did, I did not use a machete,” or “I did not engage in gang activity, and certainly not as a supervisor” to a single jury? See Apprendi, 530 U. S., at 557-558 (Breyer, J., dissenting); Monge, 524 U. S., at 729. The system can tolerate this kind of problem up to a point (consider the defendant who wants to argue innocence, and, in the alternative, second-degree, not first-degree, murder). But a rereading of the many distinctions made in a typical robbery guideline, see supra, at 334, suggests that an effort to incorporate any real set of guidelines in a complex statute would reach well beyond that point.

The majority announces that there really is no problem here because “States may continue to offer judicial factfind-ing as a matter of course to all defendants who plead guilty” and defendants may “stipulate] to the relevant facts or consent] to judicial factfinding.” Ante, at 310. The problem, of course, concerns defendants who do not want to plead guilty to those elements that, until recently, were commonly thought of as sentencing factors. As to those defendants, the fairness problem arises because States may very well decide that they will not permit defendants to carve subsets of facts out of the new, Apprendi-required 17-element robbery crime, seeking a judicial determination as to some of those facts and a jury determination as to others. Instead, States may simply require defendants to plead guilty to all *33617 elements or proceed with a (likely prejudicial) trial on all 17 elements.

The majority does not deny that States may make this choice; it simply fails to understand why any State would want to exercise it. Ante, at 310, n. 12. The answer is, as I shall explain in a moment, that the alternative may prove too expensive and unwieldy for States to provide. States that offer defendants the option of judicial factfinding as to some facts (i. e., sentencing facts), say, because of fairness concerns, will also have to offer the defendant a second sentencing jury — just as Kansas has done. I therefore turn to that alternative.

2

The second way to make sentencing guidelines Apprendicompliant would be to require at least two juries for each defendant whenever aggravating facts are present: one jury to determine guilt of the crime charged, and an additional jury to try the disputed facts that, if found, would aggravate the sentence. Our experience with bifurcated trials in the capital punishment context suggests that reqmring them for run-of-the-mill sentences would be costly, both in money and in judicial time and resources. Cf. Kozinski & Gallagher, Death: The Ultimate Run-On Sentence, 46 Case W. Res. L. Rev. 1, 13-15, and n. 64 (1995) (estimating the costs of each capital case at around $1 million more than each noncapital case); Tabak, How Empirical Studies Can Affect Positively the Politics of the Death Penalty, 83 Cornell L. Rev. 1431, 1439-1440 (1998) (attributing the greater cost of death penalty cases in part to bifurcated proceedings). In the context of noncapital crimes, the potential need for a second indictment alleging aggravating facts, the likely need for formal evidentiary rules to prevent prejudice, and the increased difficulty of obtaining relevant sentencing information, all will mean greater complexity, added cost, and further delay. See Part V, infra. Indeed, cost and delay could lead legislatures *337to revert to the complex charge offense system described in Part I-C-l, supra.

The majority refers to an amicus curiae brief filed by the Kansas Appellate Defender Office, which suggests that a two-jury system has proved workable in Kansas. Ante, at 309-310. And that may be so. But in all likelihood, any such workability reflects an uncomfortable fact, a fact at which the majority hints, ante, at 310, but whose constitutional implications it does not seem to grasp. The uncomfortable fact that could make the system seem workable— even desirable in the minds of some, including defense attorneys — is called “plea bargaining.” See Bibas, 110 Yale L. J,, at 1150, and n. 330 (reporting that in 1996, fewer than 4% of adjudicated state felony defendants have jury trials, 5% have bench trials, and 91% plead guilty). See also ante, at 310 (making clear that plea bargaining applies). The Court can announce that the Constitution requires at least two jury trials for each criminal defendant — one for guilt, another for sentencing — but only because it knows full well that more than 90% of defendants will not go to trial even once, much less insist on two or more trials.

What will be the consequences of the Court’s holding for the 90% of defendants who do not go to trial? The truthful answer is that we do not know. Some defendants may receive bargaining advantages if the increased cost of the “double jury trial” guarantee makes prosecutors more willing to cede certain sentencing issues to the defense. Other defendants may be hurt if a “single-jury-decides-all” approach makes them more reluctant to risk a trial — perhaps because they want to argue that they did not know what was in the cocaine bag, that it was a small amount regardless, that they were unaware a confederate had a gun, etc. See Bibas, 110 Yale L. J., at 1100 (“Because for many defendants going to trial is not a desirable option, they are left without any real hearings at all”); id., at 1151 (“The trial right does little good when most defendants do not go to trial”).

*338At the least, the greater expense attached to trials and their greater complexity, taken together in the context of an overworked criminal justice system, will likely mean, other things being equal, fewer trials and a greater reliance upon plea bargaining — a system in which punishment is set not by judges or juries but by advocates acting under bargaining constraints. At the same time, the greater power of the prosecutor to control the punishment through the charge would likely weaken the relation between real conduct and real punishment as well. See, e. g., Schulhofer, 29 Am. Crim. L. Rev., at 845 (estimating that evasion of the proper sentence under the Federal Guidelines may now occur in 20%-35% of all guilty plea cases). Even if the Court’s holding does not further embed plea-bargaining practices (as I fear it will), its success depends upon the existence of present practice. I do not understand how the Sixth Amendment could require a sentencing system that will work in practice only if no more than a handful of defendants exercise their right to a jury trial.

The majority’s only response is to state that “bargaining over elements . . . probably favors the defendant,” ante, at 311, adding that many criminal defense lawyers favor its position, ante, at 312. But the basic problem is not one of “fairness” to defendants or, for that matter, “fairness” to prosecutors. Rather, it concerns the greater fairness of a sentencing system that a more uniform correspondence between real criminal conduct and real punishment helps to create. At a minimum, a two-jury system, by preventing a judge from taking account of an aggravating fact without the prosecutor’s acquiescence, would undercut, if not nullify, legislative efforts to ensure through guidelines that punishments reflect a convicted offender’s real criminal conduct, rather than that portion of the offender’s conduct that a prosecutor decides to charge and prove.

Efforts to tie real punishment to real conduct are not new. They are embodied in well-established preguidelines sen-*339tenting practices — practices under which a judge, looking at a presentence report, would seek to tailor the sentence in significant part to fit the criminal conduct in which the offender actually engaged. For more than a century, questions of punishment (not those of guilt or innocence) have reflected determinations made, not only by juries, but also by judges, probation officers, and executive parole boards. Such truthseeking determinations have rested upon both adversarial and nonadversarial processes. The Court’s holding undermines efforts to reform these processes, for it means that legislatures cannot both permit judges to base sentencing upon real conduct and seek, through guidelines, to make the results more uniform.

In these and other ways, the two-jury system would work a radical change in pre-existing criminal law. It is not surprising that this Court has never previously suggested that the Constitution — outside the unique context of the death penalty — might require bifurcated jury-based sentencing. And it is the impediment the Court’s holding poses to legislative efforts to achieve that greater systematic fairness that casts doubt on its constitutional validity.

D

Is there a fourth option? Perhaps. Congress and state legislatures might, for example, rewrite their criminal codes, attaching astronomically high sentences to each crime, followed by long lists of mitigating facts, which, for the most part, would consist of the absence of aggravating facts. Ap-prendi, 530 U. S., at 541-542 (O’Connor, J., dissenting) (explaining how legislatures can evade the majority’s rule by making yet another labeling choice). But political impediments to legislative action make such rewrites difficult to achieve; and it is difficult to see why the Sixth Amendment would require legislatures to undertake them.

It may also prove possible to find combinations of, or variations upon, my first three options. But I am unaware of any *340variation that does not involve (a) the shift of power to the prosecutor (weakening the connection between real conduct and real punishment) inherent in any charge offense system, (b) the lack of uniformity inherent in any system of pure judicial discretion, or (c) the complexity, expense, and increased reliance on plea bargains involved in a “two-jury” system. The simple fact is that the design of any fair sentencing system must involve efforts to make practical compromises among competing goals. The majority’s reading of the Sixth Amendment makes the effort to find those compromises — already difficult — virtually impossible.

II

The majority rests its conclusion in significant part upon a claimed historical (and therefore constitutional) imperative. According to the majority, the rule it applies in this case is rooted in “longstanding tenets of common-law criminal jurisprudence,” ante, at 301: that every accusation against a defendant must be proved to a jury and that “ ‘an accusation which lacks any particular fact which the law makes essential to the punishment is... no accusation within the requirements of the common law, and it is no accusation in reason,’ ” ante, at 301-302 (quoting Bishop, Criminal Procedure §87, at 55). The historical sources upon which the majority relies, however, do not compel the result it reaches. See ante, at 323 (O’Connor, J., dissenting); Apprendi, 530 U. S., at 525-528 (O’Connor, J., dissenting). The quotation from Bishop, to which the majority attributes great weight, stands for nothing more than the “unremarkable proposition” that where a legislature passes a statute setting forth heavier penalties than were available for committing a common-law offense and specifying those facts that triggered the statutory penalty, “a defendant could receive the greater statutory punishment only, if the indictment expressly charged and the prosecutor proved the facts that made up the statutory offense, as opposed to simply those facts that made up *341the common-law offense.” Id., at 526 (O’Connor, J., dissenting) (characterizing a similar statement of the law in J. Arch-bold, Pleading and Evidence in Criminal Cases 51, 188 (15th ed. 1862)).

This is obvious when one considers the problem that Bishop was addressing. He provides as an example “statutes whereby, when [a common-law crime] is committed with a particular intent, or with a particular weapon, or the like, it is subjected to a particular corresponding punishment, heavier than that for” the simple common-law offense (though, of course, his concerns were not “limited to that example,” ante, at 302, n. 5). Bishop, supra, §82, at 51-52 (discussing the example of common assault and enhanced-assault statutes, e. g., “assaults committed with the intent to rob”). That indictments historically had to charge all of the statutorily labeled elements of the offense is a proposition on which all can agree. See Apprendi, supra, at 526-527 (O’Connor, J., dissenting). See also J. Archbold, Pleading and Evidence in Criminal Cases 44 (11th ed. 1849) (“[E]very fact or circumstance which is a necessary ingredient in the offence must be set forth in the indictment” so that “there may be no doubt as to the judgment which should be given, if the defendant be convicted”); 1 T. Starkie, Criminal Pleading 68 (2d ed. 1822) (the indictment must state “the criminal nature and degree of the offence, which are conclusions of law from the facts; and also the particular facts and circumstances which render the defendant guilty of that offence”).

Neither Bishop nor any other historical treatise writer, however, disputes the proposition that judges historically had discretion to vary the sentence, within the range provided by the statute, based on facts not proved at the trial. See Bishop, supra, §85, at 54 (“[W]ithin the limits of any discretion as to the punishment which the law may have allowed, the judge, when he pronounces sentence, may suffer his discretion to be influenced by matter shown in aggravation or mitigation, not covered by the allegations of the in*342dictment”); K. Stith & J. Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts 9 (1998). The modern history of preguidelines sentencing likewise indicates that judges had broad discretion to set sentences within a statutory range based on uncharged conduct. Usually, the judge based his or her sentencing decision on facts gleaned from a presentence report, which the defendant could dispute at a sentencing hearing. In the federal system, for example, Federal Rule of Criminal Procedure 32 provided that probation officers, who are employees of the Judicial Branch, prepared a presentence report for the judge, a copy of which was generally given to the prosecution and defense before the sentencing hearing. See Stith & Cabranes, supra, at 79-80, 221, n. 5. See also ante, at 315 (O’Connor, J., dissenting) (describing the State of Washington’s former indeterminate sentencing law).

In this case, the statute provides that kidnaping may be punished by up to 10 years’ imprisonment. Wash. Rev. Code Ann. §§ 9A.40.030(3), 9A.20.021(1)(b) (2000). Modern structured sentencing schemes like Washington’s do not change the statutorily fixed maximum penalty, nor do they purport to establish new elements for the crime. Instead, they undertake to structure the previously unfettered discretion of the sentencing judge, channeling and limiting his or her discretion even within the statutory range. (Thus, contrary to the majority’s arguments, ante, at 308-309, kidnapers in the State of Washington know that they risk up to 10 years’ imprisonment, but they also have the benefit of additional information about how long — within the 10-year maximum — their sentences are likely to be, based on how the kidnaping was committed.)

Historical treatises do not speak to such a practice because it was not done in the 19th century. Cf. Jones, 526 U. S., at 244 C‘[T]he scholarship of which we are aware does not show that a question exactly like this one was ever raised and resolved in the period before the framing”). This makes *343sense when one considers that, prior to the 19th century, the prescribed penalty for felonies was often death, which the judge had limited, and sometimes no, power to vary. See Lillquist, 82 N. C. L. Rev., at 628-630. The 19th century saw a movement to a rehabilitative mode of punishment in which prison terms became a norm, shifting power to the judge to impose a longer or shorter term within the statutory maximum. See ibid. The ability of legislatures to guide the judge’s discretion by designating presumptive ranges, while allowing the judge to impose a more or less severe penalty in unusual cases, was therefore never considered. To argue otherwise, the majority must ignore the significant differences between modern structured sentencing schemes and the history on which it relies to strike them down. And while the majority insists that the historical sources, particularly Bishop, should not be “limited” to the context in which they were written, ante, at 302, n. 5, it has never explained why the Court must transplant those discussions to the very different context of sentencing schemes designed to structure judges’ discretion within a statutory sentencing range.

Given history’s silence on the question of laws that structure a judge’s discretion within the range provided by the legislatively labeled maximum term, it is not surprising that oiir modern, pre-Apprendi cases made clear that legislatures could, within broad limits, distinguish between “sentencing facts” and “elements of crimes.” See McMillan, 477 U. S., at 85-88. By their choice of label, legislatures could indicate whether a judge or a jury must make the relevant factual determination. History does not preclude legislatures from making this decision. And, as I argued in Part I, supra, allowing legislatures to structure sentencing in this way has the dual effect of enhancing and giving meaning to the Sixth Amendment’s jury trial right as to core crimes, while affording additional due process to defendants in the form of sen*344tencing hearings before judges — hearings the majority's rule will eliminate for many.

Is there a risk of unfairness involved in permitting Congress to make this labeling decision? Of course. As we have recognized, the “tail” of the sentencing fact might “wa[g] the dog of the substantive offense.” McMillan, supra, at 88. Congress might permit a judge to sentence an individual for murder though convicted only of making an illegal lane change. See ante, at 306 (majority opinion). But that is the kind of problem that the Due Process Clause is well suited to cure. McMillan foresaw the possibility that judges would have to use their own judgment in dealing with such a problem; but that is what judges are there for. And, as Part I, supra, makes clear, the alternatives are worse — not only practically, but, although the majority refuses to admit it, constitutionally as well.

Historic practice, then, does not compel the result the majority reaches. And constitutional concerns counsel the opposite.

Ill

The majority also overlooks important institutional considerations. Congress and the States relied upon what they believed was their constitutional power to decide, within broad limits, whether to make a particular fact (a) a sentencing factor or (b) an element in a greater crime. They relied upon McMillan as guaranteeing the constitutional validity of that proposition. They created sentencing reform, an effort to change the criminal justice system so that it reflects systematically not simply upon guilt or innocence but also upon what should be done about this now-guilty offender. Those efforts have spanned a generation. They have led to state sentencing guidelines and the Federal Sentencing Guidelines system. E.g., ante, at 314-318 (O’Connor, J., dissenting) (describing sentencing reform in the State of Washington). These systems are imperfect and they yield far from perfect results, but I cannot believe the Constitu*345tion forbids the state legislatures and Congress to adopt such systems and to try to improve them over time. Nor can I believe that the Constitution hamstrings legislatures in the way that Justice O’Connor and I have discussed.

IV

Now, let us return to the question I posed at the outset. Why does the Sixth Amendment permit a jury trial right (in respect to a particular fact) to depend upon a legislative labeling decision, namely, the legislative decision to label the fact a sentencing fact, instead of an element of the crime? The answer is that the fairness and effectiveness of a sentencing system, and the related fairness and effectiveness of the criminal justice system itself, depend upon the legislature’s possessing the constitutional authority (within due process limits) to make that labeling decision. To restrict radically the legislature’s power in this respect, as the majority interprets the Sixth Amendment to do, prevents the legislature from seeking sentencing systems that are consistent with, and indeed may help to advance, the Constitution’s greater fairness goals.

To say this is not simply to express concerns about fairness to defendants. It is also to express concerns about the serious practical (or impractical) changes that the Court’s decision seems likely to impose upon the criminal process; about the tendency of the Court’s decision to embed further plea bargaining processes that lack transparency and too often mean nonuniform, sometimes arbitrary, sentencing practices; about the obstacles the Court’s decision poses to legislative efforts to bring about greater uniformity between real criminal conduct and real punishment; and ultimately about the limitations that the Court imposes upon legislatures’ ability to make democratic legislative decisions. Whatever the faults of guidelines systems — and there are many — they are more likely to find their cure in legislation emerging from the experience of, and discussion among, all elements of the *346criminal justice community, than in a virtually unchangeable constitutional decision of this Court.

V

Taken together these three sets of considerations, concerning consequences, concerning history, concerning institutional reliance, leave me where I was in Apprendi, i. e., convinced that the Court is wrong. Until now, I would have thought the Court might have limited Apprendi so that its underlying principle would not undo sentencing reform efforts. Today's case dispels that illusion. At a minimum, the case sets aside numerous state efforts in that direction. Perhaps the Court will distinguish the Federal Sentencing Guidelines, but I am uncertain how. As a result of today's decision, federal prosecutors, like state prosecutors, must decide what to do next, how to handle tomorrow’s case.

Consider some of the matters that federal prosecutors must know about, or guess about, when they prosecute their next case: (1) Does today’s decision apply in full force to the Federal Sentencing Guidelines? (2) If so, must the initial indictment contain all sentencing factors, charged as “elements” of the crime? (3) What, then, are the evidentiary rules? Can the prosecution continue to use, say, presentence reports, with their conclusions reflecting layers of hearsay? Cf. Crawford v. Washington, 541 U. S. 36, 63, 68 (2004) (clarifying the Sixth Amendment’s requirement of confrontation with respect to testimonial hearsay). Are the numerous cases of this Court holding that a sentencing judge may consider virtually any reliable information still good law when juries, not judges, are required to determine the matter? See, e. g., United States v. Watts, 519 U. S. 148, 153-157 (1997) (per curiam) (evidence of conduct of which the defendant has been acquitted may be considered at sentencing). Cf. Witte v. United States, 515 U. S. 389, 399-401 (1995) (evidence of uncharged criminal conduct used in determining sentence). (4) How are juries to deal with highly complex *347or open-ended Sentencing Guidelines obviously written for application by an experienced trial judge? See, e. g., USSG §3B1.1 (requiring a greater sentence when the defendant was a leader of a criminal activity that involved four or more participants or was “otherwise extensive” (emphasis added)); §§3D1.1-3D1.2 (highly complex “multiple count” rules); § IB 1.3 (relevant conduct rules).

Ordinarily, this Court simply waits for cases to arise in which it can answer such questions. But this case affects tens of thousands of criminal prosecutions, including federal prosecutions. Federal prosecutors will proceed with those, prosecutions subject to the risk that all defendants in those cases will have to be sentenced, perhaps tried, anew. Given this consequence and the need for certainty, I would not proceed further piecemeal; rather, I would call for further argument on the ramifications of the concerns I have raised. But that is not the Court’s view.

For the reasons given, I dissent.

LEVENTHAL, Circuit Judge:

Seven of the so-called “D.C. Nine” bring this joint appeal from convictions arising out of their unconsented entry into the Washington offices of the Dow Chemical Company, and their destruction *1117of certain property therein. Appellants,1 along with two other defendants who subsequently entered pleas of nolo contendere,2 were tried before District Judge John H. Pratt and a jury on a three count indictment alleging, as to each defendant, one count of second degree burglary, 22 D.C.Code § 1801(b), and two counts of malicious destruction of property valued in excess of $100, 22 D.C.Code § 403. On February 11, 1970, after a six-day trial, the seven were each convicted of two counts of malicious destruction. The jury acquitted on the burglary charges but convicted on the lesser-included offense of unlawful entry. The sentences imposed are set forth in the margin.3

Appellants urge three grounds for reversal as follows: (1) The trial judge erred in denying defendants’ timely motions to dispense with counsel and represent themselves. (2) The judge erroneously refused to instruct the jury of its right to acquit appellants without regard to the law and the evidence, and refused to permit appellants to argue that issue to the jury. (3) The instructions actually given by the court coerced the jury into delivering a verdict of guilty. On the basis of defendants’ first contention we reverse and remand for new trial. To provide an appropriate mandate governing the new trial, we consider the second and third contentions, and conclude that these cannot be accepted.

I. The Record in District Court

The undisputed evidence showed that on Saturday, March 22, 1969, appellants broke into the locked fourth floor Dow offices at 1030 - 15th Street, N.W., Washington, D.C., threw papers and documents about the office and into the street below, vandalized office furniture and equipment, and defaced the premises by spilling about a bloodlike substance. The prosecution proved its case through Dow employees who testified as to the lack of permission and extent of damage, members of the news media who had been summoned to the scene by the appellants and who witnessed the destruction while recording it photographically, and police officers who arrested appellants on the scene.

Initially, the court appointed separate counsel for each defendant. Following their arraignment on June 20, 1969, all save appellant Robert Begin elected interim joint representation by Philip Hirschkop, Addison Bowman and Caroline Nickerson. Mr. Begin was represented by Edward Bennett Williams. All attorneys were court-appointed.

At the pre-trial conference held on January 29, 1970, in his chambers, Judge Pratt indicated he had received a letter from Mr. Hirschkop to the effect that appellants JoAnn Malone, Arthur Melville, and Joseph O’Rourke no longer wished to be represented by counsel. On his own behalf, appellant Begin had written the Court requesting that the appointment of Mr. Williams be terminated, and that he be permitted a pro se defense. Judge Pratt deferred a ruling on the pro se motions in order to give the matter further consideration, observing that to waive counsel

is not quite as easy as merely getting up and saying that you want to represent yourself. You’ve got the matter of the waiver being knowing and intelligent, and we are going to take testimony on that; and furthermore, the possibility of prejudice, not only *1118to themselves but also to their co-defendants. (Tr. 3)

Later in the conference he indicated how important he felt the lawyer’s role was likely to be in achieving tranquility at trial:

[L]et me emphasize as strongly as I can that the way this ease is handled is presumably my responsibility, but the decorum in the courtroom — I’m talking particularly about the defendants themselves — will be affected to a great degree by the advice and example that they get from their own lawyers. (Tr. 14).

The day before, it seems, Judge Pratt had attended a seminar on the problems of disruption encountered in multi-de-fendant trials, and he was concerned that “there have been rumors that maybe some disruptive tactics are going to be employed.” (Tr. 15, 174). Defense counsel assured him that they knew of no such rumors and that they anticipated no disruptive behavior. In any event, Judge Pratt scheduled a hearing for February 3, 1970, on the four defendants’ requests as a matter preliminary to the trial.

At a “supplementary” conference the next day, January 30, the Judge acceded to Mr. Williams’s request that his associate, William McDaniels, be substituted as Begin’s counsel of record. Mr. Williams’s request stemmed from statements of emphatic, indeed vitriolic, dissatisfaction with Mr. Williams’s representation contained in Begin’s motion for a pro se defense.

At the February 3, 1970, hearing on the pro se motions, the four original movants were joined by appellant Dougherty. For approximately three-quarters of an hour the court heard from the five defendants and from their lawyers. The judge showed particular interest in the defendants’ education, and specifically whether any of the five had had formal legal training. None had, although appellant Begin asserted he had “taken pains to familiarize (himself) with courtroom procedure.” In general, the testimony showed that the five mov-ants were quite articulate and highly educated. It also appeared that all five movants — indeed all defendants save appellant Slaski — were associated with a religious order, either at that time or in the recent past and, in varying degrees, had been active in work among the poor and underprivileged, in this country and in Latin America.

After a brief recess, the court denied defendants’ motions in an oral opinion, set forth in the margin.5 The judge in*1119dicated that he was not troubled by defendants’ general educational background, nor, importantly, by their motivation. However, he emphasized their lack of formal legal training, the multi-defendant context of the trial, and the seriousness of the charges. The interplay of those factors he felt created too great a risk of disruption of the trial, and risk of jury prejudice against mov-ants and their co-defendants. After the judge delivered his ruling, there was some confused interchange between the court, counsel and several defendants. At Mr. Hirschkop’s request, court was recessed in order that counsel and defendants could determine how to proceed in light of the judge’s ruling.

When court reconvened after lunch, all defendants were present and a panel of prospective jurors sat in the rear of the courtroom. At that time the remaining four defendants, Catherine Melville, Dennis Moloney, Michael Slaski and Bernard Meyer, made oral motions to represent themselves. These new motions were prompted in large part by the judge’s earlier emphasis on prejudice to co-defendants with counsel as a ground for denying the original motions. After some discussion between the court and the defendants, out of the presence of the veniremen, the judge, treating the four new motions as timely made, formally denied them for the reasons given in his oral opinion.

Some collateral matters were then disposed of, including a request — denied— by Mr. Hirschkop that Judge Pratt disqualify himself. When the court began its voir dire examination of prospective jurors, there was some dispute about the judge’s decision to question the jurors himself rather than to follow the procedure of examination by counsel, but appellants do not bring that issue into this appeal. Similarly, appellants do not complain of the presence or absence of certain questions for the prospective jurors, about which there was controversy at the time.

Selection of the jury was reasonably rapid, requiring only a part of the afternoon of Feb. 3 and part of the next morning. As noted, the court conducted the voir dire. Defendants were, however, permitted to exercise their peremptory challenges in propria persona.

The trial formally began on the afternoon of February 4. Judge Pratt required that motions, objections, and examination of witnesses be made through counsel. He did, however, agree to permit each defendant to make a five minute opening statement and to testify, in narrative form, at reasonable length without a specific time limit. After the prosecution’s opening statement, Mr. Hirschkop made a brief statement on behalf of all defendants. Then five of the defendants — including appellants *1120Dougherty, Begin and Arthur Melville— made opening statements.

The prosecution’s case was completed by the end of the next day, February 5. Prior to the opening of the defense case, defendants Catherine Melville and Bernard Meyer entered pleas of nolo con-tendere to one count of malicious destruction of property; the remaining charges were dismissed; and they are not before us on this appeal.

On Friday, February 6, after an opening statement by Mr. Bowman, appellants O’Rourke and Malone made opening statements on their own behalf, as the other defendants had done prior to the Government’s case. They directed their remarks, as had the others, to an attack on the role of Dow Chemical Company and other unspecified corporations in supporting American military efforts in the Vietnam War. When Sister Malone referred to Vietnam, Judge Pratt interjected : “the war in Vietnam is not an issue in this case.” A disruption ensued. Events happened too quickly for the court reporter to provide a complete record. The court later inserted this description of what happened, Supplement to Transcript, p. 595:

The record being unclear as to what transpired in the courtroom shortly before the Court adjourned Friday, February 6, 1970, the following is a recital of those events.

Defendant Jo Ann Malone, while making her opening statement, referred to the Vietnam War. (T. 594) The Court ruled that “the War in Vietnam is not an issue in this case.” (T. 594) Defendant Arthur Melville rose to object and was ordered by the Court to be seated. Defendant Michael Sla-ski also objected and when he failed to obey the Court’s order to be seated, the Marshals were ordered to seat him. (T. 595) While this was taking place, two spectators in the rear of the courtroom then stood and shouted to the bench concerning the relevancy of the War in the case on trial. Marshals moved to eject these two persons. The first was removed without incident. While the second was being ejected with some difficulty, a woman member of the DC-Nine Defense Committee seated in the front row in back of the defendants rose and ran to the back of the courtroom to impede the Marshals and assist the two spectators being removed. When the Marshals resisted her, she screamed at them. Defendant Michael Slaski then wrestled free from the Marshals who were attempting to seat him, hurdled the rail and engaged in an altercation with the Marshals at the rear of the courtroom. During these events the jury was ushered from the courtroom. The Court ordered the courtroom cleared and took a recess. It is reported that the fighting involving defendant Slaski ceased after two or three minutes and the Marshals began clearing the courtroom amid shouts of “pigs” and obscenities. Loud shouting occurred during the entire incident. A number of spectators refused to leave the courtroom and had to be ejected forcibly. The Court returned after the courtroom had been cleared and the press, counsel and the defendants had been readmitted. The jury was recalled, admonished to disregard what it had seen, and sent home. The Court then adjourned until Monday, February 9, 1970.

When the trial resumed on Monday, February 9, defendant Slaski was cited for contempt for his role in the disturbances and the judge sternly admonished the spectators and remaining defendants against further outbreaks. Appellants Slaski and Moloney did not make opening statements. After appellant Malone finished her statement, the case for the defense began. It consisted entirely of defendants’ testimony. Appellants Arthur. Melville, O’Rourke, Malone and Begin testified. During the testimony there were several further disruptions requiring a brief recess at one point and ejection of a spectator from the court*1121room at another.6 The judge confined closing argument to counsel. He instructed the jury on the three counts of each indictment as well as on the lesser-included offense of unlawful entry under the burglary count. He refused to instruct the jury that it could disregard the law as he gave it to them, and refused to instruct the jury that “moral compulsion” or “choice of the lesser evil” constituted a legal defense.

II. The Right of Pro Se Representation

In defendants’ view, Judge Pratt violated their constitutional and statutory rights when he refused to permit them to represent themselves. They say the right to dispense with counsel is correlative to the guarantee of the right to counsel and is therefore “implicit” in the Sixth Amendment. They rely as well on 28 U.S.C. § 1654 which provides:

In all courts of the United States the parties may plead and conduct their own cases personally or by counsel as, by the rules of such courts, respectively, are permitted to manage and conduct causes therein.

The Government responds that a defendant’s right to represent himself is not protected by the Sixth Amendment but only exists by virtue of § 1654, and that this is significant (a) because statutory rights are generally subject to the “harmless error” principle, and (b) § 1654 rights can be limited when, in the judge’s view, they would be likely to lead to disruption of the trial or threaten to interfere with effective presentation of the defendant’s case.

A. Absence of Controlling Precedent on Source of Pro Se Right

The Supreme Court has never directly determined whether the Constitution guarantees the pro se right. Appellants rely on Adams v. United States ex rel. McCann, 317 U.S. 269, 279, 63 S.Ct. 236, 242, 87 L.Ed. 268 (1942) where the Court observed:

[t]he right to assistance of counsel and the correlative right to dispense with a lawyer’s help are not legal formalisms. They rest on considerations that go to the substance of an accused’s position before the law. . . [T]he Constitution does not force a lawyer upon a defendant. He may waive his Constitutional right to assistance of counsel if he knows what he is doing and his choice is made with eyes open.

However, Adams’s “correlative right” language was not an essential ingredient of its holding, which was simply that a defendant who has intelligently waived his right to counsel may also waive his right to a jury trial. Moreover, in Singer v. United States, 380 U.S. 24, 85 S.Ct. 783, 13 L.Ed.2d 630 (1965), holding a defendant’s waiver of right to a jury trial subject to the assent of the prosecution and the trial judge, the Court said that “the ability to waive a constitutional right does not ordinarily carry with it the right to insist on the opposite of that right.” 380 U.S. at 34-35, 85 S.Ct. at 789. It pointed out that

[t]he Constitution recognizes an adversary system as the proper method of determining guilt, and the Government, as a litigant, has a legitimate interest in seeing that cases in which it believes a conviction is warranted are tried before the tribunal which the Constitution regards as most likely to produce a fair result. Id. at 36, 85 S.Ct. at 790.7

*1122There are conflicting indications from the circuits. The Second Circuit in United States v. Plattner, 330 F.2d 271 (2d Cir. 1964), recognized constitutional status for the pro se right. In accord with Plattner, see Lowe v. United States, 418 F.2d 100 (7th Cir. 1969), cert. denied 397 U.S. 1048, 90 S.Ct. 1378, 25 L.Ed.2d 660 (1970); United States v. Warner, 428 F.2d 730 (8th Cir. 1970), cert. denied 400 U.S. 930, 91 S.Ct. 194, 27 L.Ed.2d 191 (1971); United States v. Pike, 439 F.2d 695 (9th Cir. 1971); Hodge v. United States, 414 F.2d 1040 (9th Cir. 1969). But compare Juelich v. United States, 342 F.2d 29 (5th Cir. 1965); Van Nattan v. United States, 357 F.2d 161 (10th Cir. 1966).

In our court, Brown v. United States, 105 U.S.App.D.C. 77, 264 F.2d 363 (en banc 1959), left the issue unresolved. Reversal was sought because the trial judge failed to instruct a defendant of his right to represent himself when his court-appointed attorney informed the court that defendant was dissatisfied with the attorney’s pessimism about the outcome of the case. There was no opinion for a majority of the court. Judge Miller, joined by Judges Prettyman, Danaher and Bastían, stated that the pro se right is statutory only, and therefore (a) defendant must assert the right in order to be entitled to it and (b) in any event no reversal was required since no prejudice could be discerned.

The opinions of the other five judges are consistent with a view that the Constitution is the basis for the right, although only one expressly discussed its source. Judge Burger, concurring in part, differed with the others voting to affirm because he treated defendant’s expression of “dissatisfaction” as the equivalent of a request for removal of counsel, but held it within the court’s discretion to deny the request so long as it was sufficiently apprised of the cause of the “dissatisfaction.”

The four dissenters joined in an opinion stating that the reasons for the “dissatisfaction” were not made sufficiently clear to the court to permit the exercise of its informed discretion on the matter, and hence the judge should have inquired further into the basis for defendant’s attitude and should have specifically informed defendant of his right to proceed alone. The pro se alternative, they felt, is one of fundamental importance, and a defendant can make an informed decision on how best to conduct his defense only if he is made aware that he is free to dispense with counsel. The principal dissent, however, made no mention of the Constitution. Only Judge Bazelon, in a separate dissent, said the pro se right was grounded in the Constitution.

B. Need for Recognition of Statutory Right — If Timely Asserted, Not Waived, And Accompanied by Waiver of Right to Counsel

The Government says the pro se right is statutory and subject to “extensive qualifications,” discerning in the decisions seven “factors” on the basis of which the pro se right may be partially or entirely denied.8

This case does not require final resolution of the constitutional question. That would be unavoidable had Congress attempted to narrow or qualify the pro se right along the lines advocated by the Government on this appeal. But that is not the case. The right of pro se representation was enacted by our very first Congress. The language declaring the pro se right is not qualified, see 28 U.S.C. § 1654. The statute was passed in a context of colonial tribunals largely manned by laymen, and of pioneer modes of thought emphasizing the *1123virtues of common sense and self-reli-anee.9 Its constitutional aura is underscored by the proposal the very next day of the Sixth Amendment.10

In sum, whether or not the right of pro se representation has a constitutional foundation it is patently a statutory right, see § 1654; this right was not only conferred by Congress in 1789 but. has wide reverberation in organic state law11 and was recognized by Congress as a fundamental right. We conclude that this right must be recognized if it is timely asserted, and accompanied by a valid waiver of counsel, and if it is not itself waived, either expressly, or constructively, as by disruptive behavior during trial.12

The precedents relied on by the Government as subjecting the pro se right to “extensive qualifications” do no more than establish these basic elements : timely assertion; need for intelligent waiver of counsel; and possibility of waiver of the pro se right. A number of cases involved the special circumstance of defendants whose mental capacity was impaired.13 The bulk of *1124the cases cited to us involved requests made after commencement of trial,14 and do no more than apply the recognized principle that the fundamental right to conduct the case pro se is one that must be claimed timely, before the trial begins. Just as a defendant who has unrestricted right to retain counsel of his own choosing must seek permission of the court once his choice has been made, to select a different retained counsel, and is subject to the sound discretion of the court when he seeks to make a change after his trial has commenced, so a defendant must obtain the court’s permission when he seeks to make a change in order to select himself as counsel.

When the pro se right is claimed after trial has begun, the court exercises its discretion. It may weigh the inconvenience threatened by defendant’s belated request against the possible prejudice from denial of defendant’s request. In exercising discretion the judge may take into account the circumstances at the time, whether there has been prior disruptive behavior by defendant, whether the trial is in an advanced stage, etc. E. g., Seale v. Hoffman, supra, note 14; United States v. Foster, 9 F.R.D. 367 (S.D.N.Y.1949). The right to self-representation, though asserted before trial, can be lost by disruptive behavior during trial, constituting constructive waiver. But that is a far different situation from that presented by the instant case, where appellants unequivocally claimed the right to represent themselves, see Brown v. United States, supra, 105 U.S. App.D.C. at 81, 264 F.2d at 367 (Burger, J.), United States ex rel. Higgins v. Fay, 364 F.2d 219 (2d Cir. 1966), well in advance of the beginning of trial and selection of the jury, see United States ex rel. Maldonado v. Denno, 348 F.2d 12 (2d Cir. 1965); cf. United States v. Thomas, 146 U.S.App.D.C. 308, 450 F.2d 1355 (1971).

C. “Possible” Disruption as a Basis for Denying Pro Se Defense

The Government seeks to sustain the denial of appellants’ pro se motions on a theory of “possible disruption.” A list of five factors is offered15 which, it is said, “taken together” support the judge’s finding of risk of disruption.

Given the general likelihood that pro se defendants have only rudimentary acquaintanceship with the rules of evidence and courtroom protocol, a measure of unorthodoxy, confusion and delay is likely, perhaps inevitable, in pro se cases.16 The energy and time toll on the *1125trial judge, as fairness calls him to articulate ground rules and reasons that need not be explained to an experienced trial counsel, can be relieved, at least in part, by appointment of an amicus curiae to assist the defendant.17 If defendant refrains from intentionally obstructive tactics, amicus would be available to provide advice on procedure and strategy. The utility of an amicus appointment is dependent on explanation to and cooperation by defendant, and on understanding, too, that he may claim with some merit that his pro se rights include his right to appear before the jury in the status of one defending himself, and that this is defeated if a too conspicuous role is played by an attorney, unless it clearly appears to the jury that he does not have the status of defense counsel.18

On the other hand, a potentially unruly defendant may and should be'clearly forewarned that deliberate dilatory or obstructive behavior may operate in effect as a waiver of his pro se rights and, in that event, amicus will be ready to assume exclusive control of the defense.19 The Supreme Court has recently emphasized that even constitutional litigation prerogatives of a defendant are available to give choice in the conduct of a trial, and do not extend so far as to permit subversion of the core concept of a trial. Illinois v. Allen, 397 U.S. 337, 90 S.Ct. 1057, 25 L.Ed. 2d 353 (1970). The same principle means that obstreperous behavior may constitute waiver of the pro se right.20

Appointment of amicus counsel would not have resolved all the problems pre*1126sented by pro se defenses in the multi-defendant context involved in this case. Thus, the prospect of repetitious interrogation of witnesses would persist. But the joint trial that the prosecution seeks in the interest of efficiency cannot set aside the fundamental right of pro se representation. The trial judge must proceed by skill and suasion, by obtaining defendants’ cooperation, not by denying their pro se rights.

We need not !here consider whether or in what circumstances withholding of reasonable cooperation may be held equivalent to unruly action as a waiver of the right of self-representation. In the case before us, defendants and counsel assured the court, on several occasions, of their lack of disruptive intent, e. g., Tr. 23, 24. The judge in his oral opinion noted that his fears of disruption did not stem from concern over defendants’ “motivations.” Furthermore, the record shows how reasonable cooperation was obtained from defendants, by a reasonable accommodation of interests. When defendants expressed concern over their exclusion from bench conferences, Judge Pratt made, and defendants accepted (Tr. 333-335), a suggestion permitting a “representative” defendant to participate in bench conferences as an observer. A similar approach might have obviated any serious problems of repetitious interrogation.21 In the last analysis, however, if the assertion of a pro se right makes a multidefendant trial unmanageable, or unfair to the other defendants, the remedy lies in severance. Rule 14, F.R.Crim.P.

In effect the unqualified right of self representation rests on an implied presumption that the court will be able to achieve reasonable cooperation. The possibility that reasonable cooperation may be withheld, and the right later waived, is not a reason for denying the right of self representation at the start.

D. Lack of Foundation for Government Claims of Prior Disruptive Behavior

The Government argues that in this case there was disruptive behavior on the part of defendants which sustains the judge’s denial of pro se representation. We assume, without deciding, that where there has been experience with the particular defendants that is plainly identifiable as disruptive in character, such as to overturn the premise of reasonable cooperation, and permit a finding of anticipatory breach and waiver, that would be a predicate for denying the pro se right. We do not think any such predicate appears in this case.

We begin by rejecting the Government’s approach of using “disruptive” incidents following the denial of the pro se motions as reasons to support that denial.22 This is like using the fruit of an unreasonable search to provide a cause making the search reasonable. Nearly all of the incidents cited by the Government concerned assertions' of the right to self-representation. It would be anomalous to hold that the denial of one’s rights can be justified by reference to the nature of subsequent complaints protesting that denial.

As to defendants’ actions prior to the denial of their pro se requests, these *1127were not the kind of “disruptive” actions that warranted denial on that basis alone. We are aware of the occasions prior to the pro se ruling when defendants interrupted the pre-trial hearing without obtaining the court’s prior leave. However, most of the interruptions stemmed from defendants’ confusion over the exclusion of the public from the pre-trial hearing, a matter that had been arranged at the pre-trial conference in which defendants had not participated, and which was subject to reasonably prompt clarification without repetition.

Thus, during Sister Malone’s testimony, in which she inquired of the court the reasons for the exclusion of the public, appellant Dougherty interrupted to observe that he thought the courtroom was large enough to accommodate the people who could be expected to attend. This followed by moments an “interruption” by Mr. O’Rourke to make a request, joined in by Sister Malone, that two of their relatives be permitted to observe the progress of the hearing— a request that the court granted.

None of the incidents can be characterized as “disruptive” in the sense of evincing defendants’ intent to upset or unreasonably delay the hearing. Indeed to some extent the defendants, not trained in courtroom decorum, had reason to suppose their behavior was within proper bounds. At the hearing and during the early trial, Judge Pratt not only took considerable care to explain his ruling on the pro se motions, but also permitted the defendants to participate personally in jury selection. The record also shows that Judge Pratt engaged defendants in colloquies on various matters directly rather than through counsel. We do not disapprove, we rather commend, Judge Pratt’s willingness to handle this case with some flexibility. The Supreme Court has emphasized that one of the most important functions of criminal trials is, within reason, to make plain to defendants and society at large that justice is done in our courts,23 and Judge Pratt’s approach likely had that effect in this case. But the latitude previously granted to appellants must be taken into account in appraising whether their later requests manifest disruptive conduct.24 We cannot agree their pre-ruling behavior can be considered as so “disruptive” as to constitute a constructive, anticipatory waiver of a fundamental right.

E. Application of the Doctrine of Harmless Error

The Government finally contends that, assuming arguendo error in the denial of pro se defense, reversal as inappropriate because no prejudice resulted. We may assume, without deciding, that the harmless error doctrine — either in its ordinary formulation, or the more refined “harmless constitutional error” version25 — applies even in cases involving denial of a fundamental statutory right approximating or equalling the rights expressly stated in the Constitution. But we do not think it applicable to this case.

The principal characteristic of “harmless error” doctrine is its “result-orientation.” Its normal operation is in cases where the challenged error concerns *1128a right given the defendant in order to permit his defense to operate at maximum competence26 or to insulate him from the effects of suspect evidence.27 In such cases there is reason to consider whether claimed error is harmless because it plainly did not affect the result adversely to defendant, for then the reason for the right lapses.

Courts have recognized a measure of result-orientation in the right of pro se representation. The Second Circuit, for example, perceives a basis for the pro se right in the need not to force a defendant to accept a lawyer in whom he has little confidence. Without such confidence, lawyer-client communication is likely to be unsatisfactory and “defendant may be better off representing himself,” United States ex rel. Maldonado v. Denno, supra, 348 F.2d at 15.

However, a salient aspect of the pro se right, in our view, is directed to considerations distinct from the objective of achieving what would be the best result in the litigation from a lawyer’s point of view. As the Supreme Court said in Adams v. United States ex rel. McCann, supra, 317 U.S. at 279, 63 S.Ct. at 241, the “right to dispense with a lawyer’s help .... rest[s] on considerations that go to the substance of an accused’s position before the law.” It is designed to safeguard the dignity and autonomy of those whose circumstances or activities have thrust them involuntarily into the criminal process. An accused has a fundamental right to confront his accusers and his “country,” to present himself and his position to the jury not merely as a witness or through a “mouthpiece,” but as a man on trial who elects to plead his own cause. He is not obliged to seek what counsel would record as a victory but what he sees as tantamount to condemnation or doubt rather than vindication. A defendant has the moral right to stand alone in his hour of trial. The denial of that right is not to be redeemed through the prior estimate of someone else that the practical position of the defendant will be enhanced through representation by another, or the subsequent conclusion that defendant’s practical position has not been disadvantaged.

In guaranteeing counsel for the accused, the Sixth Amendment conferred a right for the benefit of the accused. As implemented by Congress, this right is not an imperative requirement that may be thrust upon him when in his judgment, as a person without impaired mental capacity, it is against his interest. Even if the defendant will likely lose the case anyway, he has the right — -as he suffers whatever consequences there may be — to the knowledge that it was the claim that he put forward that was considered and rejected, and to the knowledge that in our free society, devoted to the ideal of individual worth, he was not deprived of his free will to make his own choice, in his hour of trial, to handle his own case.28

In the case at bar defendants believed they would be vindicated by their peers by presenting their positions without law-trained counsel as intermediary. They may or may not be right about the relative effectiveness of a lawyer’s presentation of their case. Ordinarily repr.e-*1129sentation by an attorney has structure and clarity that enables a jury to better understand defendants’ positions. Presentation of a case pro se will often be artless and confusing. Yet the normal disadvantage of confusion may be offset by the enhanced intensity and appearance of greater sincerity of a defendant’s presentation.29

While we cannot apply the doctrine of “harmless error” in case of denial of right of pro se representation on the ground that this would likely have resulted in the same verdict as counseled representation, a different question arises as to whether the overall format of the trial was such, in terms of the latitude given to the defendants, that they in effect had the substance though not the form of pro se representation.

This is a closer question, for as we have noted the judge made a painstaking effort to take some account of defendants’ positions in the courtroom, and indeed gave permission to defendants to make brief opening statements and to testify informally in narrative fashion.

In the last analysis, the judge’s efforts did not suffice to dissipate the erroneous denial of pro se representation. The court’s indulgence of defendants undoubtedly appeared as exactly that, a matter of grace — as something “extra” given to them beyond their due rights — which undercuts the objective of preserving defendants’ personal autonomy and responsibility in the courtroom. The format provided by the judge allowed for brief opening statements by defendants not as representing themselves, as they had a right to appear, but as a supplement to the counsel appointed to represent them. The denial of the defendants’ right to make closing statements removed them from the place of responsibility at the climax of the trial— when they would be summing up their positions in a context of a summing up by prosecution counsel. Presentation of closing statements by the defendants was originally promised and then taken away because of their disruptions. But the vast bulk of the incidents cited in the Government’s brief as such disruptions — 81 out of 89 — were essentially colloquies in which the defendants were asserting their pro se rights. And except for the eruption as to the Vietnam war “issue,” the other incidents were relatively minor in character. But the abiding difficulty remains — that one cannot fairly reason backward, from the conduct of a defendant at a trial where he was denied the right to represent himself, to what his conduct would have been if at the outset the trial judge recognized that right and at the same time clarified the responsibilities of representation.

In finding error because of the denial of the pro se right we are not unmindful of the fact that this trial presented Judge Pratt with difficult problems, and that in many respects his conduct of the trial reflected a commendable approach, a humane and flexible spirit that would have served, if the right of self-representation had been granted, to curb unwelcome consequences of that right.30 *1130If it had not been for that threshold error, the trial would have had a different character. In view, however, of both the quality of, and limitations on, the permissions granted to defendants, we cannot say that the procedure followed was the equivalent of according the right of self-representation that was erroneously denied at the outset.

III. The Issue of Jury Nullification

Our reference to the “intensity” factor underlying the pro se right should not be understood as embracing the principle of “nullification” proffered by appellants. They say that the jury has a well-recognized prerogative to disregard the instructions of the court even as to matters of law, and that they accordingly have the legal right that the jury be informed of its power. We turn to this matter in order to define the nature of the new trial permitted by our mandate.

There has evolved in the Anglo-American system an undoubted jury prerogative-in-fact, derived from • its power to bring in a general verdict of not guilty in a criminal case, that is not reversible by the court. The power of the courts to punish jurors for corrupt or incorrect verdicts, which persisted after the medieval system of attaint by another jury became obsolete, was repudiated in 1670 when Bushell’s Case, 124 Eng.Rep. 1006 (C.P. 1670) discharged the jurors who had acquitted William Penn of unlawful assembly. Juries in civil cases became subject to the control of ordering a new trial; no comparable control evolved for acquittals in criminal cases.

The pages of history shine on instances of the jury’s exercise of its prerogative to disregard uncontradicted evidence and instructions of the judge. Most often commended are the 18th century acquittal of Peter Zenger of seditious libel, on the plea of Andrew Hamilton, and the 19th century acquittals in prosecutions under the fugitive slave law. The values involved drop a notch when the liberty vindicated by the verdict relates to the defendant’s shooting of his wife’s paramour, or purchase during Prohibition of alcoholic beverages.31

Even the notable Dean Pound commented in 1910 on positive aspects of “such jury lawlessness.”32 These observations of history and philosophy are underscored and illuminated, in terms of the current place of the jury in the American system of justice, by the empirical information and critical insights and analyses blended so felicitously in H. Kalven and H. Zeisel, The American Jury.33

*1131Reflective opinions upholding the necessity for the jury as a protection against arbitrary action, such as prose-cutorial abuse of power, stress fundamental features like the jury “common sense judgment” and assurance of “community participation in the determination of guilt or innocence.” 34 Human fraility *1132being what it is, a prosecutor disposed by unworthy motives could likely establish some basis in fact for bringing charges against anyone he wants to book, but the jury system operates in fact, (see note 33) so that the jury will not convict when they empathize with the defendant, as when the offense is one they see themselves as likely to commit, or consider generally acceptable or condonable under the mores of the community.

The existence of an unreviewable and unreversible power in the jury, to acquit in disregard of the instructions on the law given by the trial judge, has for many years co-existed with legal practice and precedent upholding instructions to the jury that they are required to follow the instructions of the court on all matters of law. There were different soundings in colonial days and the early days of our Republic. We are aware of the number and variety of expressions at that time from respected sources — John Adams; Alexander Hamilton; prominent judges — that jurors had a duty to find a verdict according to their own conscience, though in opposition to the direction of the court; that their power signified a right; that, they were judges both of law and of fact in a criminal case, and not bound by the opinion of the court.35

The rulings did not run all one way, but rather precipitated “a number of classic exchanges on the freedom and obligations of the criminal jury.”36 This was, indeed, one of the points of clash between the contending forces staking out the direction of the government of the newly established Republic, a direction resolved in political terms by reforming but sustaining the status of the courts, without radical change.37 As the distrust of judges appointed and removable by the king receded, there came increasing acceptance that under a republic the protection of citizens lay not in recognizing the right of each jury to make its own law, but in following democratic processes for changing the law.

The crucial legal ruling came in United States v. Battiste, 2 Sum. 240, Fed.Cas. No. 14,545 (C.C.D.Mass. 1835). Justice Story’s strong opinion supported the conception that the jury’s function lay in accepting the law given to it by the court and applying that law to the facts. This considered ruling of an influential jurist won increasing acceptance in the nation. The youthful passion for independence accommodated itself to the reality that the former rebels were now in control of their own destiny, that the practical needs of stability and sound growth outweighed the abstraction of centrifugal philosophy, and that the judges in the courts, were not the colonial appointees projecting royalist patronage and influence but were themselves part and parcel of the nation’s intellectual mainstream, subject to the checks of the common law tradition and professional opinion, and capable, in Roscoe Pound’s words, of providing “true judicial justice” standing in contrast with the colonial experience.38

*1133The tide was turned by Battiste, but there were cross-currents. At mid-century the country was still influenced by the precepts of Jacksonian democracy, which spurred demands for direct selection of judges by the people through elections, and distrust of the judge-made common law which enhanced the movement for codification reform. But by the end of the century, even the most prominent state landmarks had been toppled ;39 and the Supreme Court settled the matter for the Federal courts in Sparf v. United States, 156 U.S. 51, 102, 15 S.Ct. 273, 39 L.Ed. 343 (1895) after exhaustive review in both majority and dissenting opinions. The jury’s role was respected as significant and wholesome, but it was not to be given instructions that articulated a right to do whatever it willed. The old rule survives today only as a singular relic.40

The breadth of the continuing prerogative of the jury, however, perseveres, as appears from the rulings permitting inconsistent verdicts. These reflect, in the words of Justice Holmes, an acknowledgment that “the jury has the power to bring in a verdict in the teeth of both law and facts,”41 or as Judge Learned Hand said: “We interpret the acquittal as no more than their assumption of a power which they had no right to exercise, but to which they were disposed through lenity.” 42

Since the jury’s prerogative of lenity, again in Learned Hand’s words (supra, note 34) introduces a “slack into the enforcement of law, tempering its rigor by the mollifying influence of current ethical conventions,” it is only just, say appellants, that the jurors be so told. It is unjust to withhold information on the jury power of “nullification,” since conscientious jurors may come, ironically, to abide by their oath as jurors to render verdicts offensive to their individual conscience, to defer to an assumption of necessity that is contrary to reality.

This so-called right of jury nullification is put forward in the name of liberty and democracy, but its. explicit avowal risks the ultimate logic of anarchy. This is the concern voiced by Judge Sobeloff in United States v. Moylan, 417 F.2d 1002, 1009 (4th Cir. 1969), cert. denied, 397 U.S. 910, 90 S.Ct. 908, 25 L.Ed.2d 91 (1970):

To encourage individuals to make their own determinations as to which laws they will obey and which they *1134will permit themselves as a matter of conscience to disobey is to invite chaos. No legal system could long survive if it gave every individual the option of disregarding with impunity any law which by his personal standard was judged morally untenable. Toleration of such conduct would not be democratic, as appellants claim, but inevitably anarchic.

• The statement that avowal of the jury’s prerogative runs the risk of anarchy, represents, in all likelihood, the habit of thought of philosophy and logic, rather than the prediction of the social scientist. But if the statement contains an element of hyperbole, the existence of risk and danger, of significant magnitude, cannot be gainsaid. In contrast, the advocates of jury “nullification” apparently assume that the articulation of the jury’s power will not extend its use or extent, or will not do so significantly or obnoxiously. Can this assumption fairly be made ? We know that a posted limit of 60 m.p.h. produces factual speeds 10 or even 15 miles greater, with an understanding all around that some “tolerance” is acceptable to the authorities, assuming conditions warrant. But can it be supposed that the speeds would stay substantially the same if the speed limit were put: Drive as fast as you think appropriate, without the posted limit as an anchor, a point of departure ?

Our jury system is a resultant of many vectors, some explicit, and some rooted in tradition, continuity and general understanding without express formulation. A constitution may be meaningful though it is unwritten, as the British have proved for 900 years.

The jury system has worked out reasonably well overall, providing “play in the joints” that imparts flexibility and avoid undue rigidity. An equilibrium has evolved — an often marvelous balance — with the jury acting as a “safety valve” for exceptional cases, without being a wildcat or runaway institution. There is reason to believe that the simultaneous achievement of modest jury equity and avoidance of intolerable caprice depends on formal instructions that do not expressly delineate a jury charter to carve out its own rules of law.43 We have taken due and wry note that those whose writings44 acclaim and invoke Roscoe Pound’s 1910 recognition of the value of the jury as safety valve, omit mention of the fact that in the same article he referred to “the extreme decentralization that allows a local jury or even a. local prosecutor to hold up instead of uphold the law of the state” as one of the conditions that “too often result in a legal paralysis of legal administration,” 45 that his writings of that period are expressly concerned with the evils of the “extravagant powers” of juries,46 and that in 1931 he joined the other distin*1135guished members of the Wickersham Commission in this comment :47

In a number of jurisdictions juries are made judges of the law in criminal cases, thus inviting them to dispense with the rules of law instead of finding the facts. The juror is made judge of the law not to ascertain what it is, but to judge of its conformity to his personal ideals and ascertain its validity on that basis. . . . It is significant that there is most satisfaction with criminal juries in those jurisdictions which have interfered least with the conception of a trial of the facts unburdened with further responsibility and instructed as to the law and advised as to the facts by the judge.

The way the jury operates may be radically altered if there is alteration in the way it is told to operate. The jury knows well enough that its prerogative is not limited to the choices articulated in the formal instructions of the court.48 The jury gets its understanding as to the arrangements in the legal system from more than one voice. There is the formal communication from the judge. There is the informal communication from the total culture— literature (novel, drama, film, and television); current comment (newspapers, magazines and television); conversation; and, of course, history and tradition. The totality of input generally convey adequately enough the idea of prerogative, of freedom in an occasional case to depart from what the judge says. Even indicators that would on their face seem too weak to notice — like the fact that the judge tells the jury it must acquit (in case of reasonable doubt) but never tells the jury in so many words that it must convict — are a meaningful part of the jury’s total input. Law is a system, and it is also a language, with secondary meanings that may be unrecorded yet are part of its life.

When the legal system relegates the information of the jury’s prerogative to an essentially informal input, it is not being duplicitous, chargeable with chicane and intent to deceive. The limitation to informal input is, rather a governor to avoid excess: the prerogative is reserved for the exceptional case, and the judge’s instruction is retained as a generally effective constraint. We “recognize a constraint as obligatory upon us when we require not merely reason to defend our rule departures, but damn good reason.”49 The practicalities of men, machinery and rules point up the danger of articulating discretion to depart from a rule, that the breach will be more often and casually invoked. We cannot gainsay that occasionally jurors uninstructed as to the prerogative may feel themselves compelled to the point of rigidity.50 The danger of the excess rigidity that may now. occasionally exist is not as great as the danger of removing the boundaries of constraint provided by the announced rules.

We should also note the inter-relation of the unanimity requirement for petit juries, which was applicable to this trial, and is still the general rule though no longer constitutionally required for state courts.51 This is an additional reason— a material consideration, though neither a necessary nor sufficient condition— to brake the wheels of those who would tell the petit jurors they are to determine the rules of law, either directly or *1136by telling them they are free to disregard the judge’s statement of the rules. . The democratic principle would not be furthered, as proponents of jury nullification claim, it would be dis-served by investing in a jury that must be unanimous the function not merely of determining facts, hard enough for like-minded resolution, but of determining the rules of law.

Rules of law or justice involve choice of values and ordering of objectives for which unanimity is unlikely in any society, or group representing the society, especially a society as diverse in cultures and interests as ours. To seek unity out of diversity, under the national motto, there must be a procedure for decision by vote of a majority or prescribed plurality — in accordance with democratic philosophy. To assign the role of mini-legislature to the various petit juries, who must hang if not unanimous, exposes criminal law and administration to paralysis, and to a deadlock that betrays rather than furthers the assumptions of viable democracy.

Moreover, to compel a juror involuntarily assigned to jury duty to assume the burdens of mini-legislator or judge, as is implicit in the doctrine of nullification, is to put untoward strains on the jury system. It is one thing for a juror to know that the law condemns, but he has a factual power of lenity. To tell him expressly of a nullification prerogative, however, is to inform him, in effect, that it is he who fashions the rule that condemns. That is an overwhelming responsibility, an extreme burden for the jurors’ psyche. And it is not inappropriate to add that a juror called upon for an involuntary public service is entitled to the protection, when he takes action that he knows is right, but also knows is unpopular, either in the community at large or in his own particular grouping, that he can fairly put it to friends and neighbors that he was merely following the instructions of the court.

In the last analysis, our rejection of the request for jury nullification doctrine is a recognition that there are times when logic is not the only or even best guide to sound conduct of government. For machines, one can indulge the person who likes to tinker in pursuit of fine tuning. When men and judicial machinery are involved, one must attend to the many and complex mechanisms and reasons that lead men to change their conduct — when they know they are being studied; when they are told of the consequences of their conduct; and when conduct exercised with restraint as an unwritten exception is expressly presented as a legitimate option.

What makes for health as an occasional medicine would be disastrous as a daily diet. The fact that there is widespread existence of the jury’s prerogative, and approval of its existence as a “necessary counter to casehardened judges and arbitrary prosecutors,” 52 does not establish as an imperative that the jury must be informed by the judge of that power. On the contrary, it is pragmatically useful to structure , instructions in such wise that the jury must feel strongly about the values involved in the case, so strongly that it must itself identify the case as establishing a call of high conscience,53 and *1137must independently initiate and undertake an act in contravention of the established instructions. This requirement of independent jury conception confines the happening of the lawless jury to the occasional instance that does not violate, and viewed as an exception may even enhance, the over-all normative effect of the rule of law. An explicit instruction to a jury conveys an implied approval that runs the risk of degrading the legal structure requisite for true freedom, for an ordered liberty that protects against anarchy as well as tyranny.

Finally, we are aware that the denial of defendants’ request for a nullification instruction will be considered by them to negative some, or perhaps most, of the value of the right of pro se representation which we have recognized. This point could be answered in terms of logic: The right of self-representation is given for reasons recognized by the law, and cannot be a springboard to establish the validity of other advantages or conditions that lie in its tactical wake. Thus, a defendant’s ability to present his demeanor and often even a kind of testimony, without exposure to impeachment or cross-examination, may be a tactical consequence of pro se representation, and even a moving cause of its invocation, but this is not to say it is an objective of the law. But defendants’ position merits a more spacious answer, that lies outside the domain of formal logic. It is this. The jury system provides flexibility for the consideration of interests of justice outside the' formal rules of law. This embraces whatever extra the defendant conveys by personal representation, whether through demeanor or sincerity of justification. But it is subject to the overriding consideration that what is tolerable or even desirable as an informal, self-initiated exception, harbors grave dangers to the system if it is opened to expansion and intensification through incorporation in the judge’s instruction.

IV. The Jury Instructions

Finally, defendants assert that one segment of the charge to the jury, set forth in the margin,54 was coercive, *1138tantamount to a directed verdict of guilty, and outside the proper scope of a judicial instruction.

For the most part, defendants’ real complaint seems to be that the court stated the law applicable to the case. There is no contention — apart from the jury nullification claim, which we have rejected — that the charge was inaccurate as a statement of the applicable law. If a judge is to instruct the jury on the ultimate facts that are material under the law, he may properly advise the jury of what matters brought forward by defendants are not material under the applicable rule of law — as surely as he may charge that voluntary intoxication is no defense to a charge of second degree murder. Since it is the essence of the judicial function to declare the applicable law, it follows that the mere declaration of the law cannot be held outside the judicial function. This is tautological, but it is often the part of wisdom to be able to recognize which propositions are true tautologies. The jury were not told they must bring in a guilty verdict nor was there the kind of language or conduct, going beyond a declaration of the applicable law, that has in other cases 55 been held coercive, and an imroper departure from the role of the judge.

* * *

The judgment must be reversed, for the reasons stated in part II of this opinion, and the case remanded for new trial, if the Government be so advised, in accordance with parts III and IV.

So ordered.

BAZELON, Chief Judge,

concurring in part and dissenting in part:

I concur in the Court’s discussion of the statutory right of self-representation in criminal cases. In view of our holding that the statutory right is “unqualified,” our decision need not rest on the asserted constitutional right to proceed pro se. I emphasize, however, that my concurrence reflects no retreat from the position I expressed thirteen years *1139ago in Brown v. United States, 105 U.S. App.D.C. 77, 84, 264 F.2d 363, 370 (1959) (en banc) (dissenting opinion). I believed then, and I believe today, that the sixth amendment guarantees a defendant the right to act on his own behalf in resisting a criminal prosecution.

My disagreement with the Court concerns the issue of jury nullification. As the Court’s opinion clearly acknowledges, there can be no doubt that the jury has “an unreviewable and unreversible power * * * to acquit in disregard of the instructions on the law given by the trial judge * * *.” Majority opinion at 1132. More important, the Court apparently concedes — although in somewhat grudging terms — that the power of nullification is a “necessary counter to case-hardened judges and arbitrary prosecutors,” 1 2and that exercise of the power may, in at least some instances, “enhance, the over-all normative effect of the rule of law.” Id. at 1137. We could not withhold that concession without scoffing at the rationale that underlies the right to jury trial in criminal cases,2 and belittling some of the most legendary episodes in our political and jurisprudential history.3

The sticking point, however, is whether or not the jury should be told of its power to nullify the law in a particular case. Here, the trial judge not only denied a requested instruction on nullification, but also barred defense counsel from raising the issue in argument before the jury. The majority affirms that ruling. I see no justification for, and considerable harm in, this deliberate lack of candor.

At trial, the defendants made no effort to deny that they had committed the acts charged. Their defense was designed to persuade the jury that it would be unconscionable to convict them of violating a statute whose general validity and applicability they did not challenge. An instruction on nullification — or at least some argument to the jury on that issue — was, therefore, the linchpin of the defense.

At the outset it is important to recognize that the trial judge was not simply neutral on the question of nullification. His instruction, set out in part in the margin,4 emphatically denied the existence of a “legal defense” based on “sincere religious motives” or a belief that *1140action was justified by “some higher law.” That charge was not directly inconsistent with the theory of jury nullification. Nullification is not a “defense” recognized by law, but rather a mechanism that permits a jury, as community conscience,5 to disregard the strict requirements of law where it finds that those requirements cannot justly be applied in a particular case. Yet the impact of the judge’s instruction, whatever his intention, was almost surely to discourage the jury from measuring the de-

fendants’ action against community concepts of blameworthiness.

Thus, we are left with a doctrine that may “enhance the over-all normative effect of the rule of law,” but, at the same time, one that must not only be concealed from the jury, but also effectively condemned in the jury’s presence. Plainly, the justification for this sleight-of-hand lies in a fear that an occasionally noble doctrine will, if acknowledged, often be put to ignoble and abusive purposes — or, to borrow the Court’s phrase, will “run *1141the risk of anarchy.” Majority opinion at 1134. A breakdown of the legal order is not a result I would knowingly encourage or enjoy. But the question cannot be resolved, at least at this stage of the argument, by asking if we are for or against anarchy, or if we are willing to tolerate a little less law and order so that we can permit a little more jury nullification. No matter how horrible the effect feared by the Court, the validity of its reasoning depends on the existence of a demonstrable connection between the alleged cause (a jury nullification instruction or argument to the jury on that issue) and that effect. I am unable to see a connection.

To be sure, there are abusive purposes, discussed below, to which the doctrine might be put. The Court assumes that these abuses are most likely to occur if the doctrine is formally described to the jury by argument or instruction. That assumption, it should be clear, does not rest on any proposition of logic. It is nothing more or less than a prediction of how jurors will react to the judge’s instruction or argument by counsel. And since we have no empirical data to measure the validity of the prediction, we must rely on our own rough judgments of its plausibility.

The Court reasons that a jury uninformed of its power to nullify will invoke that power only where it “feel[s] strongly about the values involved in the case, so strongly that it [will] itself identify the case as establishing a call of high conscience * * Majority opinion at 1136. In other words, the spontaneous and unsolicited act of nullification is thought less likely, on the whole, to reflect bias and a perverse sense of values than the act of nullification carried out by a jury carefully instructed on its power and responsibility.

It seems substantially more plausible to me to assume that the very opposite is true. The juror motivated by prejudice seems to me more likely to make spontaneous use of the power to nullify, and more likely to disregard the judge’s exposition of the normally controlling legal standards. The conscientious juror, who could make a careful effort to consider the blameworthiness of the defendant’s action in light of prevailing community values, is the one most likely to obey the judge’s admonition that the jury enforce strict principles of law.

Moreover, if it were true that nullification which arises out of ignorance is in some sense more worthy than nullification which arises out of knowledge, the Court would have to go much further. For under the Court’s assumption, the harm does not arise because a jury is told of its power to disregard the law, but because it knows of its power. Logically construed, the Court’s opinion would seem to require the disqualification at voir dire of any prospective juror who admitted to knowledge of the doctrine. By excluding jurors with knowledge of the doctrine the Court could insure that its invocation would be spontaneous. And yet, far from requiring the exclusion of jurors who are aware of the power, the Court takes comfort in the fact that informal communication to the jury “generally convey [s] adequately enough the idea of prerogative, of freedom in an occasional case to depart from what the judge says.” Majority opinion at 1135. One cannot, it seems to me, have the argument both ways. If, as the Court appears to concede, awareness is preferable to ignorance, then I simply do not understand the justification for relying on a haphazard process of informal communication whose effectiveness is likely to depend, to a large extent, on whether or not any of the jurors are so well-educated and astute that they are able to receive the message. If the jury should know of its power to disregard the law, then the power should be explicitly described by instruction of the court or argument of counsel.

My own view rests on the premise that nullification can and should serve an important function in the criminal process. I do not see it as a doctrine that exists only because we lack the power to punish *1142jurors who refuse to enforce the law or to re-prosecute a defendant whose acquittal cannot be justified in the strict terms of law. The doctrine permits the jury to bring to bear on the criminal process a sense of fairness and particularized justice. The drafters of legal rules cannot anticipate and take account of every case where a defendant’s conduct is “unlawful” but not blameworthy, any more than they can draw a bold line to mark the boundary between an accident and negligence. It is the jury — as spokesman for the community’s sense of values — that must explore that subtle and elusive boundary.

Admittedly, the concept of blameworthiness does not often receive explicit recognition in the criminal process. But it comes very close to breaking through the surface in cases where the responsibility defense is raised, see United States v. Brawner, 153 U.S.App.D.C. 1, at 62, 471 F.2d 969, at 1030 (1972) (en banc)„ (separate opinion); United States v. Bennett, 148 U.S.App.D.C. 364, 368-370, 460 F.2d 872, 876-878 (1972); United States v. Eichberg, 142 U.S.App. D.C. 110, 113, 439 F.2d 620, 623 (1971) (concurring opinion), and it is implicit in every case where criminal sanctions are imposed. More than twenty-five years ago this Court recognized that “[o]ur collective conscience does not allow punishment where it cannot impose blame.” 6 And the Supreme Court, in a well-known opinion by Justice Jackson, has pointed out that

courts of various jurisdictions, and for the purposes of different offenses, have devised working formulae, if not scientific ones, for the instruction of juries around such terms as “felonious intent,” “criminal intent,” “malice aforethought,” “guilty knowledge,” “fraudulent intent,” “wilfulness,” “scienter,” to denote guilty knowledge, or “mens rea,” to signify an evil purpose or mental culpability. By use or combination of these various tokens, they have sought to protect those who were not blameworthy in mind from conviction of infamous common-law crimes.7

The very essence of the jury’s function is its role as spokesman for the community conscience in determining whether or not blame can be imposed.8

I do not see any reason to assume that jurors will make rampantly abusive use of their power. Trust in the jury is, after all, one of the cornerstones of our entire criminal jurisprudence, and if that trust is without foundation we must re-examine a great deal more than just the nullification doctrine. Nevertheless, *1143some abuse can be anticipated. If a jury refuses to apply strictly the controlling principles of law, it may — in conflict with values shared by the larger community — convict a defendant because of prejudice against him, or acquit a defendant because of sympathy for him and prejudice against his victim. Our fear of unjust conviction is plainly understandable. But it is hard for me to see how a nullification instruction could enhance the likelihood of that result. The instruction would speak in terms of acquittal, not conviction, and it would provide no comfort to a juror determined to convict a defendant in defiance of the law or the facts of the case. Indeed, unless the, jurors ignored the nullification instruction they could not convict on the grounds of prejudice alone. Does the judge’s recitation of the instruction increase the likelihood that the jury will ignore the limitation that lies at its heart ? I hardly think so.

As for the problem of unjust acquittal, it is important to recognize the strong internal check that constrains the jury’s willingness to acquit. Where defendants seem dangerous, juries are unlikely to exercise their nullification power, whether or not an explicit instruction is offered. Of course, that check will not prevent the acquittal of a defendant who may be blameworthy and dangerous except in the jaundiced eyes of a jury motivated by a perverse and sectarian sense of values. But whether a nullification instruction would make such acquittals more common is problematical, if not entirely inconceivable. In any case, the real problem in this situation is not the nullification doctrine, but the values and prejudice that prompt the acquittal.9 And the solution is not to condemn the nullification power, but to spotlight the prejudice and parochial values that underlie the verdict in the hope that public outcry will force a re-examination of those values, and deter their implementation in subsequent cases. Surely nothing is gained by the pretense that the jurors lack the power to nullify, since that pretense deprives them of the opportunity to hear the very instruction that might compel them to confront their responsibility.

One often-cited abuse of the nullification power is the acquittal by bigoted juries of whites who commit crimes (lynching, for example) against blacks.10 That repellent practice cannot be directly arrested without jeopardizing important constitutional protections — the double jeopardy bar and the jury’s power of nullification. But the revulsion and sense of shame fostered by that practice fueled the civil rights movement, which in turn made possible the enactment of major civil rights legislation. That same movement spurred on the revitalization of the equal protection clause and, in particular, the recognition of the right to be tried before a jury selected without bias.11 The lessons we learned from these abuses helped to create a climate in which such abuses could not so easily thrive.

Moreover, it is not only the abuses of nullification that can inform our understanding of the community’s values and standards of blameworthiness. The noble uses of the power — the uses that “enhance the over-all normative effect of the rule of law”- — also provide an important input to our evaluation of the substantive standards of the criminal law. The reluctance of juries to hold defendants responsible for unmistakable violations of the prohibition laws told us much about the morality of those laws and about the “criminality” of the conduct they proscribed. And the same can be said of the acquittals returned under the fugitive slave law12 as well as con*1144temporary gaming and liquor laws.13 A doctrine that can provide us with such critical insights should not be driven underground.

On remand the trial judge should grant defendants’ request for a nullification instruction. At the very least,14 I would require the trial court to permit defendants to argue the question before the jury. But it is not at all clear that defendants would prevail even with the aid of an instruction or argument. After all, this case is significantly different from the classic, exalted cases where juries historically invoked the power to nullify. Here, the defendants have no quarrel with the general validity of the law under which they have been charged. They did not simply refuse to obey a government edict that they considered illegal, and whose illegality they expected to demonstrate in a judicial proceeding. Rather, they attempted to protest government action by interfering with others — specifically, the Dow Chemical Company. This is a distinction which could and should be explored in argument before the jury. If revulsion against the war in Southeast Asia has reached a point where a jury would be unwilling to convict a defendant for commission of the acts alleged here, we would be far better advised to ponder the implications of that result than to spend our time devising stratagems which let us pretend that the power of nullification does not even exist.

ADAMS, Circuit Judge:

This case presents several difficult and subtle issues. Their resolution requires a sensitive understanding of human dignity, American legal and political history, and the interrelationship of those factors with the present criminal justice system, for it is this system that ensures that all persons may peacefully pursue their interests without undue interference from others.

Here, the defendants, by interrupting the business of a large chemical company, attempted to publicize their dissenting views regarding the morality of the American involvement in the Vietnam War. It is apparent that the defendants attempted to exploit their criminal trial by using it as a platform for further exposition of their beliefs, and to rely on their moral position as a defense of criminal charges lodged against them. In furtherance of their efforts to transform what would otherwise be an ordinary criminal trial into a “political” fray, defendants wished to represent themselves, to deviate from the usual mode of conducting a criminal trial, and to argue to the jury that although they were in violation of the applicable statutes, they should be acquitted because their actions were morally justified.

The rulings of the district court regarding (a) the defendants’ motions to proceed pro se, (b) the defendants’ right to have the court charge the jury as to the nullification issue, and (c) the content of the charge as actually given provide the foundation of this appeal.

In addressing these questions, some of which are to a large extent as philosophical as they are legal, the Court of Appeals is at some disadvantage — especially with regard to self representation — in evaluating the district court’s action from the cold record, divorced from the climate of the time when the trial took place. We must be mindful that the events in question here — defendants’ criminal acts as well as the trial itself— occurred during a period of passionate *1145and at times bitter political dissent. Several so-called “political” trials, including those of Dr. Benjamin Spock, the Catonsville Nine, and the Chicago Seven, occurred prior to this proceeding. In particular, it is significant that this trial closely followed that of the Chicago Seven, a trial wracked with violent disruptions when the defendants there challenged the underlying structure of the criminal justice system as we know it today.

In the context of what was transpiring in the judicial process in 1968 and 1969, the trial judge here was faced with a formidable task. On one hand, he had to preserve the judicial atmosphere vital in criminal proceedings to assure that both the Government and the defendants received a fair trial on the merits. The judge could not help but be aware that as a society we believe that an orderly, adversary trial by jury is a most efficacious means of arriving at the truth of the matters into which the court is inquiring. On the other hand, he had to take account of the asserted rights of the defendants to present themselves as moral persons — possessed with dignity and freedom of spirit, attributes upon which our system of Government was founded and built — as they sought to be acquittéd of their crimes. In evaluating whether the trial judge erred, we must be cognizant of the delicate balance he had to strike between ordered tranquility and America’s irrepressible desire for liberty.. Only if the court unduly weighted the scales against the defendants, either by his application of the facts of this case to the rules governing his actions or by his interpretation of the controlling law, may we reverse the convictions.

The apposite law, in my judgment, is accurately set forth in Judge Leventhal’s opinion. Specifically, I concur in his holding that defendants have a statutory right to proceed pro se in a federal criminal trial, and that, in this case, it is not necessary to decide the question in Constitutional terms.1 I also concur in the conclusion that the right to proceed pro se is waivable, either expressly or constructively, for example, by untimely assertion or by disruptive behavior. However, I disagree with Judge Leventhal’s factual conclusion that the defendant’s behavior here did not amount to a constructive waiver of the pro se right. Because I do not consider that the trial judge violated this right of the defendants, it is not necessary to decide whether such a violation may nevertheless be considered harmless error.

Since I would not reverse these convictions based on violation of defendants’ *1146right to self-representation, I need not express an opinion on the issues of jury nullification and the charge to the jury. With regard to these latter questions, Judge Leventhal’s statement of the law is wholly acceptable.

A careful review of the transcript of the proceedings with respect to defendants’ motions to dispense with counsel indicates that the defendants on many occasions interrupted the proceeding for various reasons, despite the fact they were represented by counsel at that time. The printed record cannot inform us whether such interruptions were violent, impolite, contemptuous, or made with an impertinent or sarcastic tone of voice. However more than once the judge had to threaten intervention by the marshals to impress upon defendants that they had to obey orders of the court to be seated. At other times, the language used by the defendants appeared to be disrespectful or even rude. That the trial judge appeared to tolerate many instances of this type of behavior, and even engaged in discourse with defendants, does not prove that he acquiesced in such extraordinary tactics, but rather indicates that he was at-témpting to preserve some sort of order.

Although it may be argued that there are some things in the judicial process more important than decorum in a courtroom, we are not faced with such perplexing alternatives here. Dissent is a healthy manifestation of the freedoms we as a nation profess and cherish. And a criminal trial might serve as the seed around which a point of view may crystallize. But a courtroom is not an arena in which dissention, particularly of a disruptive nature, may supplant, or even take precedence over, the task of administering justice.

As mentioned earlier, it is not without significance that this trial occurred at a time when the judicial system was being subjected to great stress because of the actions of some defendants and attorneys. The trial judge was surely aware of the publicity that attended the commission of the charged crimes, and the notoriety the trial itself would achieve. To paraphrase a slogan of the “new left”, he did not have to be a weatherman to téll which way the wind was blowing.

Although a gale may have been raging in the courts at that time, challenging the very assumptions upon which our system of justice is based, a counter-current was also evident. Only a month after the trial here in question, the Supreme Court delivered its opinion in Illinois v. Allen, supra. There, the various justices gave notice that they were aware of the storm and the damage it could wreak, and that they were determined that the judiciary have the ability to steer clear of the danger.2

*1147Certainly, what transpired here, coupled with the real threat of further disruption, was sufficient to alert the trial judge to what could be expected if the motions to dispense with counsel were granted. Thus, his determination that to grant the motions would turn the trial into a shambles does not appear to be unreasonable.

Accordingly, I would conclude that defendants had, by their conduct, waived their right to proceed pro se, and would affirm the rulings by the district court that the defendants in these circumstances had to be represented by counsel. Based on this and on Judge Leventhal’s opinion with regard to nullification and the charge, I would affirm the judgments of convictions.

On Appellee’s Petition for Rehearing and Suggestion for Rehearing En Banc

LEVENTHAL, Circuit Judge:

The Government’s petition for rehearing leads us to amplify our opinion.

1. The Government argues that defendant’s right to defend himself at trial is not absolute, and cites the Standards Relating to the Function of the Trial Judge (ABA Project on Standards for Criminal Justice, § 6.6, Tentative Draft, June 1972, at 85), which states in pertinent commentary:

Moreover, the interest of the public in an orderly, rational trial is entitled to consideration in determining the defendant’s right to appear pro se. See United States v. Bentvena, 319 F.2d 916, 937 (2d Cir. 1963); Butler v. United States, 317 F.2d 249, 258 (8th Cir. 1963).

We agree with this statement to the extent set forth in our opinion, including e. g., that a defendant’s disruption may waive his right to pro se representation. We note that both Butler, cited in our opinion (at fn. 14), and Bentvena, are cases involving requests for pro se representation made after the commencement of trial — a situation that, as our opinion points out, is entirely different from that of a pro se claim made timely before trial begins. Indeed in Bentvena the court stated, 319 F.2d at 938:

One charged with crime has an absolute right to do without an attorney and conduct his own defense (28 U.S. C. § 1654), but that is quite different from the right to discharge counsel after trial has begun. This latter right is a qualified one.

2. The petition for rehearing submits:

With particular regard to the value of an orderly trial, we believe that a trial judge should have authority to engraft reasonable conditions on the *1148exercise of a defendant’s right of self-representation.

It suffices to say that this case did not present a situation where the trial judge sought to prescribe reasonable conditions to accompany pro se representation. We do not have before us whether, or in what manner, a trial judge could prescribe conditions for the purpose of assuring a trial without disruption — to carry out the implied premise that there will be reasonable cooperation from the defendants, as opposed to any manipulation of the trial process so as to interfere with fair administration of justice.

3. The Government stresses a contention that the court’s opinion improperly bypasses the doctrine of harmless error. Apart from the matters set forth in the opinion already filed, we observe that if the conventional doctrine of harmless error is applied, it will in effect undercut the right of pro. se representation, since rarely if ever can there be a showing of prejudice in terms of result from the conduct of a trial by counsel. There is no requirement of a showing of prejudice when counsel has been denied. A like rule is applicable, although for different reasons, when pro se representation is denied.

4. The Government objects that our opinion means that any denial of pro se representation will per se require reversal. It does not require reversal in a case where the trial judgé makes findings of defendants’ prior disruption, or refusal to assure reasonable cooperation, or inability to waive counsel, etc. The effort to cope with the problem of disruptions necessarily puts an obligation on the trial judge * to stake out the considerations explicitly and with care. In the absence of some exposition of an appropriate basis for denying his right, the defendant is entitled to a trial at which he is accorded his right to represent himself.

Petition for rehearing denied.

Circuit Judge ADAMS dissents.