GRAHAM v. FLORIDA ( No. 08-7412 )
982 So. 2d 43, reversed and remanded.

HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version

Thomas, J., dissenting



on writ of certiorari to the district court of appeal of florida, first district

[May 17, 2010]

Justice 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 nonhomicide 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 of whether this sentence can ever be “proportionat[e]” when applied to the category of offenders at issue here. Ante , at 7 (internal quotation marks omitted), ante , at 1 ( 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.


The Court recounts the facts of Terrance Jamar Graham’s case in detail, so only a summary is necessary here. At age 16 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), 810.02(2)(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 21.



The Eighth Amendment , which applies to the States through the Fourteenth, provides that “[e]xcessive bail shall not 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 understand- ing 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 8 (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 considered 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. ___ (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 under 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, “[s]ociety changes,” ante , at 1 (S tevens, 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 7 (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 10 (quoting Roper, supra , at 563); see also ante , at 8–15, 29–31. 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 (S calia, 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] legislative con- sideration” of new questions of penal policy as they emerge, Kennedy , supra , at ___ (slip op., at 2) (A lito, J., dissenting).

But the Court is not content to rely on snapshots of community consensus in any event. Ante , at 16 (“Community consensus, while ‘entitled to great weight,’ is not itself determinative” (quoting Kennedy , supra , at __ (slip op., at 24)). 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 16. 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 7 (quoting Kennedy , supra , at ___ (slip op., at 8)).

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.


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 distinction 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, 543 U. S., at 568 (quoting Atkins, 536 U. S., at 319), but also to declare that “less culpable” persons are categorically exempt from the “ second most severe penalty.” Ante , at 21 (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 excessiveness 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) (striking 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 (opinion of Kennedy, J.) (“[ M]arked 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 (“[D]iffering attitudes and perceptions of local conditions may yield different, yet rational, conclusions regarding the appropriate length of prison terms for particular crimes”) . T hat 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[s],” not the courts, Ewing , supra, at 27, 28 (plurality opinion). The Court correctly notes that those decisions were “closely divided,” ante , at 8, but so was Solem itself, and it is now fair to describe Solem as an outlier. 2

Remarkably, the Court today does more than return to Solem ’s case-by-case 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 nonhomicide offender, no matter what the circumstances. Neither the Eighth Amendment nor the Court’s precedents justify this decision.


The Court asserts that categorical proportionality review is necessary here merely because Graham asks for a categorical rule, see ante , at 10, and because the Court thinks clear lines are a good idea, see ante , at 24–25. 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 nonhomicide 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.


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 10–11 (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 33–35 (Appendix to opinion of the Court). Forty-five States, the Federal Government, and the District of Columbia expose juvenile offenders charged in adult court to the very same range of punishments faced by adults charged with the same crimes. See ante , at 33–34, 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 are measures of consensus other than legislation.’ ” Ante , at 11 (quoting Kennedy , 554 U. S., at ___ (slip op., at 22)). 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 leg-islation—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 the 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 16–17, 22–24. 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 crack[ed] 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. 1987), 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.


The Court nonetheless dismisses existing legislation, pointing out that life-without-parole sentences are rarely imposed on juvenile nonhomicide offenders—129 times in recent memory 9 by the Court’s calculation, spread out across 11 States and the federal courts. Ante , at 11–13. Based on this rarity of use, the 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 moral values of their constituents that this Court claims to have easily discerned from afar. See ante , at 11.

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 16.

As an initial matter, even accepting the Court’s theory, federal law authorizes this penalty and the Federal Government uses it. See ante , at 13 (citing Letter and Attachment from Judith Simon Garrett, U. S. Dept. of Justice, Federal Bureau of Prisons, to Supreme Court Library (Apr. 12, 2010) (available in Clerk of Court’s case file)). That should be all the evidence necessary to refute the claim of a national consensus against this penalty.

Yet even when examining the States that authorize, but have not recently employed, this sentencing practice, the Court’s theory is unsound. Under the Court’s evolving standards 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). In light of this fact, the Court is wrong to equate a jurisdiction’s disuse of a legislatively 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 juveniles 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 Oklahoma 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 21–22. 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 129 juvenile nonhomicide 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. 10

Finally, 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. 11 By the Court’s own decree, “[c]ommunity consensus … is not itself determinative.” Ante , at 16. Only the independent moral judgment of this Court is sufficient to decide the question. See ibid.


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.” Ante , at 16. 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 20 (quoting Harmelin , 501 U. S., at 999 (opinion of Kennedy, J. )), 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 nonhomicide offenders serves two “legitimate” penological goals: incapacitation and deterrence. Ante , at 20–21. 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 9 (opinion concurring in judgment). That should settle the matter, since the Court acknowledges that incapacitation is an “important” penological goal. Ante , at 21. Yet, the Court finds this goal “inadequate to justify the life-without-parole sentences here. Ante, at 22 (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 penological goal, not as illegitimate, but as insufficient. Ante , at 21 (“[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 22. Of course, the Court recognizes that rehabilitation’s “utility and proper implementation” are subject to debate. Ante, at 23. But that does not stop it from declaring that a legislature may not “forswea[r] … the rehabilitative ideal.” Ibid. In other words, the Eighth Amendment does not mandate “any one penological theory,” ante , at 20 (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 16, 20 (quoting Tison v. Arizona , 481 U. S. 137, 149 (1987) ). The Court finds that retributive purposes are not served here for two reasons.


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 17. As a general matter, this statement is entirely consistent with the evidence 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 17 (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 26. 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 “drift[ing] through successive systems aimed at curbing their deviance”). That research further suggests that the pattern of behavior in the latter group often sets in before 18. See Moffitt , supra , at 684 (“The well-documented resistance of antisocial personality disorder to treatments of all kinds seems to suggest that the life-course-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 17 (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 cold-blooded nature’ ” of a juvenile’s nonhomicide crime, will sentence him to a life-without-parole sentence for which he possesses “ ‘insufficient culpability,’ ” ante , at 27 (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 presented. 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.


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 23. 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 ante , at 17–19; cf. ante , at 9 ( 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 nonhomicide 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 18.

That 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 4. 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.


Although the 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 intra- and interjurisdictional compari- son of the defendant’s sentence with others similarly situated. Harmelin , 501 U. S., at 1000, 1005 (opinion of Kennedy, J.).

Under 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 7 ( Roberts, C. J., concurring in judgment). 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 5 ( Roberts, C. J., concurring in judgment) (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 case-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 7, apparently because it believes that Graham’s armed burglary and home invasion crimes were “certainly less serious” than murder or rape, ibid . 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 10. 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 in Florida and concludes that Graham’s sentence was “far more severe.” Ante , at 8 ( 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 ante , at 8–9, this hardly seems relevant if one takes seriously the principle that “ ‘[a]bsent 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.

*  *  *

Both 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 24. 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. 12


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 of 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 il-lustrates 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 1 (concurring opinion). Perhaps one day the Court will learn from this one.

I respectfully dissent.


1 The Chief Justice’s concurrence suggests that it is unnecessary to remark on the underlying question whether the Eighth Amendment requires proportionality in sentencing because “[n]either party here asks us to reexamine our precedents” requiring “proportionality between noncapital offenses and their corresponding punishments.” Ante, at 2 (opinion concurring in judgment). I disagree. Both the Court and the concurrence do more than apply existing noncapital proportionality precedents to the particulars of Graham’s claim. The Court radically departs from the framework those precedents establish by applying to a noncapital sentence the categorical proportionality review its prior decisions have reserved for death penalty cases alone. See Part III, infra. The concurrence, meanwhile, breathes new life into the case-by-case proportionality approach that previously governed noncapital cases, from which the Court has steadily, and wisely, retreated since Solem v. Helm, 463 U. S. 277 (1983) . See Part IV, infra. In dissenting from both choices to expand proportionality review, I find it essential to reexamine the foundations on which that doctrine is built.

2 Courts and commentators interpreting this Court’s decisions have reached this conclusion. See, e.g., United States v. Polk, 546 F. 3d 74, 76 (CA1 2008) (“[I]nstances of gross disproportionality [in noncapital cases] will be hen’s-teeth rare”); Barkow, The Court of Life and Death: The Two Tracks of Constitutional Sentencing Law and the Case for Uniformity, 107 Mich. L. Rev. 1145, 1160 (2009) (“Solem now stands as an outlier”); Note, The Capital Punishment Exception: A Case for Constitutionalizing the Substantive Criminal Law, 104 Colum. L. Rev. 426, 445 (2004) (observing that outside of the capital context, “proportionality review has been virtually dormant”); Steiker & Steiker, Opening a Window or Building a Wall? The Effect of Eighth Amendment Death Penalty Law and Advocacy on Criminal Justice More Broadly, 11 U. Pa. J. Const. L. 155, 184 (2009) (“ Eighth Amendment challenges to excessive incarceration [are] essentially non-starters”).

3 The Court ignores entirely the threshold inquiry of whether subjecting juvenile offenders to adult penalties was one of the “modes or acts of punishment that had been considered cruel and unusual at the time that the Bill of Rights was adopted.” Ford v. Wainwright, 477 U. S. 399, 405 (1986) . As the Court has noted in the past, however, the evidence is clear that, at the time of the Founding, “the common law set a rebuttable presumption of incapacity to commit any felony at the age of 14, and theoretically permitted [even] capital punishment to be imposed on a person as young as age 7.” Stanford v. Kentucky, 492 U. S. 361, 368 (1989) (citing 4 W. Blackstone, Commentaries *23–*24; 1 M. Hale, Pleas of the Crown 24–29 (1800)). It thus seems exceedingly unlikely that the imposition of a life-without-parole sentence on a person of Graham’s age would run afoul of those standards.

4 Although the details of state laws vary extensively, they generally permit the transfer of a juvenile offender to adult court through one or more of the following mechanisms: (1) judicial waiver, in which the juvenile court has the authority to waive jurisdiction over the offender and transfer the case to adult court; (2) concurrent jurisdiction, in which adult and juvenile courts share jurisdiction over certain cases and the prosecutor has discretion to file in either court; or (3) statutory provisions that exclude juveniles who commit certain crimes from juvenile-court jurisdiction. See Dept. of Justice, Juvenile Offenders and Victims: 1999 National Report 89, 104 (1999) (hereinafter 1999 DOJ National Report); Feld, Unmitigated Punishment: Adolescent Criminal Responsibility and LWOP Sentences, 10 J. Law & Family Studies 11, 38–39 (2007).

5 Alaska entitles all offenders to parole, regardless of their crime. Alaska Stat. §12.55.015(g) (2008). The other seven States provide parole eligibility to all offenders, except those who commit certain homicide crimes. Conn. Gen. Stat. §53a–35a (2009); Haw. Rev. Stat. §§706–656(1)–(2) (1993 and 2008 Supp. Pamphlet); Me. Rev. Stat. Ann., Tit. 17–a, §1251 (2006); Mass. Gen. Laws Ann., ch. 265, §2 (West 2008); N. J. Stat. Ann. §§2C:11–3(b)(2)–(3) (West 2005); N. M. Stat. Ann. §31–18–14 (Supp. 2009); Vt. Stat. Ann., Tit. 13, §2303 (2009).

6 Colo. Rev. Stat. Ann. §18–1.3–401(4)(b) (2009) (authorizing mandatory life sentence with possibility for parole after 40 years for juveniles convicted of class 1 felonies); Kan. Stat. Ann. §§21–4622, 4643 (2007); Ky. Rev. Stat. Ann. §640.040 (West 2006); Shepherd v. Commonwealth, 251 S. W. 3d 309, 320–321 (Ky. 2008); Mont. Code Ann. §46–18–222(1) (2009); Tex. Penal Code Ann. §12.31 (West Supp. 2009).

7 Although the Court previously has dismissed the relevance of the Uniform Code of Military Justice to its discernment of consensus, see Kennedy v. Louisiana, 554 U. S. ___, ___ (2008) (statement of Kennedy, J., respecting denial of rehearing), juveniles who enlist in the military are nonetheless eligible for life-without-parole sentences if they commit certain nonhomicide crimes. See 10 U. S. C. §§505(a) (permitting enlistment at age 17), 856a, 920 (2006 ed., Supp. II).

8 Kennedy, 554 U. S., at ___ (slip op., at 12, 23) (prohibiting capital punishment for the rape of a child where only six States had enacted statutes authorizing the punishment since Furman v. Georgia, 408 U. S. 238 (1972) (per curiam)); Roper v. Simmons, 543 U. S. 551, 564, 568 (2005) (prohibiting capital punishment for offenders younger than 18 where 18 of 38 death-penalty States precluded imposition of the penalty on persons under 18 and the remaining 12 States did not permit capital punishment at all); Atkins v. Virginia, 536 U. S. 304, 314–315 (2002) (prohibiting capital punishment of mentally retarded persons where 18 of 38 death-penalty States precluded imposition of the penalty on such persons and the remaining States did not authorize capital punishment at all); Thompson v. Oklahoma, 487 U. S. 815, 826, 829 (1988) (plurality opinion) (prohibiting capital punishment of offenders under 16 where 18 of 36 death-penalty States precluded imposition of the penalty on such persons and the remaining States did not permit capital punishment at all); Enmund v. Florida, 458 U. S. 782, 789 (1982) (prohibiting capital punishment for felony murder without proof of intent to kill where eight States allowed the punishment without proof of that element); Coker v. Georgia, 433 U. S. 584, 593 (1977) (holding capital punishment for the rape of a woman unconstitutional where “[a]t no time in the last 50 years have a majority of the States authorized death as a punishment for rape”).

9 I say “recent memory” because the research relied upon by the Court provides a headcount of juvenile nonhomicide offenders presently incarcerated in this country, but does not provide more specific information about all of the offenders, such as the dates on which they were convicted.

10 Because existing legislation plainly suffices to refute any consensus against this sentencing practice, I assume the accuracy of the Court’s evidence regarding the frequency with which this sentence has been imposed. But I would be remiss if I did not mention two points about the Court’s figures. First, it seems odd that the Court counts only those juveniles sentenced to life without parole and excludes from its analysis all juveniles sentenced to lengthy term-of-years sentences (e.g., 70 or 80 years’ imprisonment). It is difficult to argue that a judge or jury imposing such a long sentence—which effectively denies the offender any material opportunity for parole—would express moral outrage at a life-without-parole sentence. Second, if objective indicia of consensus were truly important to the Court’s analysis, the statistical information presently available would be woefully inadequate to form the basis of an Eighth Amendment rule that can be revoked only by constitutional amendment. The only evidence submitted to this Court regarding the frequency of this sentence’s imposition was a single study completed after this Court granted certiorari in this case. See P. Annino, D. Rasmussen, & C. Rice, Juvenile Life Without Parole for Non-Homicide Offenses: Florida Compared to Nation 2 (Sept. 14, 2009). Although I have no reason to question the professionalism with which this study was conducted, the study itself acknowledges that it was incomplete and the first of its kind. See id., at 1. The Court’s questionable decision to “complete” the study on its own does not materially increase its reliability. For one thing, by finishing the study itself, the Court prohibits the parties from ever disputing its findings. Complicating matters further, the original study sometimes relied on third-party data rather than data from the States themselves, see ibid.; the study has never been peer reviewed; and specific data on all 129 offenders (age, date of conviction, crime of conviction, etc.), have not been collected, making verification of the Court’s headcount impossible. The Court inexplicably blames Florida for all of this. See ante, at 12. But as already noted, it is not Florida’s burden to collect data to prove a national consensus in favor of this sentencing practice, but Graham’s “heavy burden” to prove a consensus against it. See supra, at 16.

11 I confine to a footnote the Court’s discussion of foreign laws and sentencing practices because past opinions explain at length why such factors are irrelevant to the meaning of our Constitution or the Court’s discernment of any longstanding tradition in this Nation. See Atkins, 536 U. S., at 324–325 (Rehnquist, C. J., dissenting). Here, two points suffice. First, despite the Court’s attempt to count the actual number of juvenile nonhomicide offenders serving life-without-parole sentences in other nations (a task even more challenging than counting them within our borders), the laws of other countries permit juvenile life-without-parole sentences, see Child Rights Information, Network, C. de la Vega, M. Montesano, & A. Solter, Human Rights Advocates, Statement on Juvenile Sentencing to Human Rights Council, 10th Sess. (Nov. 3, 2009) (“Eleven countries have laws with the potential to permit the sentencing of child offenders to life without the possibility of release”), online at (as visited May 14, 2010, and available in Clerk of Court’s case file)). Second, present legislation notwithstanding, democracies around the world remain free to adopt life-without-parole sentences for juvenile offenders tomorrow if they see fit. Starting today, ours can count itself among the few in which judicial decree prevents voters from making that choice.

12 It bears noting that Colorado, one of the five States that prohibit life-without-parole sentences for juvenile nonhomicide offenders, permits such offenders to be sentenced to mandatory terms of imprisonment for up to 40 years. Colo. Rev. Stat. §18–1.3–401(4)(b) (2009). In light of the volume of state and federal legislation that presently permits life-without-parole sentences for juvenile nonhomicide offenders, it would be impossible to argue that there is any objective evidence of agreement that a juvenile is constitutionally entitled to a parole hearing any sooner than 40 years after conviction. See Tr. of Oral Arg. 6–7 (counsel for Graham, stating that, “[o]ur position is that it should be left up to the States to decide. We think that the … Colorado provision would probably be constitutional”).