Graham v. Collins

506 U.S. 461 (1993)

GRAHAM
v.
COLLINS, DIRECTOR, TEXAS DEPARTMENT OF CRIMINAL JUSTICE, INSTITUTIONAL DIVISION

No. 91-7580.

United States Supreme Court.

Argued October 14, 1992. Decided January 25, 1993. CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT

White, J., delivered the opinion of the Court, in which Rehnquist, C. J., and Scalia, Kennedy, and Thomas, JJ., joined. Thomas, J., filed a concurring opinion, post, p. 478. Stevens, J., filed a dissenting opinion, post, p. 500. Souter, J., filed a dissenting opinion, in which Blackmun, Stevens, and O'Connor, JJ., joined, post, p. 504.

Michael E. Tigar argued the cause for petitioner. With him on the briefs was Jeffrey J. Pokorak.

Charles A. Palmer, Assistant Attorney General of Texas, argued the cause for respondent. With him on the brief were Dan Morales, Attorney General, William C. Zapalac, Assistant Attorney General, Will Pryor, First Assistant Attorney General, Mary F. Keller, Deputy Attorney General, and Michael P. Hodge, Assistant Attorney General.^[*]

Justice White, delivered the opinion of the Court.

In this case, we are asked to decide whether the jury that sentenced petitioner, Gary Graham, to death was able to give effect, consistent with the Eighth and Fourteenth Amendments, to mitigating evidence of Graham's youth, family background, and positive character traits. Because this case comes to us on collateral review, however, we must first decide whether the relief that petitioner seeks would require announcement of a new rule of constitutional law, in contravention of the principles set forth in Teague v. Lane, 489 U.S. 288 (1989). Concluding that Graham's claim is barred by Teague, we affirm.

I

On the night of May 13, 1981, Graham accosted Bobby Grant Lambert in the parking lot of a Houston, Texas, grocery store and attempted to grab his wallet. When Lambert resisted, Graham drew a pistol and shot him to death. Five months later, a jury rejected Graham's defense of mistaken identity and convicted him of capital murder in violation of Tex. Penal Code Ann. § 19.03(a)(2) (1989).

At the sentencing phase of Graham's trial, the State offered evidence that Graham's murder of Lambert commenced a week of violent attacks during which the 17-yearold Graham committed a string of robberies, several assaults, and one rape. Graham did not contest this evidence. Rather, in mitigation, the defense offered testimony from Graham's stepfather and grandmother concerning his upbringing and positive character traits. The stepfather, Joe Samby, testified that Graham, who lived and worked with his natural father, typically visited his mother once or twice a week and was a "real nice, respectable" person. Samby further testified that Graham would pitch in on family chores and that Graham, himself a father of two young children, would "buy . . . clothes for his children and try to give them food."

Graham's grandmother, Emma Chron, testified that Graham had lived with her off and on throughout his childhood because his mother had been hospitalized periodically for a "nervous condition." Chron also stated that she had never known Graham to be violent or disrespectful, that he attended church regularly while growing up, and that "[h]e loved the Lord." In closing arguments to the jury, defense counsel depicted Graham's criminal behavior as aberrational and urged the jury to take Graham's youth into account in deciding his punishment.

In accord with the capital sentencing statute then in effect,^[1] Graham's jury was instructed that it was to answer three "special issues":

"(1) whether the conduct of the defendant that caused the death of the deceased was committed deliberately and with the reasonable expectation that the death of the deceased or another would result;

(2) whether there is a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society; and

(3) if raised by the evidence, whether the conduct of the defendant in killing the deceased was unreasonable in response to the provocation, if any, by the deceased." Tex. Code Crim. Proc. Ann., Art. 37.071(b) (Vernon 1981).

The jury unanimously answered each of these questions in the affirmative, and the court, as required by the statute, sentenced Graham to death. Art. 37.071(e). The Texas Court of Criminal Appeals affirmed Graham's conviction and sentence in an unpublished opinion.

In 1987, Graham unsuccessfully sought postconviction relief in the Texas state courts. The following year, Graham petitioned for a writ of habeas corpus in Federal District Court pursuant to 28 U.S. C. § 2254, contending, inter alia, that his sentencing jury had been unable to give effect to his mitigating evidence within the confines of the statutory "special issues." The District Court denied relief and the Court of Appeals for the Fifth Circuit denied Graham's petition for a certificate of probable cause to appeal. Graham v. Lynaugh, 854 F.2d 715 (1988). The Court of Appeals found Graham's claim to be foreclosed by our recent decision in Franklin v. Lynaugh, 487 U.S. 164 (1988), which held that a sentencing jury was fully able to consider and give effect to mitigating evidence of a defendant's clean prison disciplinary record by way of answering Texas' special issues. 854 F.2d, at 719-720.

While Graham's petition for a writ of certiorari was pending here, the Court announced its decision in Penry v. Lynaugh, 492 U.S. 302 (1989), holding that evidence of a defendant's mental retardation and abused childhood could not be given mitigating effect by a jury within the framework of the special issues.^[2] We then granted Graham's petition, vacated the judgment below, and remanded for reconsideration in light of Penry. Graham v. Lynaugh, 492 U.S. 915 (1989). On remand, a divided panel of the Court of Appeals reversed the District Court and vacated Graham's death sentence. 896 F.2d 893 (CA5 1990).

On rehearing en banc, the Court of Appeals vacated the panel's decision and reinstated its prior mandate affirming the District Court. 950 F.2d 1009 (1992). The court reviewed our holdings on the constitutional requirement that a sentencer be permitted to consider and act upon any relevant mitigating evidence put forward by a capital defendant, and then rejected Graham's claim on the merits. The court noted that this Court had upheld the Texas capital sentencing statute against a facial attack in Jurek v. Texas, 428 U.S. 262 (1976), after acknowledging that "`the constitutionality of the Texas procedures turns on whether the enumerated questions allow consideration of particularized mitigating factors.' " 950 F.2d, at 1019 (quoting Jurek, supra, at 272). Noting that the petitioner in Jurek had himself proffered mitigating evidence of his young age, employment history, and aid to his family, the Court of Appeals concluded that "[a]t the very least, Jurek must stand for the proposition that these mitigating factors—relative youth and evidence reflecting good character traits such as steady employment and helping others—are adequately covered by the second special issue" concerning the defendant's risk of future dangerousness. 950 F.2d, at 1029. "Penry cannot hold otherwise," the court observed, "and at the same time not be a `new rule' for Teague purposes." Ibid. Accordingly, the court ruled that the jury that sentenced Graham could give adequate mitigating effect to his evidence of youth, unstable childhood, and positive character traits by way of answering the Texas special issues.

We granted certiorari, 504 U.S. 972 (1992), and now affirm.

II

A

Because this case is before us on Graham's petition for a writ of federal habeas corpus, "we must determine, as a threshold matter, whether granting him the relief he seeks would create a `new rule' " of constitutional law. Penry v. Lynaugh, supra, at 313; see also Teague v. Lane, 489 U. S., at 301 (plurality opinion). "Under Teague, new rules will not be applied or announced in cases on collateral review unless they fall into one of two exceptions." Penry, supra, at 313. This restriction on our review applies to capital cases as it does to those not involving the death penalty. 492 U.S., at 314; Stringer v. Black, 503 U.S. 222 (1992); Sawyer v. Smith, 497 U.S. 227 (1990); Saffle v. Parks, 494 U.S. 484 (1990); Butler v. McKellar, 494 U.S. 407 (1990).

A holding constitutes a "new rule" within the meaning of Teague if it "breaks new ground," "imposes a new obligation on the States or the Federal Government," or was not "dictated by precedent existing at the time the defendant's conviction became final." Teague, supra, at 301 (emphasis in original). While there can be no dispute that a decision announces a new rule if it expressly overrules a prior decision, "it is more difficult . . . to determine whether we announce a new rule when a decision extends the reasoning of our prior cases." Saffle v. Parks, 494 U. S., at 488. Because the leading purpose of federal habeas review is to "ensur[e] that state courts conduct criminal proceedings in accordance with the Constitution as interpreted at the time of th[ose] proceedings," ibid., we have held that "[t]he `new rule' principle. . . validates reasonable, good-faith interpretations of existing precedents made by state courts." Butler v. McKellar, 494 U. S., at 414. This principle adheres even if those goodfaith interpretations "are shown to be contrary to later decisions." Ibid. Thus, unless reasonable jurists hearing petitioner's claim at the time his conviction became final "would have felt compelled by existing precedent" to rule in his favor, we are barred from doing so now. Saffle v. Parks, supra, at 488.

B

Petitioner's conviction and sentence became final on September 10, 1984, when the time for filing a petition for certiorari from the judgment affirming his conviction expired. See Griffith v. Kentucky, 479 U.S. 314, 321, n. 6 (1987). Surveying the legal landscape as it then existed, we conclude that it would have been anything but clear to reasonable jurists in 1984 that petitioner's sentencing proceeding did not comport with the Constitution.

1

In the years since Furman v. Georgia, 408 U.S. 238 (1972), the Court has identified, and struggled to harmonize, two competing commandments of the Eighth Amendment. On one hand, as Furman itself emphasized, States must limit and channel the discretion of judges and juries to ensure that death sentences are not meted out "wantonly" or "freakishly." Id., at 310 (Stewart, J., concurring). On the other, as we have emphasized in subsequent cases, States must confer on the sentencer sufficient discretion to take account of the "character and record of the individual offender and the circumstances of the particular offense" to ensure that "death is the appropriate punishment in a specific case." Woodson v. North Carolina, 428 U.S. 280, 304-305 (1976) (plurality opinion of Stewart, Powell, and Stevens, JJ.).

Four years after Furman, and on the same day that Woodson was announced, the Court in Jurek v. Texas, supra, examined the very statutory scheme under which Graham was sentenced and concluded that it struck an appropriate balance between these constitutional concerns. The Court thus rejected an attack on the entire statutory scheme for imposing the death penalty and in particular an attack on the so-called "special issues." It is well to set out how the Court arrived at its judgment. The joint opinion of Justices Stewart, Powell, and Stevens observed that while Texas had not adopted a list of aggravating circumstances that would justify the imposition of the death penalty, "its action in narrowing the categories of murders for which a death sentence may ever be imposed serves much the same purpose." Id., at 270. The joint opinion went on to say that because the constitutionality of a capital sentencing system also requires the sentencing authority to consider mitigating circumstances and since the Texas statute did not speak of mitigating circumstances and instead directs only that the jury answer three questions, "the constitutionality of the Texas procedures turns on whether the enumerated questions allow consideration of particularized mitigating factors." Id., at 272.

The joint opinion then recognized that the Texas Court of Criminal Appeals had held:

"`In determining the likelihood that the defendant would be a continuing threat to society, the jury could consider whether the defendant had a significant criminal record. It could consider the range and severity of his prior criminal conduct. It could further look to the age of the defendant and whether or not at the time of the commission of the offense he was acting under duress or under the domination of another. It could also consider whether the defendant was under an extreme form of mental or emotional pressure, something less, perhaps, than insanity, but more than the emotions of the average man, however inflamed, could withstand.' 522 S.W.2d, at 939-940." Id., at 272-273.

Based on this assurance, the opinion characterized the Texas sentencing procedure as follows:

"Thus, Texas law essentially requires that one of five aggravating circumstances be found before a defendant can be found guilty of capital murder, and that in considering whether to impose a death sentence the jury may be asked to consider whatever evidence of mitigating circumstances the defense can bring before it. It thus appears that, as in Georgia and Florida, the Texas capital-sentencing procedure guides and focuses the jury's objective consideration of the particularized circumstances of the individual offense and the individual offender before it can impose a sentence of death." Id., at 273-274. "What is essential is that the jury have before it all possible relevant information about the individual defendant whose fate it must determine. Texas law clearly assures that all such evidence will be adduced." Id., at 276.

The joint opinion's ultimate conclusion was:

"Texas' capital-sentencing procedures, like those of Georgia and Florida, do not violate the Eighth and Fourteenth Amendments. By narrowing its definition of capital murder, Texas has essentially said that there must be at least one statutory aggravating circumstance in a first-degree murder case before a death sentence may even be considered. By authorizing the defense to bring before the jury at the separate sentencing hearing whatever mitigating circumstances relating to the individual defendant can be adduced, Texas has ensured that the sentencing jury will have adequate guidance to enable it to perform its sentencing function. By providing prompt judicial review of the jury's decision in a court with statewide jurisdiction, Texas has provided a means to promote the evenhanded, rational, and consistent imposition of death sentences under law. Because this system serves to assure that sentences of death will not be `wantonly' or `freakishly' imposed, it does not violate the Constitution. Furman v. Georgia, 408 U. S., at 310 (Stewart, J., concurring)." Ibid.

It is plain enough, we think, that the joint opinion could reasonably be read as having arrived at this conclusion only after being satisfied that the mitigating evidence introduced by the defendant, including his age, would be given constitutionally adequate consideration in the course of the jury's deliberation on the three special issues. Three other Justices concurred in the holding that the Texas procedures for imposing the death penalty were constitutional. Id., at 278— 279 (White, J., concurring in judgment).

Two years after Jurek, in another splintered decision, Lockett v. Ohio, 438 U.S. 586 (1978), the Court invalidated an Ohio death penalty statute that prevented the sentencer from considering certain categories of relevant mitigating evidence. In doing so, a plurality of the Court consisting of Chief Justice Burger and Justices Stewart, Powell, and Stevens stated that the constitutional infirmities in the Ohio statute could "best be understood by comparing it with the statutes upheld in Gregg, Proffitt, and Jurek. " Id., at 606. This the plurality proceeded to do, recounting in the process that the Texas statute had been held constitutional in Jurek because it permitted the sentencer to consider whatever mitigating circumstances the defendant could show. Emphasizing that "an individualized [sentencing] decision is essential in capital cases," the plurality concluded:

"There is no perfect procedure for deciding in which cases governmental authority should be used to impose death. But a statute that prevents the sentencer in all capital cases from giving independent mitigating weight to aspects of the defendant's character and record and to circumstances of the offense proffered in mitigation creates the risk that the death penalty will be imposed in spite of factors that may call for a less severe penalty." 438 U.S., at 605.

Obviously, the plurality did not believe the Texas statute suffered this infirmity.

The plurality's rule was embraced by a majority of the Court four years later in Eddings v. Oklahoma, 455 U.S. 104 (1982). There, the Court overturned a death sentence on the ground that the judge who entered it had felt himself bound by state law to disregard mitigating evidence concerning the defendant's troubled youth and emotional disturbance. The Court held that, "[j]ust as the State may not by statute preclude the sentencer from considering any mitigating factor, neither may the sentencer refuse to consider, as a matter of law, any relevant mitigating evidence." Id., at 113-114 (emphasis omitted); see also Hitchcock v. Dugger, 481 U.S. 393, 394 (1987); Skipper v. South Carolina, 476 U.S. 1, 4-5 (1986). The Eddings opinion rested on Lockett and made no mention of Jurek.

We cannot say that reasonable jurists considering petitioner's claim in 1984 would have felt that these cases "dictated " vacatur of petitioner's death sentence. See Teague, 489 U. S., at 301. To the contrary, to most readers at least, these cases reasonably would have been read as upholding the constitutional validity of Texas' capital sentencing scheme with respect to mitigating evidence and otherwise. Lockett expressly embraced the Jurek holding, and Eddings signaled no retreat from that conclusion. It seems to us that reasonable jurists in 1984 would have found that, under our cases, the Texas statute satisfied the commands of the Eighth Amendment: It permitted petitioner to place before the jury whatever mitigating evidence he could show, including his age, while focusing the jury's attention upon what that evidence revealed about the defendant's capacity for deliberation and prospects for rehabilitation.

We find nothing in our more recent cases, to the extent they are relevant, that would undermine this analysis. In 1988, in Franklin v. Lynaugh, 487 U.S. 164, we rejected a claim that the Texas special issues provided an inadequate vehicle for jury consideration of evidence of a defendant's clean prison disciplinary record. There, a plurality of the Court observed that "[i]n resolving the second Texas Special Issue, the jury was surely free to weigh and evaluate petitioner's disciplinary record as it bore on his `character'—that is, his `character' as measured by his likely future behavior." Id., at 178. Moreover, the plurality found

"unavailing petitioner's reliance on this Court's statement in Eddings, 455 U. S., at 114, that the sentencing jury may not be precluded from considering `any relevant, mitigating evidence.' This statement leaves unanswered the question: relevant to what? While Lockett, supra, at 604, answers this question at least in part—making it clear that a State cannot take out of the realm of relevant sentencing considerations the questions of the defendant's `character,' `record,' or the `circumstances of the offense'—Lockett does not hold that the State has no role in structuring or giving shape to the jury's consideration of these mitigating factors." Id., at 179 (citations omitted).

To be sure, Justice O'Connor's opinion concurring in the judgment in Franklin expressed "doubts" about the validity of the Texas death penalty statute as that statute might be applied in future cases. Id., at 183. The Justice agreed, however, that the special issues adequately accounted for the mitigating evidence presented in that case. Ibid.

This brings us to Penry v. Lynaugh, 492 U.S. 302 (1989), upon which petitioner chiefly relies. In that case, the Court overturned a prisoner's death sentence, finding that the Texas special issues provided no genuine opportunity for the jury to give mitigating effect to evidence of his mental retardation and abused childhood. The Court considered these factors to be mitigating because they diminished the defendant's ability "to control his impulses or to evaluate the consequences of his conduct," and therefore reduced his moral culpability. Id., at 322. The Texas special issues permitted the jury to consider this evidence, but not necessarily in a way that would benefit the defendant. Although Penry's evidence of mental impairment and childhood abuse indeed had relevance to the "future dangerousness" inquiry, its relevance was aggravating only. "Penry's mental retardation and history of abuse is thus a two-edged sword: it may diminish his blameworthiness for his crime even as it indicates that there is a probability that he will be dangerous in the future." Id., at 324. Whatever relevance Penry's evidence may have had to the other two special issues was too tenuous to overcome this aggravating potential. Because it was impossible to give meaningful mitigating effect to Penry's evidence by way of answering the special issues, the Court concluded that Penry was constitutionally entitled to further instructions "informing the jury that it could consider and give effect to [Penry's] evidence . .. by declining to impose the death penalty." Id., at 328.

We do not read Penry as effecting a sea change in this Court's view of the constitutionality of the former Texas death penalty statute; it does not broadly suggest the invalidity of the special issues framework.^[3] Indeed, any such reading of Penry would be inconsistent with the Court's conclusion in that case that it was not announcing a "new rule" within the meaning of Teague v. Lane, 489 U.S. 288 (1989). See Penry, supra, at 318-319. As we have explained in subsequent cases:

"To the extent that Penry's claim was that the Texas system prevented the jury from giving any mitigating effect to the evidence of his mental retardation and abuse in childhood, the decision that the claim did not require the creation of a new rule is not surprising. Lock- ett and Eddings command that the State must allow the jury to give effect to mitigating evidence in making the sentencing decision; Penry's contention was that Texas barred the jury from so acting. Here, by contrast, there is no contention that the State altogether prevented Parks' jury from considering, weighing, and giving effect to all of the mitigating evidence that Parks put before them; rather, Parks' contention is that the State has unconstitutionally limited the manner in which his mitigating evidence may be considered. As we have concluded above, the former contention would come under the rule of Lockett and Eddings; the latter does not." Saffle v. Parks, 494 U. S., at 491.

In our view, the rule that Graham seeks is not commanded by the cases upon which Penry rested. In those cases, the constitutional defect lay in the fact that relevant mitigating evidence was placed beyond the effective reach of the sentencer. In Lockett, Eddings, Skipper, and Hitchcock, the sentencer was precluded from even considering certain types of mitigating evidence. In Penry, the defendant's evidence was placed before the sentencer but the sentencer had no reliable means of giving mitigating effect to that evidence. In this case, however, Graham's mitigating evidence was not placed beyond the jury's effective reach. Graham indisputably was permitted to place all of his evidence before the jury and both of Graham's two defense lawyers vigorously urged the jury to answer "no" to the special issues based on this evidence. Most important, the jury plainly could have done so consistent with its instructions. The jury was not forbidden to accept the suggestion of Graham's lawyers that his brief spasm of criminal activity in May 1981 was properly viewed, in light of his youth, his background, and his character, as an aberration that was not likely to be repeated. Even if Graham's evidence, like Penry's, had significance beyond the scope of the first special issue, it is apparent that Graham's evidence—unlike Penry's—had mitigating relevance to the second special issue concerning his likely future dangerousness. Whereas Penry's evidence compelled an affirmative answer to that inquiry, despite its mitigating significance, Graham's evidence quite readily could have supported a negative answer. This distinction leads us to conclude that neither Penry nor any of its predecessors "dictates " the relief Graham seeks within the meaning required by Teague. See Stringer v. Black, 503 U. S., at 238 (Souter, J., dissenting): "The result in a given case is not dictated by precedent if it is `susceptible to debate among reasonable minds,' or, put differently, if `reasonable jurists may disagree' " (citations omitted).

Moreover, we are not convinced that Penry could be extended to cover the sorts of mitigating evidence Graham suggests without a wholesale abandonment of Jurek and perhaps also of Franklin v. Lynaugh. As we have noted, Jurek is reasonably read as holding that the circumstance of youth is given constitutionally adequate consideration in deciding the special issues. We see no reason to regard the circumstances of Graham's family background and positive character traits in a different light. Graham's evidence of transient upbringing and otherwise nonviolent character more closely resembles Jurek's evidence of age, employment history, and familial ties than it does Penry's evidence of mental retardation and harsh physical abuse. As the dissent in Franklin made clear, virtually any mitigating evidence is capable of being viewed as having some bearing on the defendant's "moral culpability" apart from its relevance to the particular concerns embodied in the Texas special issues. See Franklin, 487 U. S., at 190 (Stevens, J., dissenting). It seems to us, however, that reading Penry as petitioner urges—and thereby holding that a defendant is entitled to special instructions whenever he can offer mitigating evidence that has some arguable relevance beyond the special issues—would be to require in all cases that a fourth "special issue" be put to the jury: "`Does any mitigating evidence before you, whether or not relevant to the above [three] questions, lead you to believe that the death penalty should not be imposed?' " The Franklin plurality rejected precisely this contention, finding it irreconcilable with the Court's holding in Jurek, see Franklin, supra, at 180, n. 10, and we affirm that conclusion today. Accepting Graham's submission would unmistakably result in a new rule under Teague. See Saffle v. Parks, supra, at 488; Butler v. McKellar, 494 U. S., at 412.

In sum, even if Penry reasonably could be read to suggest that Graham's mitigating evidence was not adequately considered under the former Texas procedures, that is not the relevant inquiry under Teague. Rather, the determinative question is whether reasonable jurists reading the case law that existed in 1984 could have concluded that Graham's sentencing was not constitutionally infirm. We cannot say that all reasonable jurists would have deemed themselves compelled to accept Graham's claim in 1984. Nor can we say, even with the benefit of the Court's subsequent decision in Penry, that reasonable jurists would be of one mind in ruling on Graham's claim today. The ruling Graham seeks, therefore, would be a "new rule" under Teague.

2

Having decided that the relief Graham seeks would require announcement of a new rule under Teague, we next consider whether that rule nonetheless would fall within one of the two exceptions recognized in Teague to the "new rule" principle. "The first exception permits the retroactive application of a new rule if the rule places a class of private conduct beyond the power of the State to proscribe, see Teague, 489 U. S., at 311, or addresses a `substantive categorical guarante[e] accorded by the Constitution,' such as a rule `prohibiting a certain category of punishment for a class of defendants because of their status or offense.' " Saffle v. Parks, supra, at 494 (quoting Penry, 492 U. S., at 329, 330). Plainly, this exception has no application here because the rule Graham seeks "would neither decriminalize a class of conduct nor prohibit the imposition of capital punishment on a particular class of persons." 494 U.S., at 495.

The second exception permits federal courts on collateral review to announce "`watershed rules of criminal procedure' implicating the fundamental fairness and accuracy of the criminal proceeding." Ibid. Whatever the precise scope of this exception, it is clearly meant to apply only to a small core of rules requiring "observance of `those procedures that .. . are "implicit in the concept of ordered liberty."` " Teague, supra, at 311 (quoting Mackey v. United States, 401 U.S. 667, 693 (1971) (Harlan, J., concurring in judgments in part and dissenting in part) (in turn quoting Palko v. Connecticut, 302 U.S. 319, 325 (1937))); see also Butler v. McKellar, supra, at 416. As the plurality cautioned in Teague, "[b]ecause we operate from the premise that such procedures would be so central to an accurate determination of innocence or guilt, we believe it unlikely that many such components of basic due process have yet to emerge." 489 U.S., at 313. We do not believe that denying Graham special jury instructions concerning his mitigating evidence of youth, family background, and positive character traits "seriously diminish[ed] the likelihood of obtaining an accurate determination" in his sentencing proceeding. See Butler v. McKellar, supra, at 416. Accordingly, we find the second Teague exception to be inapplicable as well.

The judgment of the Court of Appeals is therefore

Affirmed.

Justice Thomas, concurring.

By deciding this case on the basis of Teague v. Lane, 489 U.S. 288 (1989), the Court has avoided a direct reconsideration of Penry v. Lynaugh, 492 U.S. 302 (1989). I join the Court's opinion because I agree that the holding sought by Graham is not compelled by the cases upon which Penry rests and would therefore, if adopted, be a new rule for Teague purposes. I write separately, however, to make clear that I believe Penry was wrongly decided.

Several Members of the Court have commented on the "tension" between our cases on the constitutional relevance of mitigating circumstances in capital sentencing and those decisions applying the principle, first articulated in Furman v. Georgia, 408 U.S. 238 (1972), that the Eighth and Fourteenth Amendments prohibit States from giving sentencers unguided discretion in imposing the death penalty. E. g., Franklin v. Lynaugh, 487 U.S. 164, 182 (1988) (plurality opinion); California v. Brown, 479 U.S. 538, 544 (1987) (O'Connor, J., concurring); McCleskey v. Kemp, 481 U.S. 279, 363 (1987) (Blackmun, J., dissenting). In my view, Texas had largely resolved this tension through the use of the three special issues repeatedly approved by this Court. Penry, however, is at war with the former Texas scheme. As the most extreme statement in our "mitigating" line, Penry creates more than an unavoidable tension; it presents an evident danger.

I

A

It is important to recall what motivated Members of this Court at the genesis of our modern capital punishment case law. Furman v. Georgia was decided in an atmosphere suffused with concern about race bias in the administration of the death penalty—particularly in Southern States, and most particularly in rape cases. The three petitioners were black.^[1] Lucious Jackson was a 21-year-old black man sentenced to death by Georgia for raping a white woman. Elmer Branch was sentenced to death by Texas for the rape of a 65-year-old white widow. William Henry Furman faced the death penalty in Georgia for unintentionally killing a white homeowner during a burglary. See 408 U.S., at 252— 253 (Douglas, J., concurring).^[2] In his opinion concurring in the Court's judgment that the death penalty in these cases was unconstitutional, Justice Douglas stressed the potential role of racial and other illegitimate prejudices in a system where sentencing juries have boundless discretion. He thought it cruel and unusual to apply the death penalty "selectively to minorities . . . whom society is willing to see suffer though it would not countenance general application of the same penalty across the board." Id., at 245. Citing studies and reports suggesting that "[t]he death sentence [was] disproportionately imposed and carried out on the poor, the Negro, and the members of unpopular groups," especially in cases of rape, id., at 249-250 (internal quotation marks omitted), Justice Douglas concluded that

"the discretion of judges and juries in imposing the death penalty enables the penalty to be selectively applied, feeding prejudices against the accused if he is poor and despised, and lacking political clout, or if he is a member of a suspect or unpopular minority, and saving those who by social position may be in a more protected position." Id., at 255.

Justice Marshall echoed these concerns. See id., at 364— 366 (concurring opinion). He wrote that "[r]acial or other discriminations [in sentencing] should not be surprising," because, in his view, the Court's earlier decision in McGautha v. California, 402 U.S. 183 (1971), upholding a procedure that had "committ[ed] to the untrammeled discretion of the jury the power to pronounce life or death," id., at 207, was "an open invitation to discrimination," 408 U.S., at 365. Justice Stewart also agreed that "if any basis can be discerned for the selection of these few to be sentenced to die, it is the constitutionally impermissible basis of race." Id., at 310 (concurring opinion).

The unquestionable importance of race in Furman is reflected in the fact that three of the original four petitioners in the Furman cases were represented by the NAACP Legal Defense and Educational Fund, Inc. This representation was part of a concerted "national litigative campaign against the constitutionality of the death penalty" waged by a small number of ambitious lawyers and academics on the Fund's behalf. Burt, Disorder in the Court: The Death Penalty and the Constitution, 85 Mich. L. Rev. 1741, 1745 (1987). Although their efforts began rather modestly, assisting indigent black defendants in isolated criminal cases—usually rape cases—where racial discrimination was suspected, the lawyers at the Fund ultimately devised and implemented (not without some prompting from this Court) an all-out strategy of litigation against the death penalty. See generally M. Meltsner, Cruel and Unusual: The Supreme Court and Capital Punishment (1973) (hereinafter Meltsner); Muller, The Legal Defense Fund's Capital Punishment Campaign: The Distorting Influence of Death, 4 Yale L. & Pol'y Rev. 158 (1985).^[3] This campaign was part of a larger movement carried on in the 1960's by "abolitionist lawyers" whose agenda for social and legal change depended on an activist judiciary; their "unmistakable preference for the courts, especially the federal courts," came as a direct "response to the Supreme Court's willingness to redraw America's ethical and legal map, a task state houses and executive mansions were slow to tackle." Meltsner 25, 71.^[4]

In mustering every conceivable argument—"ethical, legal, polemical, theological, speculative, [and] statistical"—for abolishing capital punishment, id., at 59, the Fund lawyers and other civil rights advocates supplied the empirical and rhetorical support for the observations of Justices Douglas, Marshall, and Stewart with respect to race bias. See Brief for Petitioner in Aikens v. California, O. T. 1971, No. 68— 5027, pp. 50-54; Brief for Petitioner in Jackson v. Georgia, O. T. 1971, No. 69-5030, p. 15 ("The racial figures for all men executed in the United States for the crime of rape since 1930 are as follows: 48 white, 405 Negro, 2 other. In Georgia, the figures are: 3 white, 58 Negro") (footnotes omitted). See also Brief for NAACP et al. as Amici Curiae in Aikens v. California, supra, at 13-18, and App. A (discussing, in particular, history of South's use of death penalty in rape cases prior to Civil War, when it was typical for rapes or attempted rapes committed by black men upon white women to be punishable by mandatory death or castration, while rapes committed by whites were not punishable by death); Brief for Synagogue Council of America et al. as Amici Curiae in Aikens v. California, supra, at 31 ("The positive relationship between the death penalty and race is strong, but where the crime involved is rape and more particularly, as in two of the present cases, the rape of white women by Negroes, the relationship is almost uncontrovertible").^[5]

In the end, Justice Douglas and the other Members of the Court concluded that "[w]e cannot say from facts disclosed in these records that these defendants were sentenced to death because they were black." Furman, 408 U. S., at 253 (Douglas, J., concurring). See id., at 310 (Stewart, J., concurring) ("racial discrimination has not been proved"). The Court focused more generally on the uncontrolled discretion placed in judges and juries. Such unbridled discretion, it was argued, practically invited sentencers to vent their personal prejudices in deciding the fate of the accused. See Brief for Petitioner in Furman v. Georgia, O. T. 1971, No. 69-5003, p. 12 ("The jury knew nothing else about the man they sentenced, except his age and race"). "Under these laws no standards govern the selection of the penalty. People live or die, dependent on the whim of one man or of 12." 408 U.S., at 253 (Douglas, J., concurring). Justice Stewart observed that "the petitioners are among a capriciously selected random handful upon whom the sentence of death has in fact been imposed," and concluded that the Eighth and Fourteenth Amendments cannot tolerate sentencing procedures that allow the penalty to be "so wantonly and so freakishly" inflicted. Id., at 309-310 (concurring opinion). The practice of delegating unguided authority—a practice "largely motivated by the desire to mitigate the harshness of the law and to bring community judgment to bear on the sentence"—actually allowed a jury, "in its own discretion and without violating its trust or any statutory policy, [to] refuse to impose the death penalty no matter what the circumstances of the crime." Id., at 313, 314 (White, J., concurring).

In sum, the Court concluded that in a standardless sentencing scheme there was no "rational basis," as Justice Brennan put it, to distinguish "the few who die from the many who go to prison." Id., at 294 (concurring opinion). See also id., at 313 (White, J., concurring) ("no meaningful basis for distinguishing"). It cannot be doubted that behind the Court's condemnation of unguided discretion lay the specter of racial prejudice—the paradigmatic capricious and irrational sentencing factor.

B

At its inception, our "mitigating" line of cases sprang in part from the same concerns that underlay Furman. In response to Furman, 35 States enacted new death penalty statutes. See Gregg v. Georgia, 428 U.S. 153, 179-180 (1976) (joint opinion of Stewart, Powell, and Stevens, JJ.). In five cases decided on a single day in 1976, we passed on the constitutionality of a representative sample of the new laws.^[6] The principal opinion in each case was a joint opinion of Justices Stewart, Powell, and Stevens. In the lead case, Gregg v. Georgia, these Justices squarely rejected the argument that the death penalty is cruel and unusual under all circumstances. Id., at 176-187. Rather, they focused on the States' capital sentencing procedures, distilling from Furman two complementary rationalizing principles about sentencing discretion: The discretion given the sentencer must be "directed and limited" to avoid "wholly arbitrary and capricious action," Gregg, 428 U. S., at 189, and this discretion must be exercised "in an informed manner," ibid. Furman was read as holding that "to minimize the risk that the death penalty [will] be imposed on a capriciously selected group of offenders, the decision to impose it ha[s] to be guided by standards so that the sentencing authority [will] focus on the particularized circumstances of the crime and the defendant." Gregg, 428 U. S., at 199. The jury should be "given guidance regarding the factors about the crime and the defendant that the State, representing organized society, deems particularly relevant to the sentencing decision." Id., at 192. "Otherwise, the system cannot function in a consistent and a rational manner." Id., at 189 (internal quotation marks omitted).

Gregg `s requirement that the sentencer be guided by information about the particular defendant and the particular circumstances of the crime—in other words, by traditionally accepted sentencing criteria, see id., at 189-190—added a second dimension to Furman `s rule against open-ended discretion. The jury's discretion must be focused on rational factors, and its decision should be based on information about the circumstances of the crime and about the accused as an individual, not merely as a member of a group. In Furman itself, for example, the jury was given almost no particularized information about the accused: "About Furman himself, the jury knew only that he was black and that, according to his statement at trial, he was 26 years old and worked at `Superior Upholstery.' It took the jury one hour and 35 minutes to return a verdict of guilt and a sentence of death." Furman, 408 U. S., at 295, n. 48 (Brennan, J., concurring) (citations omitted). Moreover, it was irrelevant to the jury's determination that the killing committed by Furman was accidental. Ibid. Without a focus on the characteristics of the defendant and the circumstances of his crime, an uninformed jury could be tempted to resort to irrational considerations, such as class or race animus.

Justices Stewart, Powell, and Stevens applied these principles in upholding the guided discretion procedures of Georgia, Florida, and Texas, and in striking down the mandatory death penalty provisions of North Carolina and Louisiana. The Georgia, Florida, and Texas schemes were held constitutional because they "guide[d] and focuse[d] the jury's objective consideration of the particularized circumstances of the individual offense and the individual offender." Jurek v. Texas, 428 U.S. 262, 273-274 (1976) (joint opinion of Stewart, Powell, and Stevens, JJ.). The "essential" factor was that "the jury ha[d] before it all possible relevant information about the individual defendant whose fate it must determine." Id., at 276. Moreover, the Georgia statute featured "an important additional safeguard against arbitrariness and caprice": a provision for automatic appeal of a death sentence that required the State Supreme Court to determine, inter alia, whether the sentence was imposed under the influence of passion or prejudice and whether it was disproportionate to other sentences imposed in similar cases. Gregg, supra, at 198.

The mandatory death penalty statutes, on the other hand, were held to violate the Eighth and Fourteenth Amendments for three reasons. First, the Justices believed, a mandatory death penalty departed from "contemporary standards" of punishment. Woodson v. North Carolina, 428 U.S. 280, 301 (1976) (plurality opinion). Second, experience had suggested that such statutes "simply papered over the problem of unguided and unchecked jury discretion" by provoking arbitrary jury nullification. Id., at 302-303. Thus, "[i]nstead of rationalizing the sentencing process, a mandatory scheme may well exacerbate the problem identified in Furman by resting the penalty determination on the particular jury's willingness to act lawlessly." Id., at 303; see Roberts v. Louisiana, 428 U.S. 325, 335 (1976) (plurality opinion). Third, the mandatory nature of the penalty prevented the sentencer from considering "the character and record of the individual offender or the circumstances of the particular offense," and thus treated all convicted persons "not as uniquely individual human beings, but as members of a faceless, undifferentiated mass." Woodson, supra, at 304. The latter concern echoed Justice Douglas' suggestion that sentences of death might have fallen disproportionately upon the "member[s] of a suspect or unpopular minority." Furman, supra, at 255.

One would think, however, that by eliminating explicit jury discretion and treating all defendants equally, a mandatory death penalty scheme was a perfectly reasonable legislative response to the concerns expressed in Furman. See Roberts, supra, at 346 (White, J., dissenting). See also Walton v. Arizona, 497 U.S. 639, 662 (1990) (Scalia, J., concurring in part and concurring in judgment). Justice White was surely correct in concluding that "a State is not constitutionally forbidden to provide that the commission of certain crimes conclusively establishes that the criminal's character is such that he deserves death." Roberts, supra, at 358. See also Roberts v. Louisiana, 431 U.S. 633, 649 (1977) (Rehnquist, J., dissenting); Sumner v. Shuman, 483 U.S. 66, 86 (1987) (White, J., dissenting). I would also agree that the plurality in Woodson and Roberts erred in equating the "raw power of [ jury] nullification" with the unlimited sentencing discretion condemned in Furman. Roberts, supra, at 347 (White, J., dissenting). The curious and counterintuitive outcomes of our 1976 cases—upholding sentences of death imposed under statutes that explicitly preserved the sentencer's discretion while vacating those imposed under mandatory provisions precisely because of a perceived potential for arbitrary and uninformed discretion—might in some measure be attributable, once again, to the powerful influence of racial concerns.^[7] Be that as it may, we are not now confronted with a mandatory sentencing provision, and I have no occasion here to flesh out my disagreement with the Court's prohibition of such schemes.

The significant point for present purposes is that Woodson and Sumner `s invalidation of the mandatory death penalty guaranteed that sentencers would exercise some degree of discretion in every capital case. And under our precedents, in turn, any such exercise of discretion is unavoidably bound up with the two requirements of Furman, as identified in Gregg: first and foremost, that the sentencing authority be "provided with standards to guide its use of the information" developed at sentencing, and second, in support of this principle, that the sentencer be "apprised of the information relevant to the imposition of sentence." Gregg, 428 U. S., at 195. By discovering these two requirements in the Constitution, and by ensuring in Woodson and its progeny that they would always be in play, the Court has put itself in the seemingly permanent business of supervising capital sentencing procedures. While the better view is that the Cruel and Unusual Punishments Clause was intended to place only substantive limitations on punishments, not procedural requirements on sentencing, see Hudson v. McMillian, 503 U.S. 1, 18-20 (1992) (Thomas, J., dissenting); Gardner v. Florida, 430 U.S. 349, 371 (1977) (Rehnquist, J.,dissenting), stare decisis requires that we make efforts to adhere to the Court's Eighth Amendment precedents, see Walton v. Arizona, supra, at 672 (Scalia, J., concurring in part and concurring in judgment).

The mitigating branch of our death penalty jurisprudence began as an outgrowth of the second of the two Furman / Gregg requirements. The plurality's conclusion in Lockett v. Ohio, 438 U.S. 586 (1978)—that the sentencer in a capital case must "not be precluded from considering, as a mitigating factor, any aspect of a defendant's character or record and any of the circumstances of the offense," id., at 604 (opinion of Burger, C. J.) (emphasis deleted)—effectively guarantees the sentencer's access to categories of information favorable to the defendant. Thus, Lockett was built on the premise, given credence in Gregg, that "where sentencing discretion is granted, it generally has been agreed that the sentencing judge's possession of the fullest information possible concerning the defendant's life and characteristics is [h]ighly relevant." 438 U.S., at 602-603 (internal quotation marks omitted). The sentencing statute at issue in Lockett failed to satisfy this requirement, in the plurality's view, because it eliminated from the jury's consideration significant facts about the defendant and her "comparatively minor role in the offense." Id., at 608.^[8] The Court's adoption in Eddings v. Oklahoma, 455 U.S. 104 (1982), of the Lockett rule and its corollary—that the sentencer may not categorically refuse to consider relevant mitigating circumstances—again drew upon Gregg `s notion that capital sentencing is less likely to be arbitrary where the jury's exercise of discretion is focused on the particularized circumstances of the offender and the crime. See Eddings, supra, at 112 (relying on Gregg, supra, at 197).

Therefore, although it is said that Lockett and Eddings represent an "about-face" and "a return to the pre-Furman days," Lockett, supra, at 622, 623 (White, J., concurring in part, dissenting in part, and concurring in judgments), there was at root a logical—if by now attenuated—connection between the rationalizing principle of Furman and the prophylactic rule of Eddings. Eddings protects the accused's opportunity to "appris[e]" the jury of his version of the information relevant to the sentencing decision. Our early mitigating cases may thus be read as doing little more than safeguarding the adversary process in sentencing proceedings by conferring on the defendant an affirmative right to place his relevant evidence before the sentencer. See Skipper v. South Carolina, 476 U.S. 1, 4 (1986). Cf. id., at 5, n. 1 (comparing Eddings with "the elemental due process requirement that a defendant not be sentenced to death `on the basis of information which he had no opportunity to deny or explain.' Gardner v. Florida, 430 U.S. 349, 362 (1977)").

Consistent with this (admittedly narrow) reading, I would describe Eddings as a kind of rule of evidence: It governs the admissibility of proffered evidence but does not purport to define the substantive standards or criteria that sentencers are to apply in considering the facts. By requiring that sentencers be allowed to "consider" all "relevant" mitigating circumstances, we cannot mean that the decision whether to impose the death penalty must be based upon all of the defendant's evidence, or that such evidence must be considered the way the defendant wishes. Nor can we mean to say that circumstances are necessarily relevant for constitutional purposes if they have any conceivable mitigating value. Such an application of Eddings would eclipse the primary imperative of Furman —that the State define the relevant sentencing criteria and provide rational "standards to guide [the sentencer's] use" of the evidence. That aspect of Furman must operate for the most part independently of the Eddings rule. This is essential to the effectiveness of Furman, since providing all relevant information for the sentencer's consideration does nothing to avoid the central danger that sentencing discretion may be exercised irrationally.

I realize, of course, that Eddings is susceptible to more expansive interpretations. See, e. g., Walton, 497 U. S., at 661, 667 (Scalia, J., concurring in part and concurring in judgment) (Eddings rule "has completely exploded whatever coherence the notion of `guided discretion' once had" by making "random mitigation" a constitutional requirement); McCleskey v. Kemp, 481 U. S., at 306 ("States cannot limit the sentencer's consideration of any relevant circumstance that could cause it to decline to impose the [death] penalty. In this respect, the State cannot channel the sentencer's discretion, but must allow it to consider any relevant information offered by the defendant"). And even under the narrow reading of Eddings, there is still a tension in our case law, because Eddings implies something of an outer boundary to the primary Furman principle: The sentencing standards chosen by the State may not be so stingy as to prevent altogether the consideration of constitutionally relevant mitigating evidence.

But with the exception of Penry v. Lynaugh, 492 U.S. 302 (1989), our most recent mitigating cases have been careful to read Eddings narrowly in an effort to accommodate the "competing commandments" of Eddings and Furman, ante, at 468. We have held that States must be free to channel and direct the sentencer's consideration of all evidence (whether mitigating or aggravating) that bears on sentencing, provided only that the State does not categorically preclude the sentencer from considering constitutionally relevant mitigating circumstances. See Walton, supra, at 652 ("[T]here is no . . . constitutional requirement of unfettered sentencing discretion in the jury, and States are free to structure and shape consideration of mitigating evidence in an effort to achieve a more rational and equitable administration of the death penalty") (internal quotation marks omitted); Boyde v.California, 494 U.S. 370, 377 (1990) (to the same effect); Franklin v. Lynaugh, 487 U. S.,at 181 (plurality opinion) (same); see also Walton, supra, at 652 (requirement of individualized sentencing in capital cases satisfied as long as State does not altogether prevent sentencer from considering any type of relevant mitigating evidence); Blystone v. Pennsylvania, 494 U.S. 299, 307-308 (1990) (same); Saffle v. Parks, 494 U.S. 484, 490-491 (1990) (same).

This understanding preserves our original rationale for upholding the Texas sentencing statute—that it "guides and focuses the jury's objective consideration of the particularized circumstances" while allowing the defendant "to bring to the jury's attention whatever [relevant] mitigating circumstances he may be able to show." Jurek, 428 U. S., at 272, 274. Thus, in reaffirming the constitutionality of Texas' system of special issues, we have expressed satisfaction that the former Texas scheme successfully reconciled any tension that exists between Eddings and Furman. See Franklin v. Lynaugh, supra, at 182 (plurality opinion). In the context of the Texas system, therefore, I am unprepared at present to sweep away our entire mitigating line of precedent. By the same token, however, if the more expansive reading of Eddings were ultimately to prevail in this Court, I would be forced to conclude that the Eddings rule, as so construed, truly is "rationally irreconcilable with Furman " and, on that basis, deserving of rejection. See Walton, supra, at 673 (Scalia, J., concurring in part and concurring in judgment).

II

Unfortunately, the narrow reading of Eddings is virtually impossible after Penry. Whatever contribution to rationality and consistency we made in Furman, we have taken back with Penry. In the process, we have upset the careful balance that Texas had achieved through the use of its special issues.

Penry held that the Texas special issues did not allow a jury to "consider and give effect to" mitigating evidence of mental retardation and childhood abuse, 492 U.S., at 328, because, even though the defendant had a full and unfettered opportunity to present such evidence to the jury, the evidence had "relevance to [Penry's] moral culpability beyond the scope of the special issues," id., at 322 (emphasis added). Thus, the Court was persuaded that the jury might have been "unable to express its `reasoned moral response ` to that evidence in determining whether death was the appropriate punishment." Ibid. (emphasis added). See post, at 518— 519. Contrary to the dissent's view, see post, at 506-512, these notions—that a defendant may not be sentenced to death if there are mitigating circumstances whose relevance goes "beyond the scope" of the State's sentencing criteria, and that the jury must be able to express a "reasoned moral response" to all evidence presented—have no pedigree in our prior holdings. They originated entirely from whole cloth in two recent concurring opinions. See Franklin, supra, at 185 (O'Connor, J., concurring in judgment); California v. Brown, 479 U.S. 538, 545 (1987) (O'Connor, J., concurring).

Together, these notions render meaningless any rational standards by which a State may channel or focus the jury's discretion and thus negate the central tenet of Furman and all our death penalty cases since 1972. Penry imposes as a constitutional imperative "a scheme that simply dumps before the jury all sympathetic factors bearing upon the defendant's background and character, and the circumstances of the offense, so that the jury may decide without further guidance" whether the defendant deserves death. Penry, 492 U. S., at 359 (Scalia, J., concurring in part and dissenting in part). "It is an unguided, emotional `moral response' that the Court demands be allowed—an outpouring of personal reaction to all the circumstances of a defendant's life and personality, an unfocused sympathy." Ibid. Justice Souter's reading of Penry bears out these fears. His dissent would require that the special issues be "construed with enough scope to allow the full consideration of mitigating potential," post, at 515, and that the jury be free to give full effect to the defendant's sympathetic evidence "for all purposes, including purposes not specifically permitted by the questions," post, at 511 (internal quotation marks and emphasis omitted).

Any determination that death is or is not the fitting punishment for a particular crime will necessarily be a moral one, whether made by a jury, a judge, or a legislature. But beware the word "moral" when used in an opinion of this Court. This word is a vessel of nearly infinite capacity— just as it may allow the sentencer to express benevolence, it may allow him to cloak latent animus. A judgment that some will consider a "moral response" may secretly be based on caprice or even outright prejudice. When our review of death penalty procedures turns on whether jurors can give "full mitigating effect" to the defendant's background and character, post, at 510, and on whether juries are free to disregard the State's chosen sentencing criteria and return a verdict that a majority of this Court will label "moral," we have thrown open the back door to arbitrary and irrational sentencing. See Penry, supra, at 360 (Scalia, J., concurring in part and dissenting in part) ("The decision whether to impose the death penalty is a unitary one; unguided discretion not to impose is unguided discretion to impose as well. In holding that the jury had to be free to deem Penry's mental retardation and sad childhood relevant for whatever purpose it wished, the Court has come full circle, not only permitting but requiring what Furman once condemned").

The Court in Penry denied that its holding signaled a return to unbridled jury discretion because, it reasoned, "so long as the class of murderers subject to capital punishment is narrowed, there is no constitutional infirmity in a procedure that allows a jury to recommend mercy based on the mitigating evidence introduced by a defendant." 492 U. S., at 327 (citing Gregg, 428 U. S., at 197-199, 203 (joint opinion), and 222 (White, J., concurring in judgment)). Cf. McCleskey v. Kemp, 481 U. S., at 311 (discussing the benefits to the defendant of discretionary leniency). Thus, the dissent suggests that once the State has sufficiently narrowed the class of death-eligible murderers, the jury's discretion to select those individuals favored to live must remain effectively unbounded. See post, at 513-515, 518-519. It turns reason on its head, however, to argue that just because we have approved sentencing systems that continue to permit juries to exercise a degree of discretionary leniency, the Eighth Amendment necessarily requires that that discretion be unguided and unlimited with respect to "the class of murderers subject to capital punishment." To withhold the death penalty out of sympathy for a defendant who is a member of a favored group is no different from a decision to impose the penalty on the basis of negative bias, and it matters not how narrow the class of death-eligible defendants or crimes. Surely that is exactly what the petitioners and the Legal Defense Fund argued in Woodson and Roberts. See n. 7, supra. It is manifest that "`the power to be lenient [also] is the power to discriminate.' " McCleskey v. Kemp, supra, at 312 (quoting K. Davis, Discretionary Justice 170 (1973)). See also Roberts, 428 U. S., at 346 (White, J., dissenting) ("It is undeniable that the unfettered discretion of the jury to save the defendant from death was a major contributing factor in the developments which led us to invalidate the death penalty in Furman v. Georgia ").^[9]

We have consistently recognized that the discretion to accord mercy—even if "largely motivated by the desire to mitigate"—is indistinguishable from the discretion to impose the death penalty. Furman, 408 U. S., at 313, 314 (White, J., concurring) (condemning unguided discretion because it allows the jury to "refuse to impose the death penalty no matter what the circumstances of the crime") (emphasis added). See also Jurek, 428 U. S., at 279 (White, J., concurring in judgment) (Texas' scheme is constitutional because it "does not extend to juries discretionary power to dispense mercy"); Roberts, supra, at 335 (joint opinion) (Louisiana's statute "plainly invites" jurors to "choose a verdict for a lesser offense whenever they feel the death penalty is inappropriate"). For that reason, we have twice refused to disapprove instructions directing jurors "`not [to] be swayed by mere . . . sympathy,' " because, we have emphasized, such instructions "foste[r] the Eighth Amendment's `need for reliability in the determination that death is the appropriate punishment in a specific case.' " California v. Brown, 479 U. S., at 539, 543 (quoting Woodson, 428 U. S., at 305 (joint opinion)). Accord, Saffle v. Parks, 494 U. S., at 493 ("Whether a juror feels sympathy for a capital defendant is more likely to depend on that juror's own emotions than on the actual evidence regarding the crime and the defendant. It would be very difficult to reconcile a rule allowing the fate of a defendant to turn on the vagaries of particular jurors' emotional sensitivities with our longstanding recognition that, above all, capital sentencing must be reliable, accurate, and nonarbitrary").

Penry reintroduces the very risks that we had sought to eliminate through the simple directive that States in all events provide rational standards for capital sentencing. For 20 years, we have acknowledged the relationship between undirected jury discretion and the danger of discriminatory sentencing—a danger we have held to be inconsistent with the Eighth Amendment. When a single holding does so much violence to so many of this Court's settled precedents in an area of fundamental constitutional law, it cannot command the force of stare decisis. In my view, Penry should be overruled.^[10]

III

The major emphasis throughout our Eighth Amendment jurisprudence has been on "reasoned" rather than "moral" sentencing. We have continually sought to verify that States' capital procedures provide a "rational basis" for predictably determining which defendants shall be sentenced to death. Furman, supra, at 294 (Brennan, J., concurring). See also Spaziano v. Florida, 468 U.S. 447, 460 (1984); California v. Brown, supra, at 541; Barclay v. Florida, 463 U.S. 939, 960 (1983) (Stevens, J., concurring in judgment) ("A constant theme of our cases . . . has been emphasis on procedural protections that are intended to ensure that the death penalty will be imposed in a consistent, rational manner"); McCleskey v. Kemp, 481 U. S., at 323 (Brennan, J., dissenting) ("[C]oncern for arbitrariness focuses on the rationality of the system as a whole, and . . . a system that features a significant probability that sentencing decisions are influenced by impermissible considerations cannot be regarded as rational"). And in the absence of mandatory sentencing, States have only one means of satisfying Furman `s demands—providing objective standards to ensure that the sentencer's discretion is "guided and channeled by . . . examination of specific factors." Proffitt v. Florida, 428 U.S. 242, 258 (1976) (joint opinion of Stewart, Powell, and Stevens, JJ.).

The rule of Eddings may be an important procedural safeguard that complements Furman, but Eddings cannot promote consistency, much less rationality. Quite the opposite, as Penry demonstrates. It is imperative, therefore, that we give full effect to the standards designed by state legislatures for focusing the sentencer's deliberations. This Court has long since settled the question of the constitutionality of the death penalty. We have recognized that "capital punishment is an expression of society's moral outrage at particularly offensive conduct" and that a process for "`channeling th[e] instinct [for retribution] in the administration of criminal justice serves an important purpose in promoting the stability of a society governed by law.' " Gregg, 428 U. S., at 183 (joint opinion) (quoting Furman, supra, at 308 (Stewart, J., concurring)). If the death penalty is constitutional, States must surely be able to administer it pursuant to rational procedures that comport with the Eighth Amendment's most basic requirements.

In my view, we should enforce a permanent truce between Eddings and Furman. We need only conclude that it is consistent with the Eighth Amendment for States to channel the sentencer's consideration of a defendant's arguably mitigating evidence so as to limit the relevance of that evidence in any reasonable manner, so long as the State does not deny the defendant a full and fair opportunity to apprise the sentencer of all constitutionally relevant circumstances. The three Texas special issues easily satisfy this standard. "In providing for juries to consider all mitigating circumstances insofar as they bear upon (1) deliberateness, (2) future dangerousness, and (3) provocation, . . . Texas had adopted a rational scheme that meets the two concerns of our Eighth Amendment jurisprudence." Penry, 492 U. S., at 358-359 (Scalia, J., concurring in part and dissenting in part).

As a predicate, moreover, I believe this Court should leave it to elected state legislators, "representing organized society," to decide which factors are "particularly relevant to the sentencing decision." Gregg, supra, at 192. Although Lockett and Eddings indicate that as a general matter, "a State cannot take out of the realm of relevant sentencing considerations the questions of the defendant's `character,' `record,' or the `circumstances of the offense,' " they do "not hold that the State has no role in structuring or giving shape to the jury's consideration of these mitigating factors." Franklin v. Lynaugh, 487 U. S., at 179 (plurality opinion). Ultimately, we must come back to a recognition that "the States, and not this Court, retain `the traditional authority' to determine what particular evidence within the broad categories described in Lockett and Eddings is relevant in the first instance," Skipper v. South Carolina, 476 U. S., at 11 (Powell, J., concurring in judgment) (quoting Lockett, 438 U. S., at 604, n. 12), since "[t]his Court has no special expertise in deciding whether particular categories of evidence are too speculative or insubstantial to merit consideration by the sentencer," 476 U.S., at 15.^[11] Accordingly, I also propose that the Court's appropriate role is to review only for reasonableness a State's determinations as to which specific circumstances—within the broad bounds of the general categories mandated under Eddings —are relevant to capital sentencing.

Every month, defendants who claim a special victimization file with this Court petitions for certiorari that ask us to declare that some new class of evidence has mitigating relevance "beyond the scope" of the State's sentencing criteria. It may be evidence of voluntary intoxication or of drug use. Or even—astonishingly—evidence that the defendant suffers from chronic "antisocial personality disorder"—that is, that he is a sociopath. See Pet. for Cert. in Demouchette v. Collins, O. T. 1992, No. 92-5914, p. 4, cert. denied, 505 U.S. 1246 (1992). We cannot carry on such a business, which makes a mockery of the concerns about racial discrimination that inspired our decision in Furman.

For all these reasons, I would not disturb the effectiveness of Texas' former system.

Justice Stevens, dissenting.

Neither the race of the defendant nor the race of the victim should play a part in any decision to impose a death sentence. As Justice Thomas points out, there is reason to believe that this imperative was routinely violated in the years before the Court first held that capital punishment may violate the Eighth Amendment, when racial discrimination infected the administration of the death penalty "particularly in Southern States, and most particularly in rape cases." Ante, at 479 (concurring opinion). And Justice Thomas is surely correct that concern about racial discrimination played a significant role in the development of our modern capital sentencing jurisprudence. Ante, at 479— 484. Where I cannot agree with Justice Thomas is in the remarkable suggestion that the Court's decision in Penry v. Lynaugh, 492 U.S. 302 (1989), somehow threatens what progress we have made in eliminating racial discrimination and other arbitrary considerations from the capital sentencing determination.

In recent years, the Court's capital punishment cases have erected four important safeguards against arbitrary imposition of the death penalty. First, notwithstanding a minority view that proportionality should play no part in our analysis,^[1] we have concluded that death is an impermissible punishment for certain offenses. Specifically, neither the crime of rape nor the kind of unintentional homicide referred to by Justice Thomas, ante, at 485, may now support a death sentence. See Enmund v. Florida, 458 U.S. 782 (1982); Coker v. Georgia, 433 U.S. 584 (1977).

Second, as a corollary to the proportionality requirement, the Court has demanded that the States narrow the class of individuals eligible for the death penalty, either through statutory definitions of capital murder, or through statutory specification of aggravating circumstances. This narrowing requirement, like the categorical exclusion of the offense of rape, has significantly minimized the risk of racial bias in the sentencing process.^[2] Indeed, as I pointed out in my dissent in McCleskey v. Kemp, 481 U.S. 279 (1987), there is strong empirical evidence that an adequate narrowing of the class of death-eligible offenders would eradicate any significant risk of bias in the imposition of the death penalty.^[3]

Third, the Court has condemned the use of aggravating factors so vague that they actually enhance the risk that unguided discretion will control the sentencing determination. See, e. g., Maynard v. Cartwright, 486 U.S. 356 (1988) (invalidating "especially heinous, atrocious, or cruel" aggravating circumstance); Godfrey v. Georgia, 446 U.S. 420 (1980) (invalidating "outrageously or wantonly vile, horrible or inhuman" aggravating circumstance). An aggravating factor that invites a judgment as to whether a murder committed by a member of another race is especially "heinous" or "inhuman" may increase, rather than decrease, the chance of arbitrary decisionmaking, by creating room for the influence of personal prejudices. In my view, it is just such aggravating factors, which fail to cabin sentencer discretion in the determination of death eligibility, that pose the "evident danger" of which Justice Thomas warns. See ante, at 479.

Finally, at the end of the process, when dealing with the narrow class of offenders deemed death eligible, we insist that the sentencer be permitted to give effect to all relevant mitigating evidence offered by the defendant, in making the final sentencing determination. See, e. g., Eddings v. Oklahoma, 455 U.S. 104 (1982); Lockett v. Ohio, 438 U.S. 586 (1978). I have already explained my view that once the class of death-eligible offenders is sufficiently narrowed, consideration of relevant, individual mitigating circumstances in no way compromises the "rationalizing principle," ante, at 490 (Thomas, J., concurring), of Furman v. Georgia, 408 U.S. 238 (1972). See Walton v. Arizona, 497 U.S. 639, 715— 719 (1990) (Stevens, J., dissenting). To the contrary, the requirement that sentencing decisions be guided by consideration of relevant mitigating evidence reduces still further the chance that the decision will be based on irrelevant factors such as race. Lockett itself illustrates this point. A young black woman,^[4] Lockett was sentenced to death because the Ohio statute "did not permit the sentencing judge to consider, as mitigating factors, her character, prior record, age, lack of specific intent to cause death, and her relatively minor part in the crime." 438 U.S., at 597. When such relevant facts are excluded from the sentencing determination, there is more, not less, reason to believe that the sentencer will be left to rely on irrational considerations like racial animus.

I remain committed to our "mitigating" line of precedent, as a critical protection against arbitrary and discriminatory capital sentencing that is fully consonant with the principles of Furman. Nothing in Justice Thomas' opinion explains why the requirement that sentencing decisions be based on relevant mitigating evidence, as applied by Penry, increases the risk that those decisions will be based on the irrelevant factor of race. More specifically, I do not see how permitting full consideration of a defendant's mental retardation and history of childhood abuse, as in Penry, or of a defendant's youth, as in this case, in any way increases the risk of race-based or otherwise arbitrary decisionmaking.

Justice Souter, in whose dissent I join, has demonstrated that the decision in Penry is completely consistent with our capital sentencing jurisprudence. In my view, it is also faithful to the goal of eradicating racial discrimination in capital sentencing, which I share with Justice Thomas.

Justice Souter, with whom Justice Blackmun, Justice Stevens, and Justice O'Connor join, dissenting.

In Penry v. Lynaugh, 492 U.S. 302 (1989), we concluded that a petitioner did not seek the benefit of a "new rule" in claiming that the Texas special issues did not permit the sentencing jury in his case to give full mitigating effect to certain mitigating evidence, and we therefore held that the retroactivity doctrine announced in Teague v. Lane, 489 U.S. 288, 301 (1989) (plurality opinion), did not bar the claim. See 492 U.S., at 314-319. The only distinctions between the claim in Penry and those presented here go to the kind of mitigating evidence presented for the jury's consideration, and the distance by which the Texas scheme stops short of allowing full effect to be given to some of the evidence considered. Neither distinction makes a difference under Penry or the prior law on which Penry stands. Accordingly, I would find no bar to the present claims and would reach their merits: whether the mitigating force of petitioner's youth, unfortunate background, and traits of decent character could be considered adequately by a jury instructed only on the three Texas special issues.^[1] I conclude they could not be, and I would reverse the sentence of death and remand for resentencing. From the Court's contrary judgment, I respectfully dissent.

I

The doctrine of Teague v. Lane, supra, that a state prisoner seeking federal habeas relief may not receive retroactive benefit of a "new rule" of law, has proven hard to apply. We have explained its crucial term a number of ways. Justice O'Connor wrote in Teague itself that "[i]n general . . . a case announces a new rule when it breaks new ground or imposes a new obligation on the States or the Federal Government. . . . To put it differently, a case announces a new rule if the result was not dictated by precedent at the time the defendant's conviction became final." 489 U. S., at 301 (plurality opinion) (emphasis in original). We have said that novelty turns on whether the rule would represent a "developmen[t] in the law over which reasonable jurists [could] disagree," Sawyer v. Smith, 497 U.S. 227, 234 (1990), and we have emphasized that reasonableness is not a wholly deferential standard, by making it clear that the existence of conflicting authority does not alone imply that any rule resolving that conflict is a new one, Stringer v. Black, 503 U.S. 222, 236-237 (1992).

One general rule that has emerged under Teague is that application of existing precedent in a new factual setting will not amount to announcing a new rule. See Wright v. West, 505 U.S. 277, 304 (1992) (O'Connor, J., joined by Blackmun and Stevens, JJ., concurring in judgment) ("If a proffered factual distinction between the case under consideration and pre-existing precedent does not change the force with which the precedent's underlying principle applies, the distinction is not meaningful, and any deviation from precedent is not reasonable"); id., at 309 (Kennedy, J., concurring in judgment) ("Where the beginning point is a rule of this general application, a rule designed for the specific purpose of evaluating a myriad of factual contexts, it will be the infrequent case that yields a result so novel that it forges a new rule, one not dictated by precedent"); id., at 313 (Souter, J., concurring in judgment) (Teague "does not mean, of course, that a habeas petitioner must be able to point to an old case decided on facts identical to the facts of his own").

That said, it can be a difficult question whether a particular holding presents simply a new setting for an old rule, or announces a new one. The question is not difficult in this case, however, for its answer is governed by Penry, supra, at 313, 329, the first case in which a majority of the Court adopted the approach to retroactivity put forward by the plurality in Teague. See 492 U.S., at 313. The circumstances in which petitioner Penry sought relief, and the rule that he sought to have applied, are virtually indistinguishable from the circumstances presented and the rule of decision sought by Graham in this case. We denied certiorari in Penry's direct appeal in 1986. Penry v. Texas, 474 U.S. 1073 (1986). The Texas Court of Criminal Appeals affirmed Graham's conviction and sentence of death in 1984, Graham v. State, No. 68,916, and Graham did not seek certiorari in this Court. In both cases, therefore, under the reasoning employed by the majority, see ante, at 467, "[t]his Court's decisions in Lockett v. Ohio, 438 U.S. 586 (1978), and Eddings v. Oklahoma, 455 U.S. 104 (1982), were rendered before [petitioners'] conviction[s] became final." Penry, 492 U. S., at 314-315. Because Penry was "entitled to the benefit of those decisions," id., at 315, so, on a comparable claim, is Graham.

Our description of Penry's claim applies, indeed, almost precisely to Graham's claim in this case. Of Penry, we said:

"[He] does not challenge the facial validity of the Texas death penalty statute, which was upheld against an Eighth Amendment challenge in Jurek v. Texas, 428 U.S. 262 (1976). Nor does he dispute that some types of mitigating evidence can be fully considered by the sentencer in the absence of special jury instructions. See Franklin v. Lynaugh, 487 U.S. 164, 175 (1988) (plurality opinion); id., at 185-186 (O'Connor, J., concurring in judgment). Instead, [he] argues that, on the facts of this case, the jury was unable to fully consider and give effect to the mitigating evidence . . . in answering the three special issues." Ibid.

In deciding whether he sought benefit of a "new rule," we went on to say:

"Lockett underscored Jurek `s recognition that the constitutionality of the Texas scheme `turns on whether the enumerated questions allow consideration of particularized mitigating factors.' Jurek, 428 U. S., at 272. The plurality opinion in Lockett indicated that the Texas death penalty statute had `survived the petitioner's Eighth and Fourteenth Amendment attack [in Jurek ] because three Justices concluded that the Texas Court of Criminal Appeals had broadly interpreted the second question—despite its facial narrowness—so as to permit the sentencer to consider "whatever mitigating circumstances" the defendant might be able to show.' 438 U.S., at 607." Id., at 317.

We then reviewed the reaffirmation in Eddings v. Oklahoma, 455 U.S. 104 (1982), of the principle that "a sentencer may not be precluded from considering, and may not refuse to consider, any relevant mitigating evidence offered by the defendant as the basis for a sentence less than death." Thus, we said, "at the time Penry's conviction became final," as at the time Graham's did,

"it was clear from Lockett and Eddings that a State could not, consistent with the Eighth and Fourteenth Amendments, prevent the sentencer from considering and giving effect to evidence relevant to the defendant's background or character or to the circumstances of the offense that mitigate against imposing the death penalty. Moreover, the facial validity of the Texas death penalty statute had been upheld in Jurek on the basis of assurances that the special issues would be interpreted broadly enough to enable sentencing juries to consider all of the relevant mitigating evidence a defendant might present." 492 U.S., at 318.

Graham contends that Jurek v. Texas, 428 U.S. 262 (1976), Lockett v. Ohio, 438 U.S. 586 (1978), and Eddings v. Oklahoma, supra, were not honored in the application of the Texas special issues on the facts of his case, and, in this respect, too, his position is identical to that of Penry, who argued that "those assurances [on which Jurek rests] were not fulfilled in his particular case because, without appropriate instructions, the jury could not fully consider and give effect to [his] mitigating evidence . . . in rendering its sentencing decision." 492 U.S., at 318 (emphasis in original). In Penry, we held that nothing foreclosed such a claim:

"The rule Penry seeks—that when such mitigating evidence is presented, Texas juries must, upon request, be given jury instructions that make it possible for them to give effect to that mitigating evidence in determining whether the death penalty should be imposed—is not a `new rule' under Teague because it is dictated by Ed- dings and Lockett. Moreover, in light of the assurances upon which Jurek was based, we conclude that the relief Penry seeks does not `impos[e] a new obligation' on the State of Texas. Teague, 489 U. S., at 301." Id., at 318-319.

Thus in Penry we held that petitioner sought nothing but the application to his case of the rule announced in Eddings and Lockett, that "a State could not, consistent with the Eighth and Fourteenth Amendments, prevent the sentencer from considering and giving effect to evidence relevant to the defendant's background or character or to the circumstances of the offense that mitigate against imposing the death penalty." 492 U.S., at 318.

The first distinction between Penry's claim and that of Graham is the type of mitigating evidence involved. Penry's went to "mental retardation and abused childhood"; Graham's involves youthfulness, unfortunate background, and traits of decent character. But any assertion that this should make any difference flies in the face of Justice Kennedy's opinion from last Term, quoted before, that "a rule of this general application, a rule designed for the specific purpose of evaluating a myriad of factual contexts [will only infrequently] yiel[d] a result so novel that it forges a new rule, one not dictated by precedent." Wright v. West, 505 U. S., at 309 (opinion concurring in judgment). Nor is the second distinction any more material, that Penry's evidence of retardation could claim no mitigating effect under the second Texas issue, which asks the jury to assess a defendant's future dangerousness, whereas Graham's evidence of youth and decency could claim some.^[2] The point under Lockett, Eddings, and Penry is that sentencing schemes must allow the sentencer to give full mitigating effect to evidence; Graham's claim that his evidence could receive only partial consideration is just as much a claim for application of the preexisting rule demanding the opportunity for full effect as was Penry's claim that his retardation could be given no effect under the second Texas special issue.

Thus, from our conclusion that the rule from which the petitioner sought to benefit in Penry was not "new," it necessarily follows that the rule petitioner Graham seeks here is not new either. Indeed, that is the conclusion reached even by respondent who concedes that "if Graham is asserting the existence of a constitutional defect that can be cured by supplemental instructions, his claim likewise is not barred." Brief for Respondent 29, n. 10.^[3]

The Court's conclusion to the contrary rests on the assumption that an additional instruction is required under Penry only where there is mitigating evidence without any "mitigating relevance" to the second, future dangerousness special issue. See ante, at 475. But that was not the holding of Penry, which reiterates the Eighth Amendment requirement expressed in Lockett and Eddings that the jury be able "to consider fully [the defendant's] mitigating evidence," Penry, 492 U. S., at 323, and requires a separate instruction whenever such evidence "has relevance to . . . moral culpability beyond the scope of the special issues," id., at 322. Indeed, Justice Scalia's dissent in Penry recognized that "[w]hat the Court means by `fully consider' (what it must mean to distinguish Jurek ) is to consider for all purposes, including purposes not specifically permitted by the questions. " Id., at 355 (opinion dissenting in relevant part) (emphasis in original). That dissent argued that this was not what was required by the Constitution, see id., at 358-360,^[4] but it correctly described the holding in the Court's opinion in Penry itself. Nothing in Penry aside from Justice Scalia's dissent, and nothing in the controlling opinions in Lockett or Eddings, suggested that this Eighth Amendment requirement will be obviated by the happenstance that a defendant's particular mitigating evidence is relevant to one of the special issues, even though it may have mitigating force beyond the scope of that issue.

Penry plainly answered the Teague question that the majority answers differently today, a question that even respondent did not see fit to raise again. Penry controls in this respect, and we should adhere to it.

II

I therefore turn to the merits of the claim,^[5] which are properly before us.^[6]Penry again controls, for reasons already anticipated in the Teague analysis, but bearing some expansion here.

A

Following the first grant of certiorari in this case, after we vacated the judgment and remanded for reconsideration in light of Penry, see Graham v. Lynaugh, 492 U.S. 915 (1989), a panel of the Court of Appeals for the Fifth Circuit decided to vacate Graham's death sentence and remand. Graham v. Collins, 896 F.2d 893 (1990). The Court of Appeals then took the case en banc, however, and, by a vote of 7 to 6, construed Penry to require no additional instruction "in instances where no major mitigating thrust of the evidence is substantially beyond the scope of all the special issues." 950 F.2d 1009, 1027 (CA5 1992) (en banc). It also limited the application of Penry to mitigating evidence of circumstances that were not "transitory," but were "uniquely severe permanent handicaps with which the defendant was burdened through no fault of his own." See 950 F.2d, at 1029. Penry lends no support for these limitations, however, and they are plainly at odds with other controlling Eighth Amendment precedents, which the Court does not purport to disturb.

B

Our cases have construed the Eighth Amendment to impose two limitations upon a state capital sentencing system. First, in determining who is eligible for the death penalty, the "State must `narrow the class of murderers subject to capital punishment,' . . . by providing `specific and detailed guidance' to the sentencer." McCleskey v. Kemp, 481 U.S. 279, 303 (1987) (quoting Gregg v. Georgia, 428 U.S. 153, 196 (1976), and Proffitt v. Florida, 428 U.S. 242, 253 (1976)). Second, "the Constitution [nonetheless] limits a State's ability to narrow a sentencer's discretion to consider relevant evidence that might cause it to decline to impose the death sentence." 481 U.S., at 304 (emphasis in original). It is this latter limitation that concerns us today.

Our cases require that a sentencer in a capital case be permitted to give a "reasoned moral response" to the defendant's mitigating evidence. See California v. Brown, 479 U.S. 538, 545 (1987) (O'Connor, J., concurring) (emphasis deleted). In so doing, "[t]he sentencer . . . [cannot] be precluded from considering as a mitigating factor, any aspect of a defendant's character or record and any of the circumstances of the offense that the defendant proffers as a basis for a sentence less than death." Lockett, 438 U. S., at 604 (plurality opinion) (emphasis in original; footnote omitted). This is understood to follow from our conclusion that "[a]ny exclusion of the `compassionate or mitigating factors stemming from the diverse frailties of humankind' that are relevant to the sentencer's decision would fail to treat all persons as `uniquely individual human beings.' " McCleskey, supra, at 304 (quoting Woodson v. North Carolina, 428 U.S. 280, 304 (1976)).

As we first described it in Jurek, the Texas scheme to be measured against this obligation assesses mitigating (as well as aggravating) evidence by looking both backward to the defendant's moral culpability for the crime itself, as distinct from strictly legal guilt, and forward to his likely behavior if his life is not taken. Thus the first issue requires the sentencer to determine whether the defendant acted deliberately, and the third asks for assessment of any provocation as mitigating the fault of any response. Each issue demands an examination of past fact as bearing on the moral significance of a past act. The second issue, on the other hand, calls for a prediction of future behavior, prompting a judgment that is moral in the utilitarian sense that society may legitimately prefer to preserve the lives of murderers unlikely to endanger others in the future, as against the lives of the guilty who pose continuing threats.

While these issues do not exhaust the categories of mitigating fact,^[7] at the time Jurek was decided the Court of Criminal Appeals of Texas had indicated that the second special issue would be given a wide enough compass to allow jury consideration of such diverse facts as prior record and the character of past crimes, duress, or emotional pressure associated with the instant crime, and the age of the defendant. Jurek, 428 U. S., at 272-273. Thus, we had a reasonable expectation that the sentencer would have authority to give comprehensive effect to each defendant's mitigating evidence. As Penry revealed, however, and as the facts of this case confirm, neither the second nor the other special issues have been construed with enough scope to allow the full consideration of mitigating potential that Lockett and Eddings confirmed are required, and challenges to the Texas statute as applied may be sustained despite the statute's capacity to withstand Jurek `s facial challenge. In its holding that a death sentence resulting from the application of the Texas special issues could not be upheld unless the jury was able "to consider fully [the defendant's] mitigating evidence," 492 U.S., at 323,^[8]Penry is a perfectly straightforward application of the Eighth Amendment's requirement of individualized sentencing.^[9]

The specific question in Penry itself was whether the mitigating evidence of Penry's mental retardation and history of abuse "as it bears on [Penry's] personal culpability" could be taken account of under the Texas special issues, ibid., and in deciding that case, we examined each special issue in turn. We concluded first that the jury instruction barred full consideration of the evidence of retardation and personal abuse under the first, or "deliberate[ness]," special issue, see ibid., and second that insofar as the evidence bore on personal culpability, it could not be given mitigating effect under the issue of "future dangerousness." As to the latter, indeed, it could have been considered only as an aggravating factor. Although we described Penry's evidence as a "two-edged sword . . . diminish[ing] his blameworthiness for his crime even as it indicates that there is a probability that he will be dangerous in the future," id., at 324, the dilemma thus presented was not essential to our conclusion that the second special issue failed to meet the State's constitutional obligations. The point was simply that the special issue did not allow the jury to give effect to the mitigating force of Penry's evidence as it bore on his personal culpability. Finally we concluded that "a juror who believed Penry lacked the moral culpability to be sentenced to death could not express that view in answering the third [`provocation'] special issue if she also concluded that Penry's action was not a reasonable response to provocation." Id., at 324-325. In sum, full consideration of the tendency of retardation and a history of abuse to mitigate moral culpability was impossible.

C

Graham's evidence falls into three distinct categories. As to each, our task is to take the same analytical steps we undertook in Penry, to see whether the sentencing jury could give it full mitigating effect.

1

First, there was the evidence of Graham's youth. He was 17 when he committed the murder for which he was convicted, and he was sentenced less than six months after the crime. Youth may be understood to mitigate by reducing a defendant's moral culpability for the crime, for which emotional and cognitive immaturity and inexperience with life render him less responsible, see Eddings v. Oklahoma, 455 U. S., at 115-116, and youthfulness may also be seen as mitigating just because it is transitory, indicating that the defendant is less likely to be dangerous in the future.

As with Penry's evidence of mental retardation, the mitigating force of Graham's youth could not be fully accounted for under the first, "deliberateness" issue, given the trial judge's explanation of that issue to the jury. While no formal jury instruction explained what "deliberate" meant, the judge emphasized at voir dire that "deliberate" meant simply "intentional," see App. 90, 127, 169, 205-206, 246, 291, 319— 320, 353, 420, a definition that hardly allowed exhaustion of the mitigating force of youth. A young person may perfectly well commit a crime "intentionally," but our prior cases hold that his youth may nonetheless be treated as limiting his moral culpability because he "`lack[s] the experience, perspective, and judgment' expected of adults." Eddings, supra, at 116 (quoting Bellotti v. Baird, 443 U.S. 622, 635 (1979)).

We have already noted that the Court of Appeals answered this difficulty by reasoning that the "major mitigating thrust" of the evidence could be given effect under the second special issue calling for assessment of future dangerousness. The errors of this view we have also seen. First, nothing in Penry suggests that partial consideration of the mitigating effect of the evidence satisfies the Constitution. Penry, like Eddings and the Lockett plurality before that, states an Eighth Amendment demand that the sentencer "consider and give effect to . . . mitigating evidence" "fully," 492 U.S., at 318, and when such evidence "has relevance to. . . moral culpability beyond the scope of the special issues," constitutional standards require a separate instruction authorizing that complete effect be given, id., at 322. See McCleskey, 481 U. S., at 304 ("[A]ny exclusion" of mitigating evidence is inconsistent with the Eighth Amendment's individualized sentencing requirements). Thus, even if the future dangerousness issue allowed the jury to recognize Graham's evanescent youth as tending to mitigate any danger if he were imprisoned for life, it would still fail the test of the Eighth Amendment because the jury could not give effect to youth as reducing Graham's moral culpability.^[10] The Eighth Amendment requires more than some consideration of mitigating evidence.

The Court of Appeals also erred in thinking the second special issue adequate even to take account of the possibility that Graham may be less dangerous as he ages. The issue is stated in terms of the statutory question "whether there is a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society." Tex. Code Crim. Proc. Ann., Art. 37.071(b)(2) (Vernon 1981). Because a boy who killed at 17 and was promptly tried (as Graham was) could well be held dangerous in the future by reason of continuing youth, it was error to limit Penry to cases in which a mitigating condition is permanent. See 950 F.2d, at 1029. It is no answer to say youth is fleeting; it may not be fleeting enough, and a sufficiently young defendant may have his continuing youth considered under the second issue as aggravating, not mitigating. In this case, moreover, the possibility of taking youth as aggravating without any recognition of mitigating effect was vastly intensified by remarks of the trial judge permitting a finding of future dangerousness based even on the probability that petitioner might commit minor acts of criminal vandalism to property such as scratching someone's car or tearing up the lawn of a high school by riding a motorcycle over it. See App. 128-129, 172, 210, 247-248, 295, 321-322, 354-355, 389-390, 422, 455.

Finally, because Graham was convicted of shooting and killing a man during a robbery, the situation with respect to the third special issue in this case is the same as it was for petitioner in Penry. The evidence of youth was irrelevant to the reasonableness of any provocation by the deceased of which there was no evidence in any event.

A juror could thus have concluded that the responses to the special issues required imposition of the death penalty even though he believed that Graham, by reason of his youth, "lacked the moral culpability to be sentenced to death." Penry, 492 U. S., at 324. Without more, the case is controlled by Penry, and additional instruction was required.

2

The next category of evidence at issue is that of Graham's difficult upbringing, of his mother's mental illness and repeated hospitalization, and his shifting custody to one family relation or another. We have specifically held that such circumstances may be considered in mitigation, particularly on the conduct of a defendant so young, see, e. g., Eddings, supra, at 115, where upbringing might be deforming enough to affect the capacity for culpability. Where, as here, however, that is not obviously the case, and deliberateness is said to turn on intention, there is no assurance that the first issue allows the full scope of its mitigating effect to be considered. As with youth itself, upbringing could be treated as aggravating under the future dangerousness issue, and it has no mitigating potential under the third issue of provocation. Again, as with youth, there is no room in the former Texas special issues as applied in this case to take full account of such mitigating relevance as the jury might find.

3

Finally, Graham argues that the jury was unable to take account of redeeming character traits revealed by evidence that growing up he had voluntarily helped his parents and grandparents with household chores, that he was a religious person who had attended church regularly with his grandmother, and that he had contributed to the support of his own children with money earned from a job with his father.

I do not accept petitioner's contention that the jury could not give adequate consideration to the testimony on these matters. Insofar as the evidence tended to paint Graham as a person unlikely to pose a future danger, the jury could consider it under the second special issue. Insofar as the jury was unable, as Graham alleges, to give the evidence further effect to diminish Graham's "moral culpability," Brief for Petitioner 36, 37, 39, it is enough to say that the relevance of the evidence to moral culpability was simply de minimis. Voluntary chores for and church attendance with a relative, and supplying some level of support for one's children have virtually no bearing on one's culpability for crime in the way that immaturity or permanent damage due to events in childhood may. Because I do not understand petitioner to be arguing that the jury should have been allowed to consider the evidence as revealing some element of value unrelated to the circumstances of the crime, see Franklin, 487 U. S., at 186 (O'Connor, J., concurring in judgment); id., at 189 (Stevens, J., dissenting), I do not address that issue.

III

I would hold that Penry and preceding Eighth Amendment cases of this Court require petitioner's death sentence to be vacated, and would remand the case for resentencing by the state courts.

NOTES

[*] Steven B. Rosenfeld and Allen Cazier filed a brief for Miguel A. Richardson as amicus curiae urging reversal.

[1] The Texas Legislature amended the statute in 1991. Those changes are set forth in the opinion of the Court of Appeals. 950 F.2d 1009, 1012, n. 1 (CA5 1992) (en banc).

[2] Penry further held that its result was dictated by the Court's prior decisions in Eddings v. Oklahoma, 455 U.S. 104 (1982), and Lockett v. Ohio, 438 U.S. 586 (1978) (plurality opinion), within the sense required by Teague v. Lane, 489 U.S. 288 (1989), and thus that its rule applied to cases on collateral review. See Penry, 492 U. S., at 314-319.

[3] To the contrary, the Court made clear in that case the limited nature of the question presented: "Penry does not challenge the facial validity of the Texas death penalty statute, which was upheld against an Eighth Amendment challenge in Jurek v. Texas, 428 U.S. 262 (1976). Nor does he dispute that some types of mitigating evidence can be fully considered by the sentencer in the absence of special jury instructions. See Franklin v. Lynaugh, 487 U.S. 164, 175 (1988) (plurality opinion); id., at 185-186 (O'Connor, J., concurring in judgment). Instead, Penry argues that, on the facts of this case, the jury was unable to fully consider and give effect to the mitigating evidence of his mental retardation and abused background in answering the three special issues." 492 U. S.,at 315.

[1] The Court decided two cases together with Furman v. Georgia, 408 U.S. 238 (1972): Jackson v. Georgia, No. 69-5030, and Branch v. Texas, No. 69-5031. A fourth case, Aikens v. California, No. 68-5027, was argued with Furman but was dismissed as moot. 406 U.S. 813 (1972).

[2] Furman was surprised to discover the victim at home and, while trying to escape, accidentally tripped over a wire, causing his pistol to fire a single shot through a closed door, thereby killing the victim. See 408 U.S., at 294-295, n. 48 (Brennan, J., concurring).

[3] According to the published account of one Legal Defense Fund lawyer who participated in the campaign, the Fund—though it had had experience with racial discrimination in rape cases in the South—did not seriously consider a broader offensive against the death penalty until three Members of this Court, in an opinion dissenting from a denial of certiorari, offered a "strong foundation" for such a strategy. Meltsner 27-35. See Rudolph v. Alabama, 375 U.S. 889 (1963) (Goldberg, J., joined by Douglas and Brennan, JJ., dissenting from denial of certiorari) (calling on the Court to decide "whether the Eighth and Fourteenth Amendments . . . permit the imposition of the death penalty on a convicted rapist who has neither taken nor endangered human life" and suggesting several lines of argument in the form of questions that "seem relevant and worthy of . . . consideration").

[4] See also Meltsner 25: "[L]awyers attempting to thrust egalitarian or humanitarian reforms on a reluctant society prefer to use the courts because lifetime-appointed federal judges are somewhat more insulated from the ebb and flow of political power and public opinion than legislators or executives."

[5] The Federal Government later acknowledged before this Court that in 11 Southern States between 1945 and 1965, "[t]he data revealed that among all those convicted of rape, blacks were selected disproportionately for the death sentence." App. to Brief for United States as Amicus Curiae in Gregg v. Georgia, O. T. 1975, No. 74-6257, p. 4a. Furthermore, the Government stated, "we do not question [the] conclusion that during the 20 years in question, in southern states, there was discrimination in rape cases." Id., at 5a. We eventually struck down the death penalty for convicted rapists under the Eighth Amendment, not on the basis of discriminatory application, but as an excessive and disproportionate punishment. Coker v. Georgia, 433 U.S. 584 (1977).

[6] Gregg v. Georgia, 428 U.S. 153 (1976); Proffitt v. Florida, 428 U.S. 242 (1976); Jurek v. Texas, 428 U.S. 262 (1976); Woodson v. North Carolina, 428 U.S. 280 (1976); Roberts v. Louisiana, 428 U.S. 325 (1976).

[7] As in Furman, the NAACP Legal Defense Fund represented the three petitioners in Woodson and Roberts, who were black. In addition to contending that the death penalty was a cruel and unusual punishment, the Fund lawyers argued in these cases that despite the mandatory nature of North Carolina's and Louisiana's statutes, the process of imposing the penalty on these petitioners was infected at key junctures with the potential for selective and discriminatory discretion, most importantly the possibility that sentencing juries in cases involving sympathetic defendants would acquit or convict on lesser charges. See Brief for Petitioners in Woodson v. North Carolina, O. T. 1975, No. 75-5491, pp. 22-39; Brief for Petitioner in Roberts v. Louisiana, O. T. 1975, No. 75-5844, pp. 30-65. The unsuccessful petitioners in Gregg, Proffitt, and Jurek were white. See Brief for United States as Amicus Curiae in Gregg v. Georgia, O. T. 1975, No. 74-6257, p. 68.

[8] Lockett aided and abetted an armed robbery that resulted in a murder. She drove the getaway car but did not carry out the robbery and did not intend to bring about the murder. See 438 U.S., at 589-591; id., at 613— 617 (Blackmun, J., concurring in part and concurring in judgment). Lockett was represented by the same lawyers from the Legal Defense Fund who had represented the petitioners in Furman, Woodson, and Roberts.

[9] The Texas special issues involved here did a considerably better job of rationalizing sentencing discretion than even the elaborate Georgia system approved in Gregg, where juries still retained power "to return a sentence of life, rather than death, for no reason whatever, simply based upon their own subjective notions of what is right and what is wrong." Woodson, 428 U. S., at 314-315 (Rehnquist, J., dissenting). As a regrettable but predictable consequence of Penry v. Lynaugh, 492 U.S. 302 (1989), the Texas Legislature has since amended its sentencing statute, which now invites the jury to react subjectively to "all" circumstances, including "the personal moral culpability of the defendant." See Tex. Code Crim. Proc. Ann., Art. 37.0711(2)(e) (Vernon Supp. 1993) (applicable to offenses committed on or after September 1, 1991).

[10] Indeed, it can be argued that we have already implicitly overruled Penry in significant respects. In Saffle v. Parks, 494 U.S. 484 (1990), we gave a dramatically narrow reading to Penry, reaffirming that under Lockett v. Ohio, 438 U.S. 586 (1978), and Eddings v. Oklahoma, 455 U.S. 104 (1982), the State is free to "limi[t] the manner in which [a defendant's] mitigating evidence may be considered." 494 U.S., at 491. And in Boyde v. California, 494 U.S. 370 (1990), we expressly rejected the significance of Penry `s conclusion that "`a reasonable juror could well have believed that there was no vehicle for expressing the view that Penry did not deserve to be sentenced to death based upon his mitigating evidence.' " 494 U.S., at 379 (emphasis in original) (quoting Penry, supra, at 326). Boyde held instead that a jury instruction will run afoul of Eddings only if "there is a reasonable likelihood that the jury has applied the challenged instruction in a way that prevents the consideration of constitutionally relevant evidence," and the Court made it clear that "a capital sentencing proceeding is not inconsistent with the Eighth Amendment if there is only a possibility of such an inhibition." 494 U.S., at 380.

[11] Under the Federal Sentencing Reform Act, for example, Congress has instructed the United States Sentencing Commission to study the difficult question whether certain specified offender characteristics "have any relevance" in sentencing. 28 U.S. C. § 994(d). In response to this directive, the Sentencing Commission has issued guidelines providing, among other things, that race, sex, national origin, creed, religion, and socioeconomic status "are not relevant in the determination of a sentence." United States Sentencing Commission, Guidelines Manual § 5H1.10 (Nov. 1992). Congress has also concluded that a defendant's education, vocational skills, employment record, and family and community ties are inappropriate sentencing factors. 28 U.S. C. § 994(e). Thus, the Sentencing Guidelines declare that these and other factors "are not ordinarily relevant in determining whether a sentence should be outside the applicable guideline range." See USSG ch. 5, pt. H, intro. comment. Similar guidelines, it seems to me, could be applied in capital sentencing consistent with the Eighth Amendment, as long as they contributed to the rationalization of the process.

[1] See Harmelin v. Michigan, 501 U.S. 957 (1991).

[2] As an indication of the difference such narrowing can make, it is worthwhile to note that at the time we decided Furman v. Georgia, 408 U.S. 238 (1972), in addition to defendants convicted of first-degree murder, almost all defendants convicted of forcible rape, armed robbery, and kidnaping were eligible for the death penalty. See Walton v. Arizona, 497 U.S. 639, 715 (1990) (Stevens, J., dissenting).

[3] "The Court's decision appears to be based on a fear that the acceptance of McCleskey's claim would sound the death knell for capital punishment in Georgia. If society were indeed forced to choose between a racially discriminatory death penalty (one that provides heightened protection against murder `for whites only') and no death penalty at all, the choice mandated by the Constitution would be plain. But the Court's fear is unfounded. One of the lessons of the Baldus study is that there exist certain categories of extremely serious crimes for which prosecutors consistently seek, and juries consistently impose, the death penalty without regard to the race of the victim or the race of the offender. If Georgia were to narrow the class of death-eligible defendants to those categories, the danger of arbitrary and discriminatory imposition of the death penalty would be significantly decreased, if not eradicated. As Justice Brennan has demonstrated in his dissenting opinion, such a restructuring of the sentencing scheme is surely not too high a price to pay." McCleskey v. Kemp, 481 U.S. 279, 367 (1987) (Stevens, J., dissenting) (internal citation omitted).

[4] See Brief for Petitioner in Lockett v. Ohio, O. T. 1977, No. 76-6997, p. 10.

[1] After Texas' capital punishment statute was invalidated in Branch v. Texas, one of the cases decided with Furman v. Georgia, 408 U.S. 238 (1972), Texas enacted a new capital sentencing statute. This statute, under which petitioner Gary Graham was sentenced, provides that: "(b) [o]n conclusion of the presentation of the evidence [at the sentencing phase of a capital murder trial], the court shall submit the following issues to the jury:

"(1) whether the conduct of the defendant that caused the death of the deceased was committed deliberately and with the reasonable expectation that the death of the deceased or another would result;

"(2) whether there is a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society; and

"(3) if raised by the evidence, whether the conduct of the defendant in killing the deceased was unreasonable in response to the provocation, if any, by the deceased.

. . . . .

"(e) If the jury returns an affirmative finding on each issue submitted under this article, the court shall sentence the defendant to death." Tex. Code Crim. Proc. Ann., Art. 37.071 (Vernon 1981).

Following our decision in Penry v. Lynaugh, 492 U.S. 302 (1989), Texas adopted a new capital sentencing procedure which is not at issue here. See Tex. Code Crim. Proc. Ann., Art. 37.071 (Vernon Supp. 1992).

[2] This distinction does not even apply to Graham's claim that the sentencing jury could not give full mitigating effect to the evidence of his unfortunate background. Of course, in this regard, despite their mitigating force, Penry's evidence of an abused childhood and Graham's evidence of an unfortunate background both have the same tendency to support only an affirmative answer to the future dangerousness special issue. The Court does not explain why, under its reasoning, Graham's claim concerning evidence of his background is barred by Teague v. Lane, 489 U.S. 288 (1989) (plurality opinion). See ante, at 475 (undifferentiated references to all of "Graham's evidence").

[3] Respondent's only argument concerning the application of Teague is that petitioner's claim is Teague -barred if "his claim is so extensive as to constitute a facial challenge to the Texas statute." Brief for Respondent 13. In other words, "if sustaining Graham's claim would necessarily require that Jurek be overruled, it is barred by Teague. " Id., at 29, n. 10. However, petitioner does not ask that Jurek v. Texas, 428 U.S. 262 (1976), be overruled. Indeed, he concedes that the Texas statute has been applied constitutionally in those cases such as Franklin v. Lynaugh, 487 U.S. 164 (1988), in which the mitigating evidence can be given "full" mitigating weight under the special issues. See Brief for Petitioner 15, and n. 12. Thus, respondent's Teague argument has no application to this case.

[4] See also Penry, 492 U. S., at 356 (Scalia, J., dissenting in part) (arguing, contrary to the holding of the Court, that after Jurek "there remains available, in an as-applied challenge to the Texas statute," only "the contention that a particular mitigating circumstance is in fact irrelevant to any of the three questions it poses, and hence could not be considered").

[5] The full Court may do the same in responding to several pending petitions for certiorari presenting the same question involved in this case, but on direct review. See, e. g., Johnson v. Texas, cert. pending, No. 92-5653; Jackson v. Texas, cert. pending, No. 91-7399; Boggess v. Texas, cert. pending, No. 91-5862.

[6] At trial petitioner did not seek the additional Penry instruction that he now says is required. Whether the failure to request such an instruction is a bar to a subsequent challenge is a question of state procedure; if the conviction were affirmed by the state appellate courts on the ground that petitioner failed to raise his claim before the trial court, that affirmance could rest on an independent and adequate state-law ground. Here, the Texas Court of Criminal Appeals appears to have addressed petitioner's challenge on the merits in a state postconviction proceeding. See App. 37. In any event, under Texas law, a Penry claim is not procedurally barred even if no additional mitigating-evidence instruction is requested or there is no objection made at trial to the jury instructions. See Selvage v. Collins, 816 S.W.2d 390, 392 (Tex. Crim. App. 1991); Black v. State, 816 S.W.2d 350, 362-369 (Tex. Crim. App. 1991); id., at 367-374 (Campbell, J., concurring). The adequacy of the Texas special issues in this case is therefore properly before us.

[7] Or, indeed, all the ways in which evidence may mitigate against imposition of a death sentence previously mentioned by Members of this Court. See Franklin v. Lynaugh, 487 U. S.,at 186 (O'Connor, J.,joined by Blackmun, J., concurring in judgment) (referring to "positive character traits that might mitigate against the death penalty");id., at 189 (Stevens, J., joined by Brennan and Marshall, JJ., dissenting) (character evidence of "redeeming features" may reveal "virtues that can fairly be balanced against society's interest in killing [a defendant] in retribution for his violent crime"). My analysis today does not require extended consideration of the category suggested in Franklin. See infra, at 521.

[8] See also Jurek, 428 U. S., at 272 (joint opinion of Stewart, Powell, and Stevens, JJ.) ("[T]he constitutionality of the Texas procedures turns on whether the enumerated questions allow consideration of particularized mitigating factors").

[9] Justice Thomas argues, ante, at 493, that the rule applied in Penry "originated entirely from whole cloth in two recent concurring opinions," California v. Brown, 479 U.S. 538, 545 (1987) (O'Connor, J., concurring), and Franklin v. Lynaugh, supra, at 185 (O'Connor, J., concurring in judgment), and that it requires "unbridled" jury discretion, even to the point that the death penalty may be withheld on the basis of race, ante, at 494.

As to the first contention, Lockett v. Ohio, 438 U. S 586 (1978), was understood at the time it was handed down to require that constitutionally relevant mitigating evidence (the definition of which is given below) be given full consideration and effect. See, e. g., id., at 623 (White, J., concurring in part, dissenting in part, and concurring in judgments) (emphasis added) (Lockett "requir[es] as a matter of constitutional law that sentencing authorities be permitted to consider and in their discretion to act upon any and all mitigating circumstances"). This is the understanding upon which Lockett and Eddings have consistently been applied by the Court. See Skipper v. South Carolina, 476 U.S. 1, 7 (1986) ("Assuming. . . that [a State Supreme Court] rule would in any case have the effect of precluding the defendant from introducing otherwise admissible evidence for the explicit purpose of convincing the jury that he should be spared the death penalty because he would pose no undue danger to his jailers or fellow prisoners and could lead a useful life behind bars if sentenced to life imprisonment, the rule would not pass muster under Eddings "); McCleskey v. Kemp, 481 U.S. 279, 306 (1987) ("States cannot limit the sentencer's consideration of any relevant circumstance that could cause it to decline to impose the [death] penalty" (emphasis added)); Franklin v. Lynaugh, supra, at 184-185 (O'Connor, J., joined by Blackmun, J., concurring in judgment); id., at 191-192 (Stevens, J., joined by Brennan and Marshall, JJ., dissenting). While one may argue that this aspect of our Eighth Amendment jurisprudence is in tension with the sentence in Gregg that the State should give the jury guidance as to what factors it "`deems particularly relevant to the sentencing decision,' " ante, at 485 (Thomas, J., concurring) (quoting Gregg v. Georgia, 428 U.S. 153, 192 (1976)), any such tension dates, at the latest, from Eddings, decided in 1982, not from Penry in 1989.

There was one novelty in the concurring opinions in Brown and Franklin, however, in the use of the phrase "reasoned moral response," see supra, at 513, to which Justice Thomas adverts in his concurring opinion. But as the concurring opinion explained in Brown, this is just a shorthand for the individual assessment of personal culpability that Lockett and Eddings mandate. See Brown, supra, at 545. It is, indeed, appropriate shorthand. Justice Thomas himself acknowledges, as I think everyone must, "that `capital punishment is an expression of society's moral outrage at particularly offensive conduct,' " ante, at 498 (quoting Gregg v. Georgia, supra, at 183 (joint opinion of Stewart, Powell, and Stevens, JJ.)), and he reminds us that "[a]ny determination that death is or is not the fitting punishment for a particular crime will necessarily be a moral one," ante, at 494.

Justice Thomas's second concern, about "sympathy for a defendant who is a member of a favored group," ante, at 495, involves an issue of very great seriousness. But the Lockett -Eddings rule is not one of "unbridled" or "unbounded" discretion. See ante, at 494-495. Constitutionally relevant mitigating evidence is limited to "any aspects of a defendant's character or record and any of the circumstances of the offense that the defendant proffers as a basis for a sentence less than death." Lockett, supra, at 604 (plurality opinion). A defendant's race as such is not mitigating as an aspect of his character or record, or as a circumstance of any offense he may have committed.

[10] I note in this regard that the trial judge's remarks at voir dire may have inappropriately left the jury to consider whether Graham would have been dangerous in the future if he were set free. See Brief for Petitioner 8, n. 4. In light of my conclusion that Graham's death sentence should be vacated, I need not address here the propriety of a sentence imposed on the basis of future dangerousness to the public when there is no possibility that a defendant will be sentenced to a term less than life without the possibility of parole.