dissenting.
It is important at the outset to remember — as the Court does not — the narrow question on which we granted certio-rari. We took care to limit our consideration to whether the Eighth and Fourteenth Amendments prohibit the imposition of a death sentence ón an offender because he was 16 years old in 1977 at the time he committed the offense; review of all other questions raised in the petition for certiorari was de-' nied. 450 U. S. 1040 (1981). Yet the Court today goes beyond the issue on which review was sought — and granted — to decide the case on a point raised for the first time in petitioner’s brief to this Court. This claim was neither presented to the Oklahoma courts nor presented to this Court in the petition for certiorari.1 Relying on this “llth-hour” claim, the Court strains to construct a plausible legal theory to support its mandate for the relief granted.
I
In Lockett v. Ohio, 438 U. S. 586 (1978), we considered whether Ohio violated the Eighth and Fourteenth Amendments by sentencing Lockett to death under a statute that “narrowly limit[ed] the sentencer’s discretion to consider the *121circumstances of the crime and the record and character of the offender as mitigating factors.” Id., at 589. The statute at issue, Ohio Rev. Code §§2929.03-2929.04(B) (1975), required the trial court to impose the death penalty upon Lockett’s conviction for “aggravated murder with specifications,”2 unless it found “that (1) the victim had induced or facilitated the offense, (2) it was unlikely that Lockett would have committed the offense but for the fact that she ‘was under duress, coercion, or strong provocation,’ or (3) the offense was ‘primarily the product of [Lockett’s] psychosis or mental deficiency.’” 438 U. S., at 593-594. It was plain that although guilty of felony homicide under Ohio law, Lockett had played a relatively minor role in a robbery which resulted in a homicide actually perpetrated by the hand of another. Lockett had previously committed no major offenses; in addition, a psychological report described her “prognosis for rehabilitation” as “favorable.” Id., at 594. However, since she was not found to have acted under duress, did not suffer from “psychosis,” and was not “mentally deficient,” the sentencing judge concluded that he had “‘no alternative, whether [he] like[d] the law or not’ but to impose the death penalty.” Ibid.
We held in Lockett that the “Eighth and Fourteenth Amendments require that the sentencer . . . 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 that the defendant proffers as a basis for a sentence less than death.” Id., at 604 (emphasis in original). We therefore found the Ohio statute flawed, be*122cause it did not permit individualized consideration of mitigating circumstances — such as the defendant’s comparatively minor role in the offense, lack of intent to kill the victim, or age. Id., at 606-608. We did not, however, undertake to dictate the weight that a sentencing court must ascribe to the various factors that might be categorized as “mitigating,” nor did we in any way suggest that this Court may substitute its sentencing judgment for that of state courts in capital cases.
In contrast to the Ohio statute at issue in Lockett, the Oklahoma death penalty statute provides:
“In the sentencing proceeding, evidence may be presented as to any mitigating circumstances or as to any of the aggravating circumstances enumerated in this act.” Okla. Stat., Tit. 21, §701.10 (1980) (emphasis added).
The statute further provides that
“[u]nless at least one of the statutory aggravating circumstances enumerated in this act is [found to exist beyond a reasonable doubt] or if it is found that any such aggravating circumstance is outweighed by the finding of one or more mitigating circumstances, the death penalty shall not be imposed.” §701.11.
This provision, of course, instructs the sentencer to weigh the mitigating evidence introduced by a defendant against the aggravating circumstances proved by the State.3
The Oklahoma statute thus contains provisions virtually identical to those cited with approval in Lockett, as examples of proper legislation which highlighted the Ohio statute’s “constitutional infirmities.” 438 U. S., at 606-607. Indeed, the Court does not contend that the Oklahoma sentencing *123provisions are inconsistent with Lockett. Moreover, the Court recognizes that, as mandated by the Oklahoma statute, Eddings was permitted to present “substantial evidence at the [sentencing] hearing of his troubled youth.” Ante, at 107.4
In its attempt to make out a violation of Lockett, the Court relies entirely on a single sentence of the trial court’s opinion delivered from the bench at the close of the sentencing hearing. After discussing the aggravated nature of petitioner’s offense, and noting that he had “given very serious consideration to the youth of the Defendant when this particular crime was committed,” the trial judge said that he could not
*124“be persuaded entirely by the . . . fact that the youth was sixteen years old when this heinous crime was committed. Nor can the Court in following the law, in my opinion, consider the fact of this young man’s violent background.” App. 189.
From this statement, the Court concludes “it is clear that the trial judge did not evaluate the evidence in mitigation and find it wanting as a matter of fact, rather he found that as a matter of law he was unable even to consider the evidence.” Ante, at 113. This is simply not a correct characterization of the sentencing judge’s action.
In its parsing of the trial court’s oral statement, the Court ignores the fact that the judge was delivering his opinion extemporaneously from the bench, and could not be expected to frame each utterance with the specificity and precision that might be expected of a written opinion or statute. Extemporaneous courtroom statements are not often models of clarity. Nor does the Court give any weight to the fact that the trial court had spent considerable time listening to the testimony of a probation officer and various mental health professionals who described Eddings’ personality and family history — an obviously meaningless, exercise if, as the Court asserts, the judge believed he was barred “as a matter of law” from “considering”, their testimony. Yet even examined in isolation, the trial court’s statement is at best ambiguous; 5 it can just as easily be read to say that, while the court *125had taken account of Eddings’ unfortunate childhood, it did not consider that either his youth or his family background was sufficient to offset the aggravating circumstances that the evidence revealed. Certainly nothing in Lockett would preclude the court from making such a determination.
The Oklahoma Court of Criminal Appeals independently examined the evidence of “aggravating” and “mitigating” factors presented at Eddings’ sentencing hearing. 616 P. 2d 1159 (1980). After reviewing the testimony concerning Eddings’ personality and family background, and after referring to the trial court’s discussion of mitigating circumstances, it stated that while Eddings’ “family history is useful in explaining why he behaved the way he did, ... it does not excuse his behavior.” Id., at 1170 (emphasis added). From this the Court concludes that “the Court of Criminal Appeals also considered only that evidence to be mitigating which would tend to support a legal excuse from criminal liability.” Ante, at 113.6 However, there is no reason to read that court’s statements as reflecting anything more than a conclusion that Eddings’ background was not a sufficiently mitigating factor to tip the scales, given the aggravating circumstances, including Eddings’ statements immediately before the killing.7 The Court of Criminal Appeals most assuredly did not, as the Court’s opinion suggests, hold that this “evidence in mitigation was not relevant,” see ibid.; indeed, had the Court of Criminal Appeals thought the evidence irrele*126vant, it is unlikely that it would have spent several paragraphs summarizing it. The Court’s opinion offers no reasonable explanation for its assumption that the Court of Criminal Appeals considered itself bound by some unstated legal principle not to “consider” Eddings’ background.
To be sure, neither the Court of Criminal Appeals nor the trial court labeled Eddings’ family background and personality disturbance as “mitigating factors.” It is plain to me, however, that this was purely a matter of semantics associated with the rational belief that “evidence in mitigation” must rise to a certain level of persuasiveness before it can be said to constitute a “mitigating circumstance.” In contrast, the Court seems to require that any potentially mitigating evidence be described as a “mitigating factor” — regardless of its weight; the insubstantiality of the evidence is simply to be a factor in the process of weighing the evidence against aggravating circumstances. Yet if this is all the Court’s opinion stands for, it provides scant support for the result reached. For it is clearly the choice of the Oklahoma courts — a choice not inconsistent with Lockett or any other decision of this Court — to accord relatively little weight to Eddings’ family background and emotional problems as balanced against the circumstances of his crime and his potential for future dangerousness.8
*127II
It can never be less than the most painful of our duties to pass on capital cases, and the more so in a case such as this one. However, there comes a time in every case when a court must “bite the bullet.”
Whether the Court’s remand will serve any useful purpose remains to be seen, for petitioner has already been given an opportunity to introduce whatever evidence he considered relevant to the sentencing determination. Two Oklahoma courts have weighed that evidence and found it insufficient to offset the aggravating circumstances shown by the State. The Court’s opinion makes clear that some Justices who join it would not have imposed the death penalty had they sat as the sentencing authority, see, e. g., ante, at 115-116. In*128deed, I am not sure I would have done so. But the Constitution does not authorize us to determine whether sentences imposed by state courts are sentences we consider “appropriate”; our only authority is to decide whether they are constitutional under the Eighth Amendment. The Court stops far short of suggesting that there is any constitutional proscription against imposition of the death penalty on a person who was under age 18 when the murder was committed. In the last analysis, the Court is forced to conclude that it is “the state courts [which] must consider [petitioner’s mitigating evidence] and weigh it against the evidence of the aggravating circumstances. We do not weigh the evidence for them.” Ante, at 117.
Because the sentencing proceedings in this case were in no sense inconsistent with Lockett v. Ohio, 438 U. S. 586 (1978), I would decide the sole issue on which we granted certiorari, and affirm the judgment.
The Court struggles to demonstrate that “the question of whether the decisions below were consistent with our decision in Lockett is properly before us.” Ante, at 113-114, n. 9. It argues that petitioner’s “Lockett claim” was somehow inherent in his general assertion that the death penalty was “excessive.” However, it is obvious that petitioner not only failed to present to this Court the question which the Court now addresses, but also never “fairly presented” the Lockett argument to the state courts so as to have afforded them the first “opportunity to apply controlling legal principles to the facts bearing upon fhis] constitutional claim.” Picard v. Connor, 404 U. S. 270, 275-277 (1971). Indeed, petitioner concedes as much, admitting that the “Lockett error was not enumerated or argued on appeal to the Oklahoma Court of Criminal Appeals . . . .” Brief for Petitioner 64.
In that case the evidence showed that while Lockett waited in a “getaway” car, her three companions robbed a store; during the robbery, the proprietor was fatally wounded. Lockett was charged with aggravated murder with two “specifications” of “aggravating circumstances”: (1) that the murder was “committed for the purpose of escaping detection, apprehension, trial, or punishment” for aggravated robbery; and (2) that the murder was “committed while . . . committing, attempting to commit, or fleeing immediately after committing or attempting to commit . . . aggravated robbery.” See Ohio Rev. Code § 2929.04(A) (1975).
It is ironic that in his petition for certiorari filed with the Oklahoma Court of Criminal Appeals, petitioner asserted that the Oklahoma sentencing scheme was constitutionally deficient, because “[t]he mitigating circumstances which may be considered are not statutorily defined or limited” (emphasis added).
Although I think it is immaterial to a correct decision of this case, it is worth noting that the Court overstates and oversimplifies the evidence presented by Eddings at the sentencing hearing. For example, it twice characterizes the testimony as indicating that, at the time of the crime, Eddings’ “mental and emotional development were at a level several years below his age.” Ante, at 107, 116. Dr. Dietsche, a psychologist, testified that if forced to extrapolate from the Wechsler Adult Intelligence Scale he would place petitioner’s “mental age” at about 14 years, 6 months; however, he then said that this mental age would have “no meaning” since “the mental age concepts break down . . . between fourteen to sixteen years of age.” He went on to state: “My opinion is that [Eddings] has the intelligence of an adult.” App. 134-136 (emphasis added). Describing a single interview with petitioner while he was awaiting trial on murder charges, Dr. Rettig, a sociologist, said that petitioner’s “responses appeared to me to be several years below his chronological age”; he “qualif[ied]” this answer, however, by noting that petitioner was “under a great deal of constraint in the atmosphere in which I saw him.” Id., at 149. Finally, Dr. Gagliano, a psychiatrist, opined on the basis of a one-hour interview — during which petitioner’s attorney was present and refused to allow questioning about petitioner’s “mental status” on the day of the shooting, id., at 177 — that at the time petitioner pulled the trigger, “he acted as a seven year old seeking revenge and rebellion” against his stepfather, a policeman. Id., at 172-173. Dr. Gagliano was also willing to state categorically, on the basis of this single interview, and without reference to the results of the psychological testing of Eddings, id., at 174, that Eddings was “preordained” to commit the murder from the time his parents were divorced, when he was five. Id., at 179-180. This sort of “determinist” approach is rejected by an overwhelming majority of psychiatrists.
It is not even clear what the trial court meant by Eddings’ “violent background." For example, Eddings’ probation officer testified that Eddings had “problems with fighting” while in school, and had once been charged with “Assault with intent to do great bodily harm.” Id., at 106-107. The State seems to concede, however, that the court was probably referring, at least in part, to Eddings’ family history. See Brief for Respondent 55 (“the inference that can be drawn is that the court did not consider petitioner’s juvenile record and family life to be a mitigating circumstance”) (emphasis added). But cf. Tr. of Oral Arg. 35 (“the remark is *125ambiguous. It could be interpreted to mean that [the trial court] was not going to consider the juvenile’s previous juvenile record in Missouri, which was extensive . .
On the other hand, the Court’s opinion concedes that petitioner’s youth was given serious consideration as a “mitigating circumstance,” although his age at the time of the offense would not “tend to support a legal excuse from criminal responsibility.”
When Eddings’ companions informed him that the officer’s patrol car was approaching, Eddings responded that if the “mother . . . pig tried to stop him he was going to blow him away.” App. 66.
Nor is this choice necessarily an unreasonable one. As the Court notes, “[ejvidence of a difficult family history and of emotional disturbance is typically introduced by defendants in mitigation.” Ante, at 115. One might even be surprised if a person capable of a brutal and unprovoked killing of a police officer did not suffer from some sort of “personality disorder.”
Indeed, Dr. Dietsche, who testified that Eddings had a “sociopathic or antisocial personality,” see ante, at 107, estimated that 91% “of your criminal element” would test as sociopathic or antisocial. App. 136. Dr. Dietsche defined “antisocial personalities” as individuals without “the usual type of companions” or “loyalties,” who are “[f]requently . . . selfish, . . . very impulsive,” showing “little in the line of responsibility” or concern “for the needs or wants of others,” and “hav[ing] little in the line of guilt or *127remorse.” Id., at 137-138. Although the Court describes Dietsche’s testimony as indicating that “approximately 30% of youths suffering from such a disorder grew out of it as they aged,” ante, at 107, Dietsche was in fact describing a study which he thought had subsequently been discredited. App. 139-141. Even that study, however, concluded that most of those who “grew out of” the disorder by the age of 35 or 40 were “more of a con-artist type” and “not. . . the assaultive type.” Ibid. A more recent study estimated that only 20% of sociopathic persons were “treatable,” id., at 141; in this study, only 9 of 255 initial participants were successfully treated, after “literally . . . thousands of hours of therapy.” Id., at 142. Thus, characterization of Eddings as a “sociopath” may connote little more than that he is egocentric, concerned only with his own desires and unremorseful, has a propensity for criminal conduct, and is unlikely to respond well to conventional psychiatric treatment — hardly significant “mitigating” factors. See Blocker v. United States, 110 U. S. App. D. C. 41, 48-49, and nn. 11, 12, 288 F. 2d 853, 860-861, and nn. 11, 12 (1961) (Burger, J., concurring in result). While the Court speaks of Eddings’ “severe emotional disturbance,” ante, at 115; see also ante, at 116, it appears to be referring primarily to the testimony that Eddings was a sociopath, and to Dr. Gagliano’s rather fantastic speculation concerning Eddings’ dissociation at the time of the crime, see n. 4, supra. The Court’s opinion exemplifies the proposition that the very occurrence of the crime functions as a powerful impetus to search for a theory to explain it. See Szasz, Psychiatry, Ethics, and the Criminal Law, 58 Colum. L. Rev. 183, 190-191 (1958).