with whom Justice Blackmun joins, dissenting.
The novel rule that the Court announces today represents a dramatic departure from the principles that have governed our capital sentencing jurisprudence for decades. Justice Marshall is properly concerned about the majority’s trivialization of the doctrine of stare decisis. But even if Booth v. Maryland, 482 U. S. 496 (1987), and South Carolina v. Gathers, 490 U. S. 805 (1989), had not been decided, today’s decision would represent a sharp break with past decisions. Our cases provide no support whatsoever for the majority’s conclusion that the prosecutor may introduce evidence that sheds no light on the defendant’s guilt or moral culpability, and thus serves no purpose other than to encourage jurors to decide in favor of death rather than life on the basis of their emotions rather than their reason.
Until today our capital punishment jurisprudence has required that any decision to impose the death penalty be based solely on evidence that tends to inform the jury about the character of the offense and the character of the defendant. Evidence that serves no purpose other than to appeal to the *857sympathies or emotions of the jurors has never been considered admissible. Thus, if a defendant, who had murdered a convenience store clerk in cold blood in the course of an armed robbery, offered evidence unknown to him at the time of the crime about the immoral character of his victim, all would recognize immediately that the evidence was irrelevant and inadmissible. Evenhanded justice requires that the same constraint be imposed on the advocate of the death penalty.
I
In Williams v. New York, 337 U. S. 241 (1949), this Court considered the scope of the inquiry that should precede the imposition of a death sentence. Relying on practices that had developed “both before and since the American colonies became a nation,” id., at 246, Justice Black described the wide latitude that had been accorded judges in considering the source and type of evidence that is relevant to the sentencing determination. Notably, that opinion refers not only to the relevance of evidence establishing the defendant’s guilt, but also to the relevance of “the fullest information possible concerning the defendant’s life and characteristics.” Id., at 247. “Victim impact” evidence, however, was unheard of when Williams was decided. The relevant evidence of harm to society consisted of proof that the defendant was guilty of the offense charged in the indictment.
Almost 30 years after our decision in Williams, the Court reviewed the scope of evidence relevant in capital sentencing. See Lockett v. Ohio, 438 U. S. 586 (1978). In his plurality opinion, Chief Justice Burger concluded that in a capital case, the sentencer must not be prevented “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 deleted). As in Williams, the character of the offense and the character of the offender constituted *858the entire category of relevant evidence. “Victim impact” evidence was still unheard of when Lockett was decided.
As the Court acknowledges today, the use of victim impact evidence “is of recent origin,” ante, at 821. Insofar as the Court’s jurisprudence is concerned, this type of evidence made its first appearance in 1987 in Booth v. Maryland, 482 U. S. 496. In his opinion for the Court, Justice Powell noted that our prior cases had stated that the question whether an individual defendant should be executed is to be determined on the basis of “ ‘the character of the individual and the circumstances of the crime,’” id., at 502 (quoting Zant v. Stephens, 462 U. S. 862, 879 (1983)). See also Eddings v. Oklahoma, 455 U. S. 104, 112 (1982). Relying on those cases and on Enmund v. Florida, 458 U. S. 782, 801 (1982), the Court concluded that unless evidence has some bearing on the defendant’s personal responsibility and moral guilt, its admission would create a risk that a death sentence might be based on considerations that are constitutionally impermissible or totally irrelevant to the sentencing process. 482 U. S., at 502. Evidence that served no purpose except to describe the personal characteristics of the victim and the emotional impact of the crime on the victim’s family was therefore constitutionally irrelevant.
Our decision in Booth was entirely consistent with the practices that had been followed “both before and since the American colonies became a nation,” Williams, 337 U. S., at 246. Our holding was mandated by our capital punishment jurisprudence, which requires any decision to impose the death penalty to be based on reason rather than caprice or emotion. See Gardner v. Florida, 430 U. S. 349, 362 (1977) (opinion of Stevens, J.). The dissenting opinions in Booth and in Gathers can be searched in vain for any judicial precedent sanctioning the use of evidence unrelated to the character of the offense or the character of the offender in the sentencing process. Today, however, relying on nothing more than those dissenting opinions, the Court abandons *859rules of relevance that are older than the Nation itself and ventures into uncharted seas of irrelevance.
I — I I — (
Today’s majority has obviously been moved by an argument that has strong political appeal but no proper place in a reasoned judicial opinion. Because our decision in Lockett, 438 U. S., at 604 (opinion of Burger, C. J.), recognizes the defendant’s right to introduce all mitigating evidence that may inform the jury about his character, the Court suggests that fairness requires that the State be allowed to respond with similar evidence about the victim. See ante, at 825-826.1 This argument is a classic non sequitur: The victim is not on trial; her character, whether good or bad, cannot therefore constitute either an aggravating or a mitigating circumstance.
*860Even if introduction of evidence about the victim could be equated with introduction of evidence about the defendant, the argument would remain flawed in both its premise and its conclusion. The conclusion that exclusion of victim impact evidence results in a significantly imbalanced sentencing procedure is simply inaccurate. Just as the defendant is entitled to introduce any relevant mitigating evidence, so the State may rebut that evidence and may designate any relevant conduct to be an aggravating factor provided that the factor is sufficiently well defined and consistently applied to cabin the sentencer’s discretion.
The premise that a criminal prosecution requires an evenhanded balance between the State and the defendant is also incorrect. The Constitution grants certain rights to the criminal defendant and imposes special limitations on the State designed to protect the individual from overreaching by the disproportionately powerful State. Thus, the State must prove a defendant’s guilt beyond a reasonable doubt. See In re Winship, 397 U. S. 358 (1970). Rules of evidence are also weighted in the defendant’s favor. For example, the prosecution generally cannot introduce evidence of the defendant’s character to prove his propensity to commit a crime, but the defendant can introduce such reputation evidence to show his law-abiding nature. See, e. g., Fed. Rule Evid. 404(a). Even if balance were required or desirable, today’s decision, by permitting both the defendant and the State to introduce irrelevant evidence for the sentencer’s consideration without any guidance, surely does nothing to enhance parity in the sentencing process.
1 — 1 h — I hH
Victim impact evidence, as used in this case, has two flaws, both related to the Eighth Amendment’s command that the punishment of death may not be meted out arbitrarily or capriciously. First, aspects of the character of the victim unforeseeable to the defendant at the time of his crime are irrel*861evant to the defendant’s “personal responsibility and moral guilt” and therefore cannot justify a death sentence. See Enmund v. Florida, 458 U. S., at 801; see also id., at 825 (O’Connor, J., dissenting) (“[P]roportionality requires a nexus between the punishment imposed and the defendant’s blameworthiness”); Tison v. Arizona, 481 U. S. 137, 149 (1987) (“The heart of the retribution rationale is that a criminal sentence must be directly related to the personal culpability of the criminal offender”); California v. Brown, 479 U. S. 538, 545 (1987) (O’Connor, J., concurring).
Second, the quantity and quality of victim impact evidence sufficient to turn a verdict of life in prison into a verdict of death is not defined until after the crime has been committed and therefore cannot possibly be applied consistently in different cases. The sentencer’s unguided consideration of victim impact evidence thus conflicts with the principle central to our capital punishment jurisprudence that, “where discretion is afforded a sentencing body on a matter so grave as the determination of whether a human life should be taken or spared, that discretion must be suitably directed and limited so as to minimize the risk of wholly arbitrary and capricious action.” Gregg v. Georgia, 428 U. S. 153, 189 (1976) (joint opinion of Stewart, Powell, and Stevens, JJ.). Open-ended reliance by a capital sentencer on victim impact evidence simply does not provide a “principled way to distinguish [cases], in which the death penalty [i]s imposed, from the many cases in which it [i]s not.” Godfrey v. Georgia, 446 U. S. 420, 433 (1980) (opinion of Stewart, J.).
The majority attempts to justify the admission of victim impact evidence by arguing that “consideration of the harm caused by the crime has been an important factor in the exercise of [sentencing] discretion.” Ante, at 820. This statement is misleading and inaccurate. It is misleading because it is not limited to harm that is foreseeable. It is inaccurate because it fails to differentiate between legislative determinations and judicial sentencing. It is true that an evaluation of *862the harm caused by different kinds of wrongful conduct is a critical aspect in legislative definitions of offenses and determinations concerning sentencing guidelines. There is a rational correlation between moral culpability and the foreseeable harm caused by criminal conduct. Moreover, in the capital sentencing area, legislative identification of the special aggravating factors that may justify the imposition of the death penalty is entirely appropriate.2 But the majority cites no authority for the suggestion that unforeseeable and indirect harms to a victim’s family are properly considered as aggravating evidence on a case-by-case basis.
The dissents in Booth and Gathers and the majority today offer only the recent decision in Tison v. Arizona, 481 U. S. 137 (1987), and two legislative examples to support their contention that harm to the victim has traditionally influenced sentencing discretion. Tison held that the death penalty may be imposed on a felon who acts with reckless disregard for human life if a death occurs in the course of the felony, even though capital punishment cannot be imposed if no one dies as a result of the crime. The first legislative example is that attempted murder and murder are classified as two different offenses subject to different punishments. Ante, at 819. The second legislative example is that a person who drives while intoxicated is guilty of vehicular homicide if his actions result in a death but is not guilty of this offense if he has the good fortune to make it home without killing anyone. See Booth, 482 U. S., at 516 (White, J., dissenting).
*863These three scenarios, however, are fully consistent with the Eighth Amendment jurisprudence reflected in Booth and Gathers and do not demonstrate that harm to the victim may be considered by a capital sentencer in the ad hoc and post hoc manner authorized by today’s majority. The majority’s examples demonstrate only that harm to the victim may justify enhanced punishment if the harm is both foreseeable to the defendant and clearly identified in advance of the crime by the legislature as a class of harm that should in every case result in more severe punishment.
In each scenario, the defendants could reasonably foresee that their acts might result in loss of human life. In addition, in each, the decision that the defendants should be treated differently was made prior to the crime by the legislature, the decision of which is subject to scrutiny for basic rationality. Finally, in each scenario, every defendant who causes the well-defined harm of destroying a human life will be subject to the determination that his conduct should be punished more severely. The majority’s scenarios therefore provide no support for its holding, which permits a jury to sentence a defendant to death because of harm to the victim and his family that the defendant could not foresee, which was not even identified until after the crime had been committed, and which may be deemed by the jury, without any rational explanation, to justify a death sentence in one case but not in another. Unlike the rule elucidated by the scenarios on which the majority relies, the majority’s holding offends the Eighth Amendment because it permits the sen-tencer to rely on irrelevant evidence in an arbitrary and capricious manner.
The majority’s argument that “the sentencing authority has always been free to consider a wide range of relevant material,” ante, at 820-821 (emphasis added), thus cannot justify consideration of victim impact evidence that is irrelevant because it details harms that the defendant could not have foreseen. Nor does the majority’s citation of Gregg v. Geor*864gia concerning the “wide scope of evidence and argument allowed at presentence hearings,” 428 U. S., at 203 (joint opinion of Stewart, Powell, and Stevens, JJ.), support today’s holding. See ante, at 821. The Gregg joint opinion endorsed the sentencer’s consideration of a wide range of evidence “[s]o long as the evidence introduced and the arguments made at the presentence hearing do not prejudice a defendant.” 428 U. S., at 203-204. Irrelevant victim impact evidence that distracts the sentencer from the proper focus of sentencing and encourages reliance on emotion and other arbitrary factors necessarily prejudices the defendant.
The majority’s apparent inability to understand this fact is highlighted by its misunderstanding of Justice Powell’s argument in Booth that admission of victim impact evidence is undesirable because it risks shifting the focus of the sentencing hearing away from the defendant and the circumstances of the crime and creating a “ ‘mini-trial’ on the victim’s character.” 482 U. S., at 507. Booth found this risk insupportable not, as today’s majority suggests, because it creates a “tactical” “dilemma” for the defendant, see ante, at 823, but because it allows the possibility that the jury will be so distracted by prejudicial and irrelevant considerations that it will base its life-or-death decision on whim or caprice. See 482 U. S., at 506-507.
IV
The majority thus does far more than validate a State’s judgment that “the jury should see ‘a quick glimpse of the life petitioner chose to extinguish,’ Mills v. Maryland, 486 U. S. 367, 397 (1988) (Rehnquist, C. J., dissenting).” Ante, at 830-831 (O’Connor, J., concurring). Instead, it allows a jury to hold a defendant responsible for a whole array of harms that he could not foresee and for which he is therefore not blameworthy. Justice Souter argues that these harms are sufficiently foreseeable to hold the defendant accountable because “[e]very defendant knows, if endowed with the mental competence for criminal responsibility, that *865the life he will take by his homicidal behavior is that of a unique person, like himself, and that the person to be killed probably has close associates, ‘survivors,’ who will suffer harms and deprivations from the victim’s death.” Ante, at 838 (Souter, J., concurring). But every juror and trial judge knows this much as well. Evidence about who those survivors are and what harms and deprivations they have suffered is therefore not necessary to apprise the sentencer of any information that was actually foreseeable to the defendant. Its only function can be to “divert the jury’s attention away from the defendant’s background and record, and the circumstances of the crime.” See Booth, 482 U. S., at 505.
Arguing in the alternative, Justice Souter correctly points out that victim impact evidence will sometimes come to the attention of the jury during the guilt phase of the trial. Ante, at 840. He reasons that the ideal of basing sentencing determinations entirely on the moral culpability of the defendant is therefore unattainable unless a different jury is empaneled for the sentencing hearing. Ante, at 841. Thus, to justify overruling Booth, he assumes that the decision must otherwise be extended far beyond its actual holding.
Justice Souter’s assumption is entirely unwarranted. For as long as the contours of relevance at sentencing hearings have been limited to evidence concerning the character of the offense and the character of the offender, the law has also recognized that evidence that is admissible for a proper purpose may not be excluded because it is inadmissible for other purposes and may indirectly prejudice the jury. See 1 J. Wigmore, Evidence § 13 (P. Tillers rev. 1983). In the case before us today, much of what might be characterized as victim impact evidence was properly admitted during the guilt phase of the trial and, given the horrible character of this crime, may have been sufficient to justify the Tennessee Supreme Court’s conclusion that the error was harmless because the jury would necessarily have imposed the death sentence even absent the error. The fact that a good deal of *866such evidence is routinely and properly brought to the attention of the jury merely indicates that the rule of Booth may not affect the outcome of many cases.
In reaching our decision today, however, we should not be concerned with the cases in which victim impact evidence will not make a difference. We should be concerned instead with the cases in which it will make a difference. In those cases, defendants will be sentenced arbitrarily to death on the basis of evidence that would not otherwise be admissible because it is irrelevant to the defendants’ moral culpability. The Constitution’s proscription against the arbitrary imposition of the death penalty must necessarily proscribe the admission of evidence that serves no purpose other than to result in such arbitrary sentences.
V
The notion that the inability to produce an ideal system of justice in which every punishment is precisely married to the defendant’s blameworthiness somehow justifies a rule that completely divorces some capital sentencing determinations from moral culpability is incomprehensible to me. Also incomprehensible is the argument that such a rule is required for the jury to take into account that each murder victim is a “unique” human being. See ante, at 823; ante, at 830-831 (O’Connor, J., concurring); ante, at 838 (Souter, J., concurring). The fact that each of us is unique is a proposition so obvious that it surely requires no evidentiary support. What is not obvious, however, is the way in which the character or reputation in one case may differ from that of other possible victims. Evidence offered to prove such differences can only be intended to identify some victims as more worthy of protection than others. Such proof risks decisions based on the same invidious motives as a prosecutor’s decision to seek the death penalty if a victim is white but to accept a plea bargain if the victim is black. See McCleskey v. Kemp, 481 U. S. 279, 366 (1987) (Stevens, J., dissenting).
*867Given the current popularity of capital punishment in a crime-ridden society, the political appeal of arguments that assume that increasing the severity of sentences is the best cure for the cancer of crime, and the political strength of the “victims’ rights” movement, I recognize that today’s decision will be greeted with enthusiasm by a large number of concerned and thoughtful citizens. The great tragedy of the decision, however, is the danger that the “hydraulic pressure” of public opinion that Justice Holmes once described3 — and that properly influences the deliberations of democratic legislatures — has played a role not only in the Court’s decision to hear this case,4 and in its decision to reach the constitutional question without pausing to consider affirming on the basis of the Tennessee Supreme Court’s rationale,5 but even in its resolution of the constitutional issue involved. Today is a sad day for a great institution.
Justice Scalia accurately described the argument in his dissent in Booth v. Maryland, 482 U. S. 496 (1987):
“Recent years have seen an outpouring of popular concern for what has come to be known as ‘victims’ rights’ — a phrase that describes what its proponents feel is the failure of courts of justice to take into account in their sentencing decisions not only the factors mitigating the defendant’s moral guilt, but also the amount of harm he has caused to innocent members of society. Many citizens have found one-sided and hence unjust the criminal trial in which a parade of witnesses comes forth to testify to the pressures beyond normal human experience that drove the defendant to commit his crime, with no one to lay before the sentencing authority the full reality of human suffering the defendant has produced — which (and not moral guilt alone) is one of the reasons society deems his act worthy of the prescribed penalty.” Id., at 520.
In his concurring opinion today, Justice Scalia again relies on the popular opinion that has “found voice in a nationwide ‘victims’ rights’ movement.” Ante, at 834. His view that the exclusion of evidence about “a crime’s unanticipated consequences” “significantly harms our criminal justice system,” ibid., rests on the untenable premise that the strength of that system is to be measured by the number of death sentences that may be returned on the basis of such evidence. Because the word “arbitrary” is not to be found in the constitutional text, he apparently can find no reason to object to the arbitrary imposition of capital punishment.
Thus, it is entirely consistent with the Eighth Amendment principles underlying Booth and South Carolina v. Gathers, 490 U. S. 805 (1989), to authorize the death sentence for the assassination of the President or Vice President, see 18 U. S. C. §§ 1751, 1111, a Congressman, Cabinet official, Supreme Court Justice, or the head of an executive department, § 351, or the murder of a policeman on active duty, see Md. Ann. Code, Art. 27, § 413(d)(1) (1987). Such statutory provisions give the potential offender notice of the special consequences of his crime and ensure that the legislatively determined punishment will be applied consistently to all defendants.
Northern Securities Co. v. United States, 193 U. S. 197, 400-401 (1904) (Holmes, J., dissenting).
See Payne v. Tennessee, 498 U. S. 1076 (1991) (Stevens, J., dissenting).
Rust v. Sullivan, 500 U. S. 173, 223 (1991) (O’Connor, J., dissenting).